table of content no2 - oradea - descriptionenergy-cie.ro/archives/2012/jse_vol_3_no_2.pdf ·...

Editorial Board

JOURNAL OF SUSTAINABLE ENERGY EDITOR IN CHIEF Felea Ioan – Member of I.E.E.E.

University of Oradea, Department of Energy Engineering, [email protected]

EDITORS Gleb Drăgan

Member of Romanian Academy Florin Gheorghe Filip

Member of Romanian Academy [email protected]

Cornel Antal University of Oradea, [email protected]

Anatolie Carabulea Politehnica University of Bucharest

Gianfranco Chicco Politecnico de Torino , Italia [email protected]

Roberto Cipollone(details) University of L’Aquila

[email protected]

Fiodor Erchan State Agricultural University, Moldova

[email protected]

Mihai Jădăneanţ Politehnica University of Timisoara

[email protected]

Ştefan Kilyeni Politehnica University of Timisoara

[email protected]

Gheorghe Lăzăroiu Politehnica University of Bucharest

[email protected]

Carlo Mazetti La Sapienza di Roma

[email protected]

Victori�a Rădulescu Politehnica University of Bucharest

[email protected]

Florin Popenţiu University of Oradea

[email protected]

Jacques Padet Universite de Reims, France [email protected]

Paulo F. Ribeiro (details) Grand Rapids, Michigan

[email protected]

Marcel Roşca University of Oradea [email protected]

Saroudis J. AECL, CANDU Services

[email protected]

Takács János (details) Technical of University Bratislava

[email protected]

Victor Vaida University of Oradea [email protected]

Varju György (details) Budapest University of Technology&Economics

[email protected]

Kalmar Ferenc University of Debrecen [email protected]

Mircea Vereş University of Oradea [email protected]

Alexandru Vasilievici Politehnica University of Timisoara

[email protected]

Badea Gabriela University of Oradea, [email protected]

Nikolai Voropai Energy Systems Institute, Russia

[email protected]

Irena Wasiak (details) Technical University of Lodz

[email protected]

Dan Zlatanovici ICEMENERG

[email protected]

Gabriel Bendea University of Oradea

[email protected]

Nicolae Coroiu University of Oradea

[email protected]

Laurenţiu Popper University of Oradea

[email protected]

Călin Secui University of Oradea [email protected]

Zétényi Zsigmond

University of Oradea [email protected]

EXECUTIVE STAFF

Executive Editor: Dziţac Simona

University of Oradea, [email protected]

Editorial secretary Albuţ-Dana Daniel


Technical Secretary Vasile Moldovan


Editorial Activities Barla Eva


PUBLISHER & EDITORIAL OFFICE University of Oradea Editing House,

Str. Universitatii Nr. 1, Oradea, jud. Bihor, România, Zip code: 410087, Tel.: 00-40-259-408171, Fax: 00-40-259-408404 ISSN: 2067-5534 (print version)

74

The Journal of Sustainable Energy (JSE) had its first appearance under this name in 2010. The history of JSE is the following: is formed in 2010 by transforming of 1993-2009 Analele Universitatii din Oradea. Fascicula de Energetica, 1224-1261. Which superseded in part (1991-1992): Analele Universitatii din Oradea. Fascicula Electrotehnica si Energetica (1221-1311); Which superseded in part (1976-1990): Lucrari Stiintifice - Institutul de Invatamant Superior Oradea. Seria A, Stiinte Tehnice, Matematica, Fizica, Chimie, Geografie (0254-8593);

The Journal of Sustainable Energy (JSE) publishes original contributions in the field of the following topics: Energy engineering education System reliability and power service quality Generation of electric and thermal power Energy policy and economics Energy development (solar power, renewable energy, waste-to-energy systems) Energy systems operation Energy efficiency, reducing consumption for conservation of energy Energy sustainability as related to energy and power production, distribution and usage Waste management and environmental issues Energy infrastructure issues (power plant safety, security of infrastructure network) Energetic equipments The articles quality increases with every issue. Apart from its improved technical contents, a special care is given to

its structure. The guidelines for preparing and submitting an article were modified and developed in order to meet high quality standards requirements. With every passing issue the peer-review process was developed too, being now a double-blinded peer-review process.

Authors who wish to submit a manuscript to the Journal of Sustainable Energy (JSE) are kindly asked to send their

manuscripts as DOC and PDF format to [email protected], according to the formatting instructions. Once the manuscript is submitted, the authors will shortly receive a feedback regarding the status of their submission. As the review process is completed, the author will be informed about the reviewers’ comments and the changes the paper should suffer in order to satisfy the journal quality requirements.

Journal of Sustainable Energy JSE is covered/indexed/abstracted in: Index Copernicus Ulrich's Update - Periodicals Directory DOAJ - Directory of Open Access Journals EBSCO Publishing - EBSCOhost Online Research Databases Engineering village (Pending)

The JSE may be purchased based on annual subscription (4 issues) at the following prices: paper – 50 � ,

electronic – 20 �, or individually (each number) at the following prices: paper - 20 �, electronic – 8 �. To purchase the complete “control” standard (www.energy-cie.ro) and submitted electronically or by fax. In order to generate the invoice (www.energy-cie.ro) is transmitted to the applicant. After payment of the amount stated in the bill the applicant receives and invoice numbers ordered JSE. Publishing House name/address: University of Oradea Editing House Universitatea din Oradea, Universitatii Str., No. 1, 410087, Oradea, Bihor, Romania

ISSN: 2067-5534 Tel.: 00-40-259-408171 (231, 288) Fax: 00-40-259-408404 Place of publishing: Oradea, Romania Year of the foundation of publication in domain of power engineering: 1976 Releasing frequency: 4 / year Language: English

75

CONTENTS RELIABILITY AND SYSTEMS ENERGY QUALITY SERVICES ANALISING SOME OF THE EXISTING RISK ASSESSMENT AND MANAGEMENT STANDARDS APPLIED WORLDWIDE, FOR ENERGY COMPANIES VORONCA S. L....................................................................................................................................................................................77 ARTIFICIAL INTELLIGENCE TECHNIQUES FOR DIAGNOSING POWER TRANSFORMERS’ FAULTS CIULAVU C.A. ....................................................................................................................................................................................85 RENEWABLE SOURCES OF ENERGY. SUSTAINABLE ENERGY TECHNOLOGIES RECENT DEVELOPMENTS IN SITE INVESTIGATION, DESIGN AND APPLICATION FOR GROUND SOURCE HEAT PUMPS SANNER B., MANDS D., SAUER M., GRUNDMANN E. ................................................................................................................91 ASPECTS RELATED TO WIND POWER PLANTS OPERATION IN ROMANIA OPREA S. V., PETRESCU D.E., BOLBORICI D., STĂNESCU O.R...............................................................................................101 GROUND SOURCE HEAT PUMPS FOR MEDITERRANEAN CLIMATE: PRESENT STATUS AND EXPECTATIONS FROM THE GROUND-MED PROJECT MENDRINOS D., KARYTSAS C. .....................................................................................................................................................107 MODELING OF WIND POWER PLANTS GENERATORS IN TRANSIENT STABILITY ANALYSIS STANESCU O., BOLBORICI D., OPREA S. ....................................................................................................................................114 NITROGEN COMPOUNDS REMOVAL FROM MUNICIPAL WASTEWATER CRET P., LOLEA M., HODISAN S.. .................................................................................................................................................120

EVOLUTION OF POWER ELECTRIC SYSTEMS TRANSPORT AND DISTRIBUTION. ENERGY SYSTEM’S PERFORMANCE AN INNOVATIVE MODEL FOR HIGH ACCURACY WIND POWER PREDICTION ALBU R.-D. POPENŢIU-VLĂDICESCU F.......................................................................................................................................125 INFLUENCE FACTORS ON ENERGY INTENSITY AGRICULTURAL PRODUCTS ROTARI V. .........................................................................................................................................................................................131 MARKET AND STOCK-MARKET OF POWER. MANAGEMENT OF POWER SYSTEMS SIMULATION AND ENERGY EFFICIENCY EVALUATION OF A LOW-ENERGY BUILDING VLAD. G.E., IONESCU C., NECULA H. ..........................................................................................................................................133

JOURNAL OF SUSTAINABLE ENERGY VOL. III, NO. 2, JUNE, 2012

ISSN 2067-5534 © 2012 JSE

ANALISING SOME OF THE EXISTING RISK ASSESSMENT AND MANAGEMENT STANDARDS APPLIED WORLDWIDE, FOR

ENERGY COMPANIES

VORONCA S. L. CN Transelectrica SA,Bucuresti [email protected]

Abstract - The paper reviews and categorizes the most relevant standards, guidelines, frameworks and methodologies about risk assessment and management, that could be useful to enhance energy companies risk assessment practices, allowing further on to determine a holistic, integrated method, applicable to operators of energy critical infrastructures. It will be presented the criteria used in evaluation methodologies - especially on how to handle threats and vulnerabilities, standards and prescription coverage, along all risk management process - identification, quantification, prioritization, treatment and control, and areas of specific application. Keywords: risk, energy companies, standards. 1. INTRODUCTION

The objective of this paper is to review and categorize the most relevant standards, guidelines, frameworks and methodologies about risk assessment and management, that could be useful to enhance energy companies risk assessment practices. The paper summarised a survey realised in 2011, in the framework of the EU project, Energy Control Center Risk Assessment and Management Methodology - ECCRAMM, developed by Symantec and Booz & Company, with Transelectrica acting as partner and end-user, project related to the enhancement of risk assessment methodologies against emerging threats for the protection of Energy SCADA systems, with specific focus on Energy Control Centers (ECC).

In performing this analysis, it has been considered and reviewed a great variety of different types of documents and publications, so to have a perspective as comprehensive as possible. It was considered standards and guidelines issued by global, regional and national organizations, consequently, it was signalled the geographic applicability of the various documents.

There were assessed documents which perform analyses at different levels of detail and specificity. Some of the analyzed standards and guidelines have a generalist approach and provide risk management methodologies applicable to virtually every kind of industry. Others focus either on the risks related to a certain type of industry (e.g. energy) or on a specific type

of risk (e.g. IT risk). It was taken into account both more general

standards and documents with deeper technical content. Thus, the standards and guidelines analyzed may be directed to different audiences (e.g. manufacturers, operators, security managers) and may have different purposes (e.g. dealing with technical Issues; evaluation and certification).

Focused on risk assessment and management, the analyse presents, for the documents, the areas and the phases of the risk management process that are covered; then are considered whether and to what extent the standards and guidelines assess the threats and vulnerabilities which may affect an energy companies and whether or not they take into account the possible controls which are implemented to reduce the vulnerabilities.

The main results and findings of this review of the existing standards, guidelines and frameworks about risk assessment and management for energy companies, could be summarized, as follows:

1.1. Areas and fields covered, general information

over the main topics and areas covered by the analyzed standards and guidelines. The documents are grouped and classified by the industry / field addressed. The standards/guidelines which have a general purpose and that are applicable across industries have been classified as “Global”. The other relevant elements included in the table are the geographic relevance, the type of risk covered and the audience targeted which allow having a first understanding of the relevance and of the applicability of the documents analyzed.

1.2. Risk management process coverage: analyzes

for each document the phases of the risk management process (i.e. risk analysis, risk evaluation, risk reporting and risk monitoring & control) that it covers and whether it addresses them through qualitative or quantitative methodologies. Thus, allows to quickly find information on what documents may be relevant for dealing with a certain aspect of the risk management process and what standards/guidelines propose mitigation strategies and tools to deal with different types of risks and it also indicates the documents which may be used for auditing purposes.

77


ISSN 2067-5534 © 2012 JSE

1.3. Threats and vulnerabilities coverage, allowing to go more in detail in the process of risk assessment. The survey signals what documents presented categorizations or at least examples of threats and vulnerabilities which are relevant for energy companies. It is indicated whether the documents provide only a general categorization of threats and vulnerabilities or whether they cover more in depth some specific types of threats and vulnerabilities; also analysed what standards and guidelines evaluate the possible controls which are in place to reduce the vulnerabilities.

2. MAIN ORGANIZATIONS DEFINING STANDARDS OR GUIDELINES RELEVANT FOR ENERGY COMPANIES RISK ASSESSMENT AND MANAGEMENT

The following section lists and gives a short

description of the main institutions relevant for energy critical infrastructures risk assessment and management. Beside international activities, also national activities of the United States and of some European countries are reported. Table 1. Main institutions relevant for energy companies risk assessment and management Institution Geographic

Relevance CIGRE (International Council on Large Electric Systems)

Global

IEC (International Electrotechnical Commission)

Global

IEEE (Institute of Electrical and Electronics Engineers

Global

IRGC (International Risk Governance Council)

Global

ISA (International Society of Automation

Global

ISACA (Information Systems Audit and Control Association

Global

ISO (International Organization for Standardization

Global

ACC (American Chemistry Council) United States AGA (American Gas Association) United States API (American Petroleum Institute) United States COSO (Committee of Sponsoring Organizations of the Treadway Commission

United States

DOE (US Department of Energy) United States DHS (US Department of Homeland Security)

United States

ES-ISAC (Electricity Sector - Information Sharing and Analysis Center)

United States

Idaho National Laboratory United States NERC (North American Electric Reliability Corporation)

United States

NIST (National Institute of Standards and Technology)

United States

BSI (Bundesamt für Sicherheit in der Informationstechnik)

Europe

CPNI (Centre for the Protection of National Infrastructure)

United Kingdom

ENTSO-E (European Network of Transmission System Operators for Electricity)

Europe

EURELETRIC (Union of the Electricity Industry)

Europe

SEMA (Swedish Emergency Management Agency)

Sweden

ITSEAG (Information Technology Security Expert Advisory Group)

Australia

The institutions were classified according to this template: Table 2. Template Name Organizational form

e.g. non-profit organization; industry association; government agency/department

Geographic relevance

e.g. International, EU, US

Year of foundation

e.g. 2002

Composition Number and type of members; sponsors; main bodies/organizational units

Key executives e.g. President, CEO, General Secretary, General Director

Industry / field

e.g. oil and gas; cross industry; automation; risk governance; security

Areas of activity

Short description of the main activities

Table 3. An example Name ISO (International Organization

for Standardization) Organizational form

non-governamental organization

Geographic relevance

Global

Year of foundation

1947

Composition Members are the national standards institutes of 163 countries.

Key executives Alan Morrison (President); Rob Steele (CEO); Jacob Holmblad (Vice-President, technical management); Sadao Takeda (Vice-President, policy).

Industry / field

Cross industry

Areas of activity

ISO is the world's largest developer and publisher of International Standards. It has developed over 17500 International Standards on a variety of subjects and more than 1000 new ISO standards are published every year. ISO also publishes technical reports, technical

78


ISSN 2067-5534 © 2012 JSE

specifications, publicly available specifications and guides. ISO, together with IEC, created the ISO/IEC Joint Technical Committee which deals with all matters of information technology. Its purpose is to develop, maintain, promote and facilitate IT standards required by global markets, meeting business and user requirements. Among the standards published by ISO, the series ISO 2700x defined a set of standards for information security management, while ISO 31000 established general principles for risk management.

3. LIST OF IDENTIFIED STANDARDS AND GUIDELINES Standards, guidelines and other publications relevant for energy companies are reported in the following table:

Table 4. Relevant documents

Institution Title Type ACC Guidance for Addressing

Cyber Security in the Chemical Industry Version 3.0

Guideline

AGA AGA Report No. 12, Cryptographic Protection of SCADA Communications

Guideline

AIRMIC, ALARM, IRM

A risk management standard

Standard

API Security Vulnerability Assessment Methodology for the Petroleum and Petrochemical Industries

Guideline

API API Std 1164 - SCADA Security, First Edition

Standard

API Security Guidelines for the Petroleum Industry

Guideline

Austrian Federal Chancellery

Austrian IT Security Handbook

Guideline

British Office of Government Commerce

CRAMM (CCTA Risk Analysis and Management Method

Frame-work

BSI IT-Grundschutz Guideline CIGRE Management of

Information Security for an Electric Power Utility - On Security Domains and Use of ISO/IEC17799 Standard

Guideline

CLUSIF Méthode Harmonisée d’Analyse de Risques Informatiques (MEHARI)

Frame-work

COSO COSO - Enterprise Risk Management - Integrated

Frame-work

Framework CPNI Good Practice Guide -

Process Control and SCADA Security, Overview, Parts 1 to 6

Guideline

DOE Roadmap to Secure Control Systems in the Energy Sector

Guideline

DOE Risk Management Guide Guideline DOE / ES-ISAC

DOE Vulnerability and Risk-Assessment Methodology

Guideline

DOE / ES-ISAC

Energy Infrastructure Risk Management Checklists for Small and Medium Sized Energy Facilities

Guideline

DHS Catalogue of Control Systems Security: Recommendations for Standards Developers

Guideline

DHS Risk Analysis and Management for Critical Assets Protection

Guideline

Dutch Government

DHM Security Management

Guideline

Dutch Ministry of the Interior and Kingdom Relations

NRB Guideline

EURAM European Risk Assessment Methodology

Guideline

EURELECTRIC

Operational Risk Methodology - Scenario Analysis

Frame-work

French Government

EBIOS (Expression des Besoins et Identification des Objectifs de Sécurité)

Guideline

HM Treasury

The Orange Book - Management of Risk - Principles and concepts

Frame-work

IEC IEC 62531 Data and Communication Security

Standard

IEC IEC 61508 Functional safety of electrical/electronic/ programmable electronic safety related systems

Standard

IEC IEC 62210 - Power system control and associated communications - Data and communication security

Standard

IEEE IEEE Guide for Electric Power Substation Physical and Electronic Security

Standard

IRGC Risk governance - Toward an integrative

Frame-work

79


ISSN 2067-5534 © 2012 JSE

approach ISA ISA TR99.03.01 -

Security Technologies for Industrial Automation and Control Systems

Standard

ISA ISA 99.02.01 - Establishing an IACS Security Program

Standard

ISACA The Risk IT Framework Frame-work

ISO ISO 31000 - Risk management – Principles and guidelines

Standard

ISO ISO/IEC 27005:2008 Information technology - Security techniques - Information security risk management

Standard

ISO ISO 15408 Common criteria for information technology security evaluation

Standard

ISO ISO/IEC 13335-1:2004 Information technology – Guidelines for the Management of IT security

Standard

ITSEAG Generic SCADA Risk Management Framework

Frame-work

NAVI Good Practices for Risk Analysis

Guideline

NERC CIP 001-009 - Reliability Standards for the Bulk Electric Systems in North America

Regulation

NERC Security Guidelines for the Electricity Sector

Guideline

NIST NIST 800-30 Risk Management Guide for Information Technology Systems

Guideline

NIST NIST 800-82 Guide to Industrial Control Systems (ICS) Security

Guideline

NIST System Protection Profile - Industrial Control Systems

Specifica-tion

ENTSO-E Octavio Control Center Security

Workshop

SEMA Guide to Increased Security in Process Control Systems for Critical Societal Functions Guideline

Standards

Australia/ Standards New Zealand Commit-tee

AS/NZS 4360:2004: Risk Management

Standard

4. RISK RELATED CRITERIA Standards and guidelines were categorized by considering: Process coverage: Evaluation of whether or not the document covers one or more of these three main processes: Risk Management; Audit; Maintenance & Alignment. Risk management is further divided in sub-phases: Risk assessment: It includes risk analysis and risk

evaluation which be qualitative or quantitative; Risk reporting: Covered or not; Risk Monitoring & Control: Covered or not; Mitigations: Covered or not; just examples or more

detailed mitigation strategies. Risk Type: Type of risk addressed by the standard or guideline. E.g. general, IT-related, SCADA-related; Area coverage: Considers whether the document treats or at least gives examples of the following aspects: Threats: They may be classified in Categories;

Categories+ list; Natural hazard; Manmade /fraudulent; Cyber Attacks;

Vulnerabilities: They may be classified as: Personal; Information; ICT assets; Physical assets;

Controls: Mechanisms already in place to reduce the vulnerabilities.

5. EXAMPLES OF DETAILED DESCRIPTION OF A STANDARD / GUIDELINES Further on, we present some examples of detailed description form some of the standards analysed:

80


ISSN 2067-5534 © 2012 JSE

82


ISSN 2067-5534 © 2012 JSE

6. FINDINGS In total were identified 47 relevant standards, guidelines or frameworks in the area of energy companies risk assessment and management. The survey has included both more general publications and documents dealing with more specific aspects of risk management and security. Classification based on general criteria Geographic relevance: 15 standards / guidelines were developed by global organizations, 15 by US ones, 5 by EU ones, 3 by UK ones, 3 by Dutch, 2 by Australian, 2 by German and 2 by French. Even if some standards were created by national or regional institutions on the basis of their geographical specificities, they have a broader relevance and are applicable also to other countries. Industry / field addressed: 17 standards / guidelines are generic and so can be applied across industries and fields; 9 address information and communication technology; 11 are about energy; 5 are about critical infrastructure; 4 are about oil & gas; 1 is about the chemical industry. Classification based on risk related criteria Risk assessment: Virtually all the standards/guidelines address risk analysis and risk evaluation through qualitative methodologies. One of the few exceptions is represented by EURELECTRIC - “The Operational Risk Methodology - A Scenario Analysis” which provides a detailed methodology to estimate in a quantitative way the operational risk for the electricity industry. DOE - “Risk Management Guide” also covers some quantitative techniques for risk assessment, but does not go so much in depth in its analysis. Mitigations: 12 standards/guidelines give concrete examples of mitigations and suggest applicable mitigation tools and strategies. Some standards (e.g. ISA 99.02.01) propose an integrated mitigation system, others (e.g. IRGC - “Risk governance - Toward an integrative approach”) present a variety of mitigation techniques which may be suitable for dealing with different types of risks. In any case, the mitigation strategies illustrated in

these standards are not thought to be exhaustive; rather they aim at providing the organizations with some practical examples and suggestions in order to help them define the mitigation system which is more suitable for their specific characteristics and operational context. Audit: 5 standards/guidelines are suitable for auditing purposes. They are developed by institutions that deal specifically with audit such as COSO (Committee of Sponsoring Organizations of the Treadway Commission) and ISACA (Information System Audit and Control Association). These standards provide organizations with a detailed guidance to implement a risk management methodology and a security system which are perfectly compliant with the existing regulations. They also help the organizations clearly define responsibilities and the accountability for the different phases of the risk management process. Risk Type: 21 standards/guidelines cover general risks, while 26 address IT-related risks. Some standards (e.g. ISO 31000) consider generic risks across industries, while other standards (e.g. DOE - “Risk Management Guide”) look at generic risks for a specific sector (e.g. energy). EURELECTRIC - “The Operational Risk Methodology – A Scenario Analysis” estimates operational risk (which may come in a variety of forms) for the electricity industry. Regarding IT-related risks, some standards (e.g. ISO/IEC 27005:2008) focus on IT risk in general, while others (e.g. AGA - “Cryptographic Protection of SCADA Communications”) deal specifically with SCADA security. Threats: 26 standards/guidelines provide some kind of categorization, description and practical examples of threats. They present a different level of detail: some standards just identify some general typologies of threats, while some others (e.g. ISA TR99.03.01) describe more in depth a specific type of threat (e.g. cyber threats). Vulnerabilities: 14 standards/guidelines provide some kind of categorization, description and practical examples of vulnerabilities. Some standards present cases of vulnerabilities just to better illustrate their general framework using some practical examples. For instance,

83


ISSN 2067-5534 © 2012 JSE

ISO/IEC 27005:2008 develops a general methodology widely applicable across industries and then, in one its annexes, it gives many examples of vulnerabilities together with the threats which may exploit them. Other guidelines (e.g. DOE - “Vulnerability and Risk Assessment Methodology”) follow a bottom-up approach and start from a list of possible vulnerabilities to create a risk assessment framework. Controls: 15 standards/guidelines examine the control mechanisms which organizations may have in place to reduce the likelihood of a negative event and its possible impact. DHS - “Catalog of Control Systems Security” gives the most detailed list of recommended security controls that organizations from various industries may have implemented to counteract various types of threats. 7. CONCLUSIONS Some of the analyzed standards and guidelines have a generalist approach and provide risk management methodologies applicable to virtually every kind of industry (e.g. HM Treasury - “The Orange Book-Management of Risk - Principles and concepts”; ISO 31000. They could be used in the energy sector, but they would not be able to fully capture its peculiarities. In developing a specific methodology for energy companies, they might be utilized as a reference to make sure to proceed in a rational and systematic way. Some standards and guidelines (e.g. DOE - “Risk Management Guide”; NERC - “Reliability Standards for the Bulk Electric Systems in North America”; NERC -“Security Guidelines for the Electricity Sector”) offer a guidance to deal with risk in the energy sector, covering various aspects and types of risk. However, they lack a clear focus on energy companies to be considered as the ideal methodologies for performing energy critical infrastructure risk assessment and management. Moreover, most of them focus on the US context rather than on the European one. The key finding of this paper, resulting from the review of the existing standards and guidelines, is that, at the moment, there is not a comprehensive methodology for risk assessment and management for energy companies. There are general frameworks which are not able to fully address the needs of the Transmission System Operators and to capture all the specificities of energy critical infrastructures. There are more technical standards that lack the broader perspective which is required to perform an adequate risk assessment and management for energy companies. CN Transelectrica is participating in many international projects, financed under EU Seventh Framework Programme for Research and Technological development FP7: Energy Control Center Risk Assessment and

Mitigation Methodology ECCRAMM, Securing the European Electricity Supply Against Malicious and Accidental Threats SESAME, Critical Response in Security and Safety Emergencies CRISYS, European Risk Assessment and Contingency Planning Methodologies for Interconnected Energy Networks EURACOM. Reviewing the most relevant standards, guidelines, frameworks and methodologies about risk assessment and management, is a common baseline in developing, in international projects, more integrated and comprehensive risk assessment methodologies for energy critical infrastructures, useful to enhance energy critical infrastructure companies risk assessment practices. REFERENCES [1]. EURACOM, European Risk Assessment and Contingency

Planning Methodologies for Interconnected Energy Networks, Common areas of Risk Assessment Methodologies, FP7, 2010;

[2]. ECCRAMM, Energy Control Center Risk Assessment and Mitigation Methodology, Symantec, Stefano BUSCHI, Booz&Co, 2011;

[3]. Energy Control Centers & SCADA Protection Against Emerging Information Security Threats, Symantec, Booz&Co, 2011;

[4]. Generating a European risk assessment methodology for critical infrastructures, EURAM, European Risk Assessment Methodology, 2008;

[5]. EURELECTRIC Working Group on Risk, Risk Management în the Electricity Industry – White Paper, 2007;

[6]. HM Treasury The Orange Book - Management of Risk - Principles and concepts, 2004;

[7]. Australia/Standards AS/NZS 4360:2004: Risk Management New Zealand Committee, 2004;

[8]. A risk management standard, AIRMIC, ALARM, IRM, 2002;

[9]. CIGRE Management of Information Security for an Electric Power Utility - On Security Domains and Use of ISO/IEC17799 Standard, 2005;

[10]. COSO - Enterprise Risk Management - Integrated Framework, 2004;

[11]. DOE, Risk Management Guide, 2009; [12]. DOE/ES-ISAC, DOE Vulnerability and Risk-Assessment

Methodology, 2002; [13]. DOE/ES-ISAC, Energy Infrastructure Risk Management

Checklists for Small and Medium Sized Energy Facilities, 2002;

[14]. IRGC, Risk governance - Toward an integrative approach, 2006;

[15]. ISO ISO 31000 - Risk management – Principles and guidelines, 2009;

[16]. ISO ISO/IEC 27005:2008 Information technology - Security techniques - Information security risk management.

84


ISSN 2067-5534 © 2012 JSE

ARTIFICIAL INTELLIGENCE TECHNIQUES FOR DIAGNOSING POWER TRANSFORMERS’ FAULTS

CIULAVU C.A

University Transilvania of Brasov, Brasov, [email protected]

Abstract – The paper is structured in five parts. The first part includes the importance of the power transformers’ in the power system. The second part presents the type of incipient faults which may decrease the electrical and mechanical integrity of the insulation system. The third part presents the methods for detecting the incipient faults of power transformers. In the fourth part it is introduced a diagnosis program which identifies different types of faults using artificial intelligence and dissolved gas analysis. Keywords: diagnosis, dissolved gas analysis, fault, artificial intelligence. 1. INTRODUCTION

Power transformers are major power system equipment. The major concern of power transformers’ incipient faults is that they may decrease the electrical and mechanical integrity of the insulation system. Transformer faults can have a significant economic impact due to long lead times in procurement, manufacturing and installation in addition to high equipment cost. Extending the useful life of the power transformers is the single most important strategy for increasing life of power transmission and distribution infrastructures.

The fault of a power transformer may leave thousands of homes without heat and light, and the fault of a step-up transformer in a power generation plant may cause the shutdown of the attached generation unit.

Power transformers are very expensive devices and therefore monitoring and diagnosing systems will be valuable for preventing damage to the transformers.

It is important to know about transformer failures because: Largest financial losses due to failures of power

transformer; The main responsible for failure of power transformer

is insulation; The cost of insulation failures alone accounts for more

than half of all failure costs. The purpose of this article is to develop a diagnosis system using artificial intelligence techniques.

2. TYPES OF FAULTS The faults that occur within the transformer

protection zone are internal faults. Transformer internal faults can be divided into two

classifications: internal short circuit faults and internal incipient faults. Internal short circuits faults are generally turn-to-turn short circuits or turn to earth short circuits in transformer windings.

Internal incipient transformer faults usually develop slowly, often in the form of a gradual deterioration of the insulation due to some causes.

Statistics show that winding failures most frequently cause transformer faults (ANSI/IEEE 1985). Insulation deterioration, often the result of moisture, overheating, vibration, voltage surges, mechanical stress created during transformer through faults, are major reason for winding failure.

Voltage regulating load tap changers, when supplied, rank as the second most likely cause of a transformer fault. Tap changer failures can be caused by a malfunction of the mechanical switching mechanism, high resistance load contacts, insulation tracking, overheating, or contamination of the insulation oil.

Transformer bushings are the third most likely cause of failure. General aging, contamination, cracking, internal moisture and loss of oil can all cause a bushing to fail.

Two other possible reasons are vandalism and animals that externally flash over the bushing.

Transformer core problems have been attributed to core insulation failure, an open ground strap, or shorted laminators.

Other miscellaneous failures have been caused by current transformers, oil leakage due to inadequate tank welds, oil contamination from metal particles, overloads and over voltage.

The factors responsible for failures and accelerated deterioration can be categorized as: Operating environment (electrical): load current, short

circuits, lightening and switching surges; Operating environment (physical): temperature, wind,

rain, pollution; Operating time: time in service and time under

abnormal conditions; Number of operations of tapchanger; Vibration effect: sound and material fatigue; Contaminants: moisture, presence of oxygen and

particles in oil. A correlation between the causes and the effects

85


ISSN 2067-5534 © 2012 JSE

produced at the flaw is presented in Table 1 [2], [4-16].

Table 1. Correlation between power transformer internal faults and causes

Faults Causes Arcing Corona Overheating

of cellulose Overheating of oil

Winding turn-to-turn short-circuit

X X

Winding open circuit

X X

Operation of build-in LTC

X

Winding distortion or displacement

X X

Lead distortion or displacement

X X

Loose connection to bushing terminals, tap leads, terminal boards

X X X

Free water or excessive moisture in oil

X X

Floating metal particles

X X

Loose connection to corona shields

X

Loose collars, spacers, core ground straps, core hold down angle (Braces)

X

Through fault X

Overloading X X

Damaged yoke bolt insulation

X

Rust or other damage on core

X

Damaged shunt packs of tank

X

Jammed oil circulating path

X

Cooling system malfunction

X

Usually, one fault type may have more than one

cause. Example: arching and/or overheating of solid insulation may have as cause winding turn-to-turn short-circuit; arching and corona discharges may have as cause free water or excessive moisture in oil, etc. This makes fault location very difficult.

Nevertheless, fault diagnosis is good enough to provide information to a maintenance program, and serve as the basis of a preventive maintenance strategy. 3. METHODOLOGY OF INCIPIENT FAULT DIAGNOSIS

Dissolved gas analysis has become a very popular

technique for monitoring the overall health of a transformer. As various faults develop, it is known that different gases are generated.

By taking samples of the mineral oil inside a transformer, one can determine what gases are present and their concentration levels.

Researches have been done to connect theoretically the gaseous hydrocarbon formation mechanism with the thermodynamic equilibrium.

Some studies indicated that the hydrocarbon gases with the fastest rate of evolution would be methane, ethane, ethylene and acetylene.

Some studies have focused on key gases and what faults they can identify.

In Table 2 [2,3,4,6] the relationship between fault types and the key gases is shown. In the case of key gas analysis, a fault condition is indicated when there is excessive generation of any of these gases.

For this to be effective, much expert experience is still needed.

Table 2. The relationship between fault types and key gases

Key gas Chemical symbol

Fault type

Hydrogen H2 Corona Carbon monoxide and carbon dioxide

CO CO2

Cellulose insulation Breakdown

Methane and ethane

CH4 C2H6

Low temperature Oil Breakdown

Acetylene C2H2 Arcing Ethylene C2H4 High temperature oil

breakdown

For example, acetylene concentrations that exceed

the ethylene concentrations indicate that extensive arcing is occurring in the transformer, since arcing produces acetylene.

In addition to gas in the oil, it is an accepted fact that the presence of water is not healthy for power transformers. Water in the oil indicates paper aging, since the cellulose insulation used in power transformers is known to produce water when it degrades.

Water and oxygen in the mineral oil further increases the rate at which the insulation will degrade. This means that a high concentration of water in the oil not only indicates that the insulation has been degrading but it will degrade more quickly in the future due to increased presence of water in the oil.

Water in the oil is also a sign that the mineral oil itself is deteriorating.

When the mineral oil deteriorates, the dielectric constant of the oil decreases.

The key gas method identifies the key gas for each type of fault and uses the percent of this gas to diagnose the fault. It interprets dissolved gas analysis results based on a simple set of facts.

In Table 3 is summarized the diagnostic criteria of the key gas method.

Table 3. Diagnostic criteria of key gas method

Fault Key gas Criteria Arcing Acetylene (C2H2) Large amount of H2 and

C2H2 and minor quantities of CH4 and C2H4. CO and CO2 may also exist if cellulose is involved.

Corona (PD) Hydrogen (H2) Large amount of H2, some CH4, with small quantities of C2H6 and C2H4. CO and CO2 may be comparable if cellulose is involved.

Overheating of oil

Ethylene (C2H4) Large amount of C2H4, less amount of C2H6, some quantities of CH4 and H2. Traces of CO.

Overheating of cellulose

Carbon monoxide (CO) Large amount of CO and CO2. Hydrocarbon gases may exist.

In conclusion, the transformer oil analysis can give

86


ISSN 2067-5534 © 2012 JSE

very important information about the state of the transformer and its electrical insulation, like: Internal fault: there can be detected electrical or

thermal faults and their type (partial discharge, hot spots, arcing);

Transformer aging: the actual ageing of the transformer is given by the state of the winding insulation;

Quality of the oil: it must be verified if it achieves the insulating and cooling functions.

It can be added that the advantages of the oil analyses: The cost is very small in comparison wih the cost of a

transformer failure; No need to interrupt the production process.

Though moisture and dissolved gas analysis are helpful in detecting many types of failures that can occur in a transformer, the measurement of partial discharges is the most effective method to detect pending failure in the electrical system.

As the electrical insulation in a transformer begins to degrade and breakdown, there are localized discharges within the electrical insulation. Every discharge deteriorates the insulation material by the impact of high-energy electrons, thus causing chemical reactions.

Partial discharges may occur only right before failure but may also be present for years before any type of failure. A high occurrence of partial discharges can indicate voids, cracking, contamination or abnormal electrical stress in the insulation [2], [4-16].

The most common method for on-line detection of partial discharges is the use of acoustical sensors mounted external to the transformer.

The main difficulty with using acoustical sensors in the field, however, is in distinguishing between internal transformer partial discharges and external partial discharges sources, such as discharges from surrounding power equipment.

An alternative method has been proposed recently to differentiate between internal and external partial discharges and is based on the combined use of signals from a capacitive tap and signals from an inductive coil fitted around the base of the bushing.

The advantage of partial discharges sensors is the ability to detect the actual location of insulation deterioration, unlike the dissolved gas sensors. The one disadvantage to partial discharge sensors is that they are greatly affected by the electromagnetic interference in the substation environment.

One of the simplest and most effective ways to monitor a transformer externally is through temperature sensors. Abnormal temperature readings almost always indicate some type of failure in a transformer.

It is known that as a transformer begins to heat up, the winding insulation begins to deteriorate and the dielectric constant of the mineral oil begins to degrade.

In order to make on-line monitoring possible, thermocouples are placed externally on the transformer and provide real-time data on the temperature at various locations on the transformer. In many applications, temperature sensors have been placed externally on transformers in order to estimate the internal state of the transformer.

Though the breakdown of the insulation can cause

catastrophic failure in a transformer, the life of a transformer is predominantly shortened by the deterioration of its accessories. These accessories include the bushings, load tap changers and cooling system.

Some of the causes of bushing failures include changing dielectric properties with age, oil leaks, design or manufacturing flaws, or the presence of moisture. Sensors have now been created to monitor the health of bushings. Transformer bushings have a finite life.

Overheated load tap changers can result from many different phenomena. These causes include coking, misalignment, and loss of spring pressure. Though the contact temperature cannot easily be measured directly, the overheating will generally result in an increase in the load tap changer oil temperature.

By monitoring the load tap changer temperature closely, the flashover between the contacts can be avoided, which usually results in a short circuit of the regulating winding and subsequent failure of the transformer [2].

Vibration analysis by itself cannot predict many faults associated with transformers, but it is another useful tool to help determine transformer condition. Vibration can result from loose transformer core segments, loose windings, shield problems, loose parts or bad bearings on oil cooling pumps or fans.

Every transformer is different, therefore, to detect this, baseline vibration tests should be run and data recorded for comparison with future tests.

Vibration analyzers are used to detect and measure the vibration. Information gained from these tests supplements ultrasonic and sonic fault detection tests and dissolved gas analysis.

Information from these tests may indicate maintenance is needed on pumps/fans mounted external to the tank. It may also show when an internal transformer inspection is necessary.

If wedging has been displaced due to paper deterioration or through faults, vibration will increase markedly. 4. ARTIFICIAL INTELLIGENCE TECHNIQUES

The most achievements of diagnosis systems are

incorporated using expert systems, in which the methods based on analytical models are combined with those based on qualitative models. This reasoning stays at the base of drawing up the architecture of such a system, which combines the heuristic knowledge obtained through experience with the qualitative and quantitative knowledge about the model.

The ordering and diagnosis unit constitutes the central part of the system, which incorporates both the method based on analytical model with the one that uses heuristic information, and the method based on knowledge about the system. The analytical and heuristic knowledge from the knowledge base results after a generating process of symptoms. The analytical knowledge is used to obtain analytical information, quantized through the measured variables processing of the process and a set of characteristic values is obtained which allow: the control of the admissible bound values of the

87


ISSN 2067-5534 © 2012 JSE

directly measured signals; the directly measured signals analysis through the

utilization of analysis methods; the process analysis through the utilization of the

mathematical models of the process with the estimation methods of parameters, state and the method of parity equations. Beside the symptoms with quantized information, the heuristic symptoms can be obtained through the utilization of the information supplied by the human operators:

based on the observations and technical inspections (reviews), the values of the heuristic characteristics are obtained (noise, color, smell, vibration, wear);

a set of statistical data obtained on the basis of experience in process use or another similar processes and the operations of maintenance, repair, life time. In the case of complex systems, with the help of

heuristic knowledge stated as heuristic models (qualitative models), the causality fault-symptom can be established and it can be done a balance of different diagnosis strategy. In this way, fault trees or IF – THEN reasoning can be used, so the abnormal working and failures, that appear in a system, can be determined.

Figure 1 represents a block diagram showing the steps that are taken for the knowledge-based.

Fig.1. Approach to fault diagnosis

In the beginning, most attempts in transformer

diagnostic focused on analytic models. Analytic models attempt to represent the system through mathematical equations. Depending on the complexity of the system and the desired model accuracy, both linear and nonlinear models have been developed. These models attempt to use physical principals to model a system.

For transformers, many different types of models have been developed to try to identify the system and detect failures. The transformer system is very complex. It contains thermal, mechanical, electrical and fluid systems. For protection against overloading, transformer thermal models have also been developed, which use two exponential equations and non-linear time constants determined from transformer data. These models are used to predict top-oil rise over ambient temperature, hottest-

spot conductor over top-oil temperature, and hottest-spot winding temperature given a specific load. Many models have been formed that combine temperature measurements with current, voltage and other transformer measurements. State estimation methods were used to provide accurate estimates using oil, tank and ambient temperatures in addition to the voltages and currents on all the phases. In addition to the thermal models, many mechanical, electrical and even fluid models have been developed for the transformer. On the mechanical side, a mathematical model has been derived to express the mechanical stresses due to forces on the transformer windings. This model provides critical information on the possible damage that is caused from radial short circuit forces and gives an assessment of the possibility that a catastrophic fault from a winding short circuit could occur. Likewise, diagnostics of the electrical system have been developed using the transfer function method. Though the method uses ratios of the transformers electrical voltages and currents, the method actually detects defects in the mechanical system. The transfer function method is a quotient of the Fourier transformed input and output signals. These quotients are used to model the system electrically, and through comparison with previous fingerprints, can detect developing defects. The artificial intelligence trains itself to the system and provides diagnostic information based on a set of inputs and outputs. The actual mapping that the artificial intelligence develops or how this relationship relates to any physical principles is usually not defined. The most common forms of artificial intelligence used for transformer diagnosis are neural networks and fuzzy logic. Due to the complexity of the numerous phenomena, it is difficult to formulate a precise relationship relating the different contributing factors. This uncertainty naturally lends itself to fuzzy set theory. It has been developed a transformer diagnostic system that utilized both an expert system and a neural network to detect failures in a transformer. The knowledge of the expert system has many uncertainties, and therefore fuzzy logic is employed. The two techniques are integrated by comparing the expert system conclusion with the neural network reasoning using a consultative mechanism. A block diagram for this type of hybrid system is presented below. There was also developed a comprehensive system that included fuzzy logic, expert system and an artificial neural network to detect faults in the insulation system. In this case, fuzzy logic is implemented in coordination with the neural network. The outputs of the neural network are numerical values between 0 and 1, which are placed in membership functions based on a set of fuzzy rules. The combination of an expert system with neuro-fuzzy techniques is not the only diagnostic tool used in transformer systems. An integration of an artificial neural network and an expert system has been developed for power equipment diagnosis. The system uses the neural network to form implicit diagnostic rules and has the added benefit of logic regression analysis for fault location.

Knowledge based diagnosis

Human Knowledge rule base for sensor outputs

Artificial intelligence

Expert system Fault detection

88


ISSN 2067-5534 © 2012 JSE

Fig. 2. Strategy for combined fuzzy logic, expert system and neural network

Though many methods have employed some

combination of fuzzy logic, artificial neural networks and expert systems, this is not always the case. A highly accurate two-step artificial neural network has been used for transformer fault diagnosis using dissolved gas data.

Fuzzy logic has also been used to diagnose the health of a transformer and foresee any developing failures. Fuzzy logic has been used to smooth out some of the problems that can appear when using the cut and dry rules of expert system knowledge. By forming fuzzy membership functions for the different measurements (gases, generation rates, electric current, temperature), it is possible to overlap the individual membership functions into one large fuzzy matrix that can be used for diagnosis. A fuzzy logic diagnostic system has also been developed for the transformer that utilizes evolutionary programming and different shaped membership functions to get a more accurate fuzzy diagnostic system.

Another artificial intelligence based approach utilizes a genetic algorithm in coordination with an artificial neural network. One of the weaknesses of artificial neural network approach is the tendency to find only a local minimum in its training due to improper initial value. The genetic algorithm is used to optimize the initial value and thus increase the accuracy of the neural network training. Likewise, genetic algorithms have been used in the training of a fuzzy controller that forms diagnostic rules based on dissolved gas data. In this case, the fuzziness helps define diagnostic operating conditions and the genetic algorithm decreases the amount of rules needed.

For a diagnosis system of transformers, it is needed to obtain some rules. For guidance there can be use IEC or IEEE. These are revised from time to time and need a correct interpretation of the rules. Figure 3 represents a part of a power transformer diagnosis engine using the IEEE conditions and ratio method. The rules are represented in IF-THEN structure.

Fig. 3. Power transformer fault diagnosis engine

The ratio CO/CO2 is used as a main factor to diagnose the degradation of cellulose. 5. CONCLUSION

To manage the life of transformers, to reduce failures and to extend the life of the transformer, some tests must be taken. The tests are carried out to prove that the transformers are ready to operate or to find the faults.

Equipment failures do occur even with the best equipment designs available and using the best utility practices. In order to operate a power system reliably, transformer failures must be anticipated. Dissolved gas analysis is very important to determine the condition of a transformer, it can identify a problem such as: deteriorating insulation oil, overheating, partial discharge and arcing.

The transformers have different gassing characteristics because of their size, structure, manufacture, loading and maintenance history. The new diagnosis systems have expert systems incorporated, in which the methods based on analytical models are combined with those based on qualitative models.

Expert systems and artificial intelligence techniques have already been proposed to understand the obvious and non-obvious relationships between transformer failures and the causes of failures. Preliminary results, obtained from the application of these techniques, are encouraging, however some limitations exist. Knowledge acquisition, knowledge representation and maintenance of a great number of rules in the expert systems require plenty of efforts. Artificial intelligence techniques were studied. These techniques include expert system, fuzzy logic, evolutionary algorithm and artificial neural network. The effectiveness of the expert system and fuzzy logic depends on the precision and completeness of human knowledge accumulated over the years. Both methods

89


ISSN 2067-5534 © 2012 JSE

need a large knowledge base that must be constructed manually and cannot adjust their diagnostic rules automatically thus cannot acquire knowledge from the new data samples through a self-learning process. The quality of the data is especially important as poor data can lead to a poor diagnosis or misdiagnosis. REFERENCES [1]. Van Harmelen, F., Ten Teije, A. - Approximations in

diagnosis: motivations and techniques, SARA, 1995. [2]. *** Technical manual – Power Transformer maintenance

and acceptance testing, 1998. [3]. Ciulavu, C. - Diagnoza echipamentelor electrice din

centrale si statii de inalta si medie tensiune, Referat doctorat, Brasov, 2007.

[4]. Lance R. Lewand - Condition Assessment of Transformers – Analysis of oil Data and its Quality.

[5]. Alff, J.J., Der Houhanessian, V., Zaengl, W.S., Kachler A. - A Novel Compact Instrument for Measurement and Evaluation of Polarization /Depolarization Currents Conceived for On Site Diagnosis of Electric Power Apparatus. In: IEEE Intern. Sympos. Electr. Insul. USA, pp. 161-167, 2000.

[6]. Saha, T. K. - Review of Modern Diagnostic Techniques for Assessing Insulation Condition in Aged Transformers. In: IEEE Transactions on Electrical Insulation, vol 10, no. 5, pp. 903- 916, 2003.

[7]. Gubanski, S. M., Boss, P., Csepes, G., Der Houhanessian,V., Filippini, J., W. Guuinic, P., Gafvert, P., Karius, V. Lapworth, J., Urbani, G. Wertelius, P. Zaengl. - Dielectric Response Methods or Diagnostics of Power Transformers. In: IEEE Electrical Insulation Magazine, vol. 19, no. 3, pag. 12-17, 2003.

[8]. Duval, M., Dukarm, J. - Improving the Reability of Transformer Gas – in Oil Diagnosis. In: IEEE Electrical Insulation Magazine, vol. 21, no. 4, pp. 21-27, 2005.

[9]. Baird, P.J., Herman H., Stevens, G .C. - Spectroscopic

Measurement and Analysis of Water and Oil in Transformer Insulating Paper. In: IEEE Transactions on Dielectrics and Electrical Insulation vol. 13, no.1, pp.293-308, 2006

[10]. Oomen, T.V., Prevost, Th.A. - Cellulose Insulation in Oil-Filled Pover Transformers: Part II Maintaining Insulation Integrity and Life. In: IEEE Electrical Insulation Magazine, vol. 22, no. 2, pp. 5-14, 2006

[11]. Helerea E., Munteanu A., Lungoci S. - Aging of Electroinsulating Materials. Part I: Life Modeling Theories, In: Proceedings of International Conference on Materials Science & Engineering BRAMAT 2003, Braşov, pp.203-208, 2003

[12]. Helerea E., Sângeorzan L. - Aging Curves of Transformer Insulation, In: Proceedings of the 5th International Conference on Optimization of Electric and Electronic Equipment, Optim'96, Braşov, 1-8, 1996

[13]. Leondes, Cornelius T. - Knowledge – Based systems: Techniques and Applications, ISBN – 13: 9780124438750, Academic Press, 2000, Vol. 4 – Cap. 33 – Expert systems in power system control.

[14]. Sinha, Naresh K., Gupta, Madan N., Zadeh, Lotti A. - Soft computing and intelligent systems: theory and applications, Academic Press, 2000, Cap 13 – Expert systems in process diagnosis and control.

[15]. Bayliss, C.R., Hardy, B. J. - Transmission and distribution electrical engineering, Newnes, 2007, Cap. 5 – Current and voltage transformers.

[16]. Laughton, M.A., Warne, D.F. - Electrical engineer’s reference book, Elsevier Newnes, 2002, Cap. 33 – Power transformers.

[17]. Mobley, R. Keith - Plant engineer’s handbook, Elsevier Butterworth – Heinemann, 2001, Cap. 44 – Vibration monitoring and analysis.

[18]. Hopgood, A. - Intelligent Systems for Engineers and Scientists, CRC Press LLC, USA, 2001.

[19]. Liebowitz, J. - The Handbook of Applied Expert Systems, CRC Press LLC, ISBN – 0849331064.

90


ISSN 2067-5534 © 2012 JSE

RECENT DEVELOPMENTS IN SITE INVESTIGATION, DESIGN AND APPLICATION FOR GROUND SOURCE HEAT PUMPS

SANNER B., MANDS D., SAUER M., GRUNDMANN, E. UBeG GbR, Reinbergstrasse 2, 35580 Wetzlar, Germany

[email protected]

Abstract – Ground Source Heat Pumps (GSHP) are a proven technology to harness shallow geothermal energy. Nevertheless, the market introduction of GSHP is rather different among the European countries, as is the industry infrastructure, knowledge and installation skills. Standardisation and clear strategies and rules for exploration, design and installation are crucial to deploy GSHP in countries with an embryonic market. The paper presents recent developments in GSHP technology in several fields, in particular: - Site investigation technology TRT - Design software and its validation - Application examples with new approaches The constraints controlling GSHP design for a certain building and site are discussed and the role of the developments presented is explained. Keywords: geothermal, heat pumps, site investigation, design software

1. INTRODUCTION

First Ground Source Heat Pumps (GSHP) are reported from USA in 1945, in Europe the technology dates back to the 1960s. An overview over the historical development is given in [1]. GSHP thus could achieve meanwhile a high degree of evolution and refinement, with proven concepts and reliable components. In 2010, the number of heat pumps operating within the EU exceeded 1 mio. units (figure 1). As this does not include Switzerland and Norway, two countries with a substantial number of installations, the total for all Europe might today be close to 1.5 mio. units.

The main technologies used today for coupling the heat pump to the ground are: - Borehole heat exchangers (“vertical loops”) - Horizontal heat exchangers (“ground loops”) - Compact heat exchangers (slinky, spiral, cages…) - Energy piles, foundation walls, etc. - Groundwater wells

0

2000

4000

6000

8000

10000

12000

14000

2003 2004 2005 2006 2007 2008 2009 2010

installed heating capacity (M

Wth)

0

200.000

400.000

600.000

800.000

1.000.000

1.200.000

1.400.000Number of units in operation

MW ins ta l led

tota l number of GSHP

Fig. 1: Number of GSHP units and their installed heating capacity inside EU (after data from [2]) The market in some of the countries is reaching a

mature or even saturated phase, as in Sweden and Switzerland. On the other hand, there is a huge market potential in several other countries. A big obstacle in these countries is the lack of awareness in the public (and with the authorities), and the lack of knowledge and specific skills with the installers. Training and education schemes must help in overcoming this obstacle, like the promising approach given in the Geotrainet program [3]. Material and curricula elaborated in the Geotrainet project are available at www.geotrainet.eu.

2. SITE INVESTIGATION / TRT

The Thermal Response Test (TRT) is a tool to investigate ground thermal parameters required for design of borehole heat exchangers (BHE), as used in GSHP. While the theory and the use of the basic principles date back to the 1970s, the first mobile application of TRT is reported in 1995 [4], and the first TRT in Germany was done in 1999 [5]. Since then, a wealth of practical experience could be sampled both in the practical setup of the test (accuracy, reliability, site

91


ISSN 2067-5534 © 2012 JSE

accessibility, insulation, etc.) as in the understanding of the different evaluation methods. A global review was done in 2005 [6].

The measurement signal of a TRT is a temperature change due to heat injection or extraction. Concerning the basic physics, it does not matter if heat is injected or extracted. The parameter to be determined, thermal conductivity is not dependent from the direction of heat flow. The most important impact is given by the uniformity of the thermal load. This must be as constant as possible, in order to achieve an undisturbed signal.

Practical experiences on test operation has been gained during more than a dozen years of TRT tests throughout many European countries. One result is the importance of having a verification of the final results by using methods of sequential (step-wise) evaluation. Through this method a sufficient length of test time and the prevalence of conductive heat transport can be checked. Beside this verification, an awareness of the overall accuracy of the results as depending on accuracy of data collection is required. For the validity of temperature logs, the bottom heat dissipation can be used as an indicator.

A lot of additional information can be obtained from the test data, in particular if a temperature log inside the borehole is combined with the test. Examples of such information comprise layers with groundwater flow, distinction of layers of different thermal conductivity, quality of grouting, geothermal gradient, etc. It is also possible to use the TRT for investigating the actual depth of the borehole heat exchanger by use of the Thermo-Impulse Method [7].

a) Test procedures and reliability check Based upon experience, some mandatory routine

procedures are suggested to be performed before the start of the response test, in order to avoid unpleasant incidents:

Power supply check. The test can of course not be performed without electric power, be it from the grid or from a generator. Considering the required power levels, typically 3-phase AC is the source. Wrong phasing of this power supply can result in shunt fault, controller failure, overheating of the device and even smouldering of the test rig. Power breakdown or instable power supply may lead to inconsistent development of the temperatures, and thus makes it difficult or impossible to evaluate the test.

Sufficient de-aeration. Without proper de-aeration, the flow inside the borehole can collapse after an unknown amount of time, and the test will come to an unexpected early end.

Insulation of the test rig and connections. The ambient influence (heat or cold) should be kept as low as possible, as it cannot be controlled and heavily affects the test in a similar way as fluctuating power supply.

The so-called “Stepwise Evaluation” (sequential data analysis) allows for cross-checking if any of the effects mentioned above have had an influence on the test operation (fig. 2).

p

.12.08.06:00

12.08.00:00

11.08.18:00

11.08 .12:00

11.08.06:00

11.08.00:00

10.08.18:00

10.08.12:00

10.08.06:00

10 .08.00 :00

09 .08.18 :00

09.08.12:00

09.08 .06:00

09.08.00:00

08.08.18:00

08.08.12:00

08.08.06:00

08.08.00:00

Lam

bda

[W /

(m x

K)]

3

2

1

0

p

.18.12.06:00

18.12.00:00

17.12.18:00

17.12.12:00

17.12.06:00

17.12.00:00

16.12.18:00

16.12.12:00

16.12.06:00

16.12.00:00

15.12.18:00

15.12.12:00

15.12.06:00

15.12.00:00

Lam

bda

[W /

(m x

K)]

5,0

4,5

4,0

3,5

3,0

2,5

2,0

1,5

1,0

0,5

0,0

.04.06.06:00

04.06.00:00

03.06.18:00

03.06.12:00

03.06.06:00

03.06.00:00

02.06.18:00

02.06.12:00

02.06.06:00

02.06.00:00

Lam

bda

[W /

(m x

K)]

5,0

4,5

4,0

3,5

3,0

2,5

2,0

1,5

1,0

0,5

0,0

Fig. 2. Examples of stepwise evaluation of TRT: Dominated by conductivity, good reliability (top), dominated by advection and not usable (middle), and high fluctuations and low reliability (bottom)

92


ISSN 2067-5534 © 2012 JSE

An evaluation of the recorded data is performed in a stepwise evaluation with a fixed start time and increasing length of the data set, until the full duration to the end time. The resulting thermal conductivity for each time-span can be calculated and plotted over time. Usually in the first part of such a curve the thermal conductivity swings up and down, converging to a steady value and a horizontal curve in the case of a perfect test (figure 2, top). This procedure is a useful tool to check the quality of the data collected and the validity of the results.

With substantial influence of flowing groundwater, the curve rises upwards steadily after some time (figure 2, middle). Thus the test result value () is determined by the duration of the test, and the longer the testing time is,

the higher the will be. There is no reliable result for such a test. In case of influence of fluctuating power supply or environmental influences (e.g. solar radiation), the test result is not stable, and testing time must be extended (figure 2, bottom).

Beside the thermal conductivity of the underground, the undisturbed ground temperature (as average over the BHE length) is of crucial importance when calculating a BHE-field design. This parameter can be drawn from the temperatures recorded with the TRT device before the heating phase started with just the circulation pump running (fig. 3). Another method is the temperature log (fig. 4).

Temperatures

.26.02.22:00

26.02.21:00

26.02.20:00

26.02.19:00

26.02.18:00

26.02.17:00

26.02.16:00

26.02.15:00

26.02.14:00

26.02.13:00

26.02.12:00

26.02.11:00

26.02.10:00

232221

201918

171615

141312

1110

average (over BHE length)undisturbed ground temperature

but

influenced by movement of water,heat input of circulation pumpand heat capacity of test device

Fig. 3. : Undisturbed ground temperature from TRT

120

100

80

60

40

20

0

10 12 14Temperature [°C]

Undisturbed groundtemperature =arithmetic ave. ofall values belowzone of seasonalfluctuation andbottom of BHE

Zone of seasonalfluctuation canbe left out forcalculation of average temperature

Fig. 4: Temperature profile before performing TRT

When running the circulation without heating,

however, due to the (very small) heat input from the circulation pump, a small increase of the value might occur over time. An observation of temperature

development without heating over some hours (as in fig. 3) also can help in detecting any residual heat from drilling or solidifying of the grout, given away by a temperature decreasing over time.

The temperature log (fig. 4) in addition allows for exclusion of the zone of annual variation and gives much more details that could be used for further information, There are several tools available for taking a log, even inside a 32-mm-pipe as used for most BHE. Logging requires some patience to allow for full thermal equilibrium at each depth level to be measured, and some handling skills not to get a tool stuck in a well or pipe.

b) Extended range of applications

The temperature log before the test as shown in fig. 4 should be complemented with temperature logs after the end of the TRT (a recommendation could be a log directly after, one about 1 hour later, and another one 2-3 hours after the end of the test). These logs will show the gradual cooling of the fluid inside the pipes and allows for various conclusions as shown in figures 5 and 6. It should be taken into account that the exact time of the temperature measurement is not the same over the depth of the BHE, as the logging takes some time (up to 30 minutes for 100 m). So the signals might by slightly different with depth.

Among the features visible (figures 5 and 6) are groundwater flow, missing grout (to cool down so quickly as in figure 6, the BHE must have a direct contact to flowing groundwater, i.e. not being encased by

93


ISSN 2067-5534 © 2012 JSE

the grouting), or layers with different conductivity. Sometimes it is not clear if the temperature sensor went all the way to the bottom of the BHE, or if the BHE is just blocked (e.g. by a pinch). The “Bottom Heat Dissipation” (figure 5 right) gives a confirmation for

having reached the bottom, as at this point the heat is also transported in vertical direction downwards, and a faster cooling can be seen.

9m

100m

Brown Coal

vor Testbeginn 1h nach Testende 2h nach Testende 3h nach Testende

11 12 13 14 15 16 17 18 19

Temperature [°C]

100

90

80

70

60

50

40

30

20

10

0

16 18 20 22 24

Temperature [°C] Fig. 5: Features as elucidated by temperature log before and after TRT:

presence of a low-conductivity layer (left), prove of final depth (right)

100

80

60

40

20

0

8 12 16 20

3,0 h nach Testende 2,0 h nach Testende 1,0 h nach Testende vor Testbeginn

Tie

fe [m

]

Temperatur [°C]

100

90

80

70

60

50

40

30

20

10

0

8 9 10 11 12

2,0 h after TRT 1,0 h after TRT 0,5 h after TRT before TRT

Temperature [°C]

Fig. 6: Features as elucidated by temperature log before and after TRT: Influence of groundwater-flow (left), poor or non-existent grouting (right)

94


ISSN 2067-5534 © 2012 JSE

c) Thermo-Impulse Method Using the Thermo-Impulse Method, a practical issue

can be solved in shallow geothermal installations. Sometimes disputes arise over the question if the BHE actually has the full length as contracted. The TRT rig can offer a convenient method of determining the actual BHE-depth within a narrow margin of error. The method was first published in Sauer et al. (2010). It comprises the following steps (fig. 7):

• A strong thermal signal (impulse) is injected into the BHE circuit

• The time the impulse needs to return is measured.

• With the (measured) flow rate and pulse-time-delay the volume of the BHE can be calculated.

• With the known diameter of the BHE tube and the volume the length can be calculated.

Tests with recurring Thermo-Impulse measurement at the same borehole heat exchanger confirmed the reproducibility of the depth measurement (table 1).

time [s]

658 s

1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

18,0

18,5

19,0

19,5

20,0

1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 7000 7500

19,0

20,0

21,0

22,0

23,0

24,0

25,0 forward flow return flow

Fig. 7: Principle of Thermo-Impulse method (recurrence of impulse)

Table 1: Reproducibility of Thermo-Impulse

measurement

Measurement Time delay to

recurrence Depth (m)

1st: 658 s 129.0

2nd: 662 s 130.2

3rd: 659 s 129.7

Average 129.6

Maximum deviation

±0.6 m

(±0.5 %)

d) Possible use of TRT for investigation of deep

geothermal potential As a side note, some thoughts are presented here on

measurements in TRT that could be of interest for deep geothermal projects. From temperature logs before TRT, the geothermal gradient (temperature increase with depth) and, with knowledge of the thermal conductivity as a result of the TRT, the geothermal heat flux can be determined as:

Qg = kg *

with: Qg = Geothermal heat flux (W/m2)

kg = geothermal gradient, in K/m

= thermal conductivity (W/m/K)

Estimates on the expected lithology under the site allow for extrapolation of these values down to the depth required for deep geothermal projects (heat and/or power). Naturally, such extrapolation will not sufficiently reflect deep groundwater movements and other factors contributing to geothermal anomalies, but it can be a first hint to the geothermal character of an area where no deep boreholes yet exist. 3. DESIGN SOFTWARE AND ITS VALIDATION

Design of a geothermal heat pump system requires provision of sufficient heat extraction capacity from the ground for heating, or heat injection capacity (for cooling). With groundwater wells, this will be the well yield, to be determined by classical hydrogeological

95


ISSN 2067-5534 © 2012 JSE

methods (well test / pumping test), and some calculation of the thermal influence zones.

For systems with borehole heat exchangers (BHE), the temperature development in the BHE in response to heat extraction or injection is the key issue. To calculate this response, the Earth Energy Designer (EED) is a typical software. Being around for quite some years [8], EED now is in version 3.16 from 2010, and can be considered one of the standard tools for design of BHE.

A monitoring project [9] provided an opportunity for validation of geothermal design tools with actual measured data. A large office building with GSHP and BHE in Langen, Germany, built in 2000 [10], was used for reference. For the use of EED, the measured heat loads had to be summarised into monthly values (figure 8). The values in table 2 and figure 8 are those actually extracted from or injected into the underground, not the loads on the building side.

Table 2: Measured ground-side heat loads in the Langen project

design 2008 2009 2010 2011

Heat extraction (heating, MWh/a)

658 575 533 594 469

Heat injection (cooling, MWh/a)

572 461 480 423 432

Ratio extract./inject.

1.15 (1 : 0.87)

1.25 (1 : 0.80)

1.11 (1 : 0.90)

1.40 (1 : 0.71)

1.09 (1 : 0.92)

-150

-100

-50

0

50

100

150

1.1.

08

1.4.

08

1.7.

08

1.10

.08

1.1.

09

1.4.

09

1.7.

09

1.10

.09

1.1.

10

1.4.

10

1.7.

10

1.10

.10

1.1.

11

1.4.

11

1.7.

11

1.10

.11

MW

h /

mo

nth

460.9 MWh/a for cooling


574.9 MWh/a for heating

533.2 MWh/a for heating

594.3 MWh/afor heating


gro

und

hea

t ext

ract

ion

gro

und

heat

inje

ctio

n

468.7 MWh/afor heating


Figure 8: Monthly heat extraction from the ground (for heating) and injection into the ground (for cooling) in Langen GSHP

EED is programmed for calculation of the same

heat/cold loads recurring every year. Using EED for calculating annually differing heat loads is only possible in plants with quasi-balanced energy flows at the ground side. In such cases, the surrounding ground temperature will be stable over the years. Long-term decreasing or increasing ground temperatures could not be addressed as input parameters within EED. For the ground thermal parameters of the Langen project, values from first Thermal Response Tests (TRT) in Germany in 1999-2000 could be used [5]. The undisturbed ground temperatures, however, under the greenfield in 1999 were about 1 K lower than those measured today in some observation wells outside the BHE field. This can be attributed to a general heating up of the underground from the buildings etc. over the past decade.

Using the measured temperature from the wells of 12.7 °C as the mean value over BHE depth, the

comparison of EED-calculation with the measured values as given in figure 9 and 10 can be drawn. The measured values are taken at two points, at the forward/return pipes from the mechanical room, and in a sensor chain inside one BHE in the field. For comparison with EED, the mean value between forward and return was used, and the sensor at 35 m depth (half of the BHE depth) in the field. The monthly averaged values from the BHE match well with the EED base load curve (which represents the monthly average as well). There is a deviation in summer 2008 and January-March 2009, which can be attributed to a substantial number of BHE isolated from the system in the search for a leakage. The percentage of active BHE was considered in the load input for EED, however, there might be some inaccuracy of representation of the actual situation. Since autumn 2009, the system is operating normally again, with just 2 BHE isolated permanently (i.e. 98.7 % of total BHE length available). Another

96


ISSN 2067-5534 © 2012 JSE

deviation is with the values at the building during summertime. While these values match well in autumn and winter, they are substantially higher in summer (and also higher than those measured at the BHE). This discrepancy still needs to be explained; most probable reasons comprise influences of ambient room temperature, from ground-side circulation pump, or from external sources (e.g. heat emissions of pumps etc. near sensors).

Beside the monthly averages shown in figure 9, EED allows also for calculating the maximum and minimum temperatures to be expected during full-load operation of the BHE system. However, this is not given as an actual

temperature, but as a kind of envelope within which the temperature will swing according to actual load patterns. The design just has to make sure that the extremes of this envelope are within allowed ranges for temperature both concerning the technical operation constraints as well as environmental issues in the underground. In figure 10 this min-max-envelope is shown for 2008-2011, for which consistent values for the hourly temperatures at the BHE in 35 m depth during the period May 2008 – October 2011 could be used for comparison. The prediction given by EED is rather well matching the actual temperature development.

-5

0

5

10

15

20

25

30

1.1.

08

1.4.

08

1.7.

08

1.10

.08

1.1.

09

1.4.

09

1.7.

09

1.10

.09

1.1.

10

1.4.

10

1.7.

10

1.10

.10

1.1.

11

1.4.

11

1.7.

11

1.10

.11

Tem

per

atu

re (

°C)

Ambient air mean value fluid at building

BHE in field, at 35 m depth mean value fluid calcul. EED

Figure 9: Measured temperatures in ambient air and in the Langen BHE (monthly averages), compared with EED-calculation of BHE

‐5

0

5

10

15

20

25

30

Temperature (°C)

BHE in field, at 35 m depth fluid peak cooling calc. EED

mean value fluid calcul. EED fluid peak heat calcul. EED

1.1.2008 1.1.2009 1.1.2010 1.1.2011 1.1.2012

Figure 10: EED-calculation showing the development of monthly averages of mean fluid temperature on the ground side in Langen and minimum and maximum values for temperature during peak-load

conditions, compared with the annual averages of temperature at a BHE in the field

97


ISSN 2067-5534 © 2012 JSE

4. Application and operation strategies

Monitoring allows for understanding the functioning of a GSHP plant. Since the earliest experiments, detailed monitoring campaigns have been executed in many countries. The earliest monitoring report found in literature was published exactly 60 years ago [11]. 20 years ago, a listing of monitoring results and literature for the first phase of GSHP application in Europe and North America was given in [12]. a) Monitoring energy flows

Data from monitoring were already used in the previous chapter, for validation of design software. The main reason for monitoring is, however, the check of function and efficiency of a certain project. For the Langen GSHP project from the previous chapter, some energy figures are given here (Table 3, cf. [9]). As the

GSHP in this relatively large project (154 BHE) is not the only source for heat and cold, the whole building system had to be considered for evaluation and the geothermal share to be determined. Figure 11 shows the total specific heating and cooling loads and the part covered by the ground.

In figure 11 also the electricity consumption of the building and the share of electric power used by the GSHP system (ground-side pumps and heat pumps) can be seen. The specific energy loads are calculated using the net floor area (NFA) of the building. Some constraints are given from a number of drinking-water wells about 1 km away in the direction of groundwater flow; heating up of the groundwater was not allowed, and thus heat extraction must be higher than heat injection on the long term. Table 2 shows that this goal (given with a ratio 1 : 0.87 in the design) was achieved in all years covered.

Heating

0

10

20

30

40

50

60

70

80

2009 2010 2011

kW

h/m

2 /a

district heating

geothermal heat pump

Cooling

0

10

20

30

40

50

60

70

80

2009 2010 2011

kW

h/m

2 /a

conventional chiller

direct geothermal

Electricity

0

10

20

30

40

50

60

70

80

2009 2010 2011

kWh

/m2 /a

electric consumption

used for geothermal

?

Figure 11: Annual specific energy use (kWh/m2/a, for NFA) in the Langen building, and geothermal

contribution or share in the case of electricity consumption, respectively; dotted lines: design values (after [9])

Table 3: Annual performance (SPF) and geothermal share of the heating and cooling

energy supplied to the Langen building (after [9])

design 2009 2010 2011

SPF total H/C --- 8.2 7.1 7.9

SPF heating 5 6.5 5.6 6.1

SPF cooling > 8 9.9 9.9 12.0

geoth. share heat 75 % 23.1 % 25.3 % 26.3 %

geoth. share cold 82 % 53.6 % 54.0 % 49.5 %

b) Example of design for specific climate conditions

In the southwest of Spain, a new retail outlet was planned in Jerez de la Frontera [13]. A little more than fifteen km from the Atlantic Ocean, Jerez is characterized by mild winters and very hot and dry summers, with 17.7 °C annual average. The extreme temperatures in August in a long-term average rise to 33.1 °C maximum and fall to 18.4 °C minimum, and the actual readings exceed 38 °C each year on several occasions. Thus cooling demand in this region exceeds any heating demand by far, in particular in commercial buildings with lot of internal heat sources. Designing a GSHP for cooling under these

conditions requires unconventional solutions; seasonal storage is hardly feasible, with mean temperatures in winter not lower than 10 °C.

The company owning the retail outlet has equipped already a number of its large stores with GSHP, mainly in Northern Europe, but the one in Jerez is quite different for the specific climatic conditions it has to deal with. Given the climate of Jerez and the building design and concept used for the retail building, there is a totally unbalanced thermal energy demand:

Heating demand: 75 MWh/a Cooling demand: 4’104 MWh/a

Thus heat accounts for only 1.8% of the demand for cooling. The monthly building loads are given in figure 12; even in winter, the monthly cooling demand is higher than heating demand!

The design target under these conditions was to create a geothermal HVAC system that covers the full (small) heating demand and a part of the total cooling demand as large as possible. For this extremely unbalanced situation, a substantial part of the cooling can only be covered if sufficient cold is stored in the underground, or in other terms, surplus heat is extracted from the underground. The final design hence did not only include cold storage in wintertime for a seasonal

98


ISSN 2067-5534 © 2012 JSE

balancing, but also short-term cold storage during night in summer. With using all time available for heat extraction, considering the periods when ambient air temperature is sufficiently lower than ground temperature, a maximum annual cooling supply of about 700 MWh/a might be achieved.

Heating

0

100

200

300

400

500

600

700

800

900

1 3 5 7 9 11

month

MW

h /

mo

nth

Cooling

0

100

200

300

400

500

600

700

800

900

1 3 5 7 9 11

month

Figure 12: Monthly heating and cooling loads as to building design for Jerez retail outlet (after [13])

For the BHE design, several thermal response tests (TRT) had been done in advance with a resulting thermal conductivity of 1.5 W/m/K. The undisturbed under-ground temperature was 19.8 °C, a rather high value compared to classical GSHP countries like Sweden or Germany.

From economic considerations, the maximum number of BHE was limited to 50, with a maximum distance of 8 m among each, and a maximum depth of 130 m. So the primary design task was to check what would be the maximum cooling that could be provided by a BHE-field of this size. Calculations using a standard approach resulted in the possible loads as shown in table 4; of the total annual cooling demand of >4 GWh, only about 7 % could be covered from the ground that way.

As the percentage of geothermal coverage of the cooling load is so small, an almost steady operation over the whole year for this very base load can be assumed. The heating in wintertime is only able to reduce the heat injection into the BHE field, but not to turn it into heat extraction. As a result, the operation would be dominated by continuous heat dissipation into the underground, and in consequence the ground temperature would rise constantly.

Table 4: Load data on building and ground side for two different scenarios for Jerez retail outlet (cf. [13])

supply to building

geothermal coverage *

expected SPF

BHE extraction for heating

BHE extraction from re-cooling

total BHE extract. / inject.

Standard case

Heating 75 MWh/a 100 % 5 60 MWh/a - 60 MWh/a

Cooling 300 MWh/a 7 % 3 450 MWh/a

Maximum cooling case

Heating 75 MWh/a 100 % 5 60 MWh/a 420 MWh/a 480 MWh/a

Cooling 530 MWh/a 13 % 3 - - 795 MWh/a

* percentage of total building loads

Even in summertime, ambient air at night can be colder than the temperature in the BHE field. As temperature in the underground will rise steadily over the years also when active re-cooling is done (the increase just being slower than in the standard case), the opportunities for re-cooling with nighttime ambient air will improve over time.

Weather data from nearby Cadiz were used to assess the amount of re-cooling that could be done during spring, summer and autumn (example for July given in figure 13). In order to use the cold from the ground efficiently, no geothermal cooling was assumed from November to March, as the lower ambient air temperatures in wintertime will allow for efficient use of air coolers. Using the ground for cooling is more desirable in summer, when ground temperatures are much lower than cooling water from air coolers. The software EED was also used here to calculate the temperature development, and eventually the load data as given in table 4 were deemed feasible.

The complete geothermal system consists of borehole heat exchangers (BHE), heat pump and dry cooler(s). The 50 BHE were finished in 2010, and the

underground thermal storage volume around the BHE now extends to about 553’000 m3. Alas, by the time of writing, no monitoring data could be evaluated yet.

With this innovative design concept, adapted to Mediterranean climate and combining both diurnal and seasonal cold storage, the cooling output from BHE can be increased in a sustainable way. In summer, the underground works as a store of cold during the night and as a sink of heat during the day (diurnal storage). In wintertime, the regular operation of the heat pump for heating extracts some heat from the ground, and additional heat extraction (or re-cooling) is done by dry cooler (seasonal cold storage). 5. Conclusions

Ground Source Heat Pumps (GSHP) are used throughout Europe, in small applications (residential houses) as well as in large projects for commercial or institutional buildings, and in various climatic zones from Northern Scandinavia to the Mediterranean Sea. In particular for large installations, good knowledge of the thermal parameters of the underground and thorough

99


ISSN 2067-5534 © 2012 JSE

forecasting of building loads are crucial in achieving highly efficient and long-term sustainable heating and cooling systems. Recent developments in investigation of ground thermal parameters and design software have been presented above.

With suitable investigation and design tools, the behaviour of a certain shallow geothermal system can be predicted quite accurately. However, depending on building type and climate, further considerations need to be made in order to match building, geology and climate. The example of a retail building in Jerez in southernmost Spain given in this paper shows how GSHP can be adapted even to rather extreme conditions.

In order to understand better the real behaviour of large GSHP systems, and to check if the expected efficiency and planned operation strategy could be met in reality, more monitoring of this kind of installations is required. Alas, the benefits of monitoring typically are not apparent to the building owners, and thus they avoid the related cost. The authors hope that research funds from governments and foundations will be granted more to support monitoring campaigns. The need for this kind of funding was expressed recently within the document on Research Priorities for Geothermal Heating and Cooling [14].

10

15

20

25

30

35

July

Te

mp

era

ture

(°C

)

1 7 14 21 28

average undisturbed ground temperature around BHE (before GSHP operation)

expected temperature development around BHE during GSHP operation

9 K difference

7 K difference

7 K difference

Figure 13: Hourly dry air temperature in July (data for Cadiz, from Spanish Meteorological Service)

and ground temperatures in undisturbed situation and during GSHP operation (cf. [13]) REFERENCES [1]. EurObserv´er – Ground-source heat pump barometer.

Systèmes solaires, N° 205, pp. 82-101, Paris2011 [2]. Sanner, B. – 50 years of ground-source heat pumping. -

Newsletter IEA Heat Pump Center 13/4, p. 4, Sittard. 1995 [3]. Sanner, B., Dumas, P., Fernandez, I., Regueiro, M. –:

Geotrainet, A New European Initiative for Training and Education of Planners, Drillers and Installers of Geothermal Heat Pumps. Proc. WGC 2010 Bali, paper #0905, 5 p.

[4]. Eklöf, C., Gehlin, S. – TED, a mobile equipment for thermal response test. MSc-thesis 1996:198E, LuTH, 62 p., Luleå 1996

[5]. Sanner, B., Reuss, M., Mands, E. – Thermal Response Test, eine Methode zur In-Situ-Bestimmung wichtiger thermischer Eigenschaften bei Erdwärmesonden. Geothermische Energie 24-25/99, pp. 29-33, Geeste 1999

[6]. Sanner, B., Hellström, G., Spitler; J., Gehlin, S. – Thermal Response Test, current status and world-wide application. Proc. WGC 2005 Antalya, paper #1436, pp. 1-9

[7]. Sauer, M., Mands, E., Grundmann, E., Sanner. B. – Erweiterte Anwendungsmöglichkeiten des Geothermal Response Test: Bestimmung der Erdwärmesondenlänge mittels Thermoimpuls. Tagungsband Geothermiekongress 2010 Karlsruhe, paper F11.4, 7 p.

[8]. Hellström, G., Sanner, B., Klugescheid, M., Gonka, T., Mårtensson, S. – Experiences with the borehole heat

exchanger software EED. Proc. Megastock 1997 Sapporo, pp. 247-252

[9]. Bohne, D., Wohlfahrt, M., Harhausen, G., Sanner, B., Mands, E., Sauer, M., Grundmann, E. – Geothermal Monitoring of eight non-residential buildings with heat and cold production, experiences, results and optimization. Proc. Innostock 2012 Lleida, paper #INNO-U-26, 10 p.

[10]. Sanner, B., Mands, E., Sauer, M. – Larger geothermal heat pump plants in the central region of Germany. Geothermics 32, pp. 589-602, Elsevier, Amsterdam 2003

[11]. Harlow, J.H., Klapper, G.E. – Residential Heat Pump Experiments in Philadelphia, Installation and Operating Experience. AIEE Trans 71/II, pp. 366-375, New York 1952

[12]. Sanner, B. – Erdgekoppelte Wärmepumpen, Geschichte, Systeme, Auslegung, Installation. 328 p., IZW-Berichte 2/92, Karlsruhe 1992

[13]. Fernández, A., Mands, E., Sanner, B., Sauer, M., Novelle, L. – Underground diurnal and seasonal energy storage for a cooling and heating system in a retail building in Jerez de la Frontera / Spain. Proc. Innostock 2012 Lleida, paper #INNO-U-24, 8 p.

[14]. ETP-RHC – Strategic Research Priorities for Geothermal Technologies. European Technology Platform for Renewable Heating and Cooling, 70 p., Brussels 2012 download from: http://www.rhc-platform.org/publications/

100


ISSN 2067-5534 © 2012 JSE

ASPECTS RELATED TO WIND POWER PLANTS OPERATION IN ROMANIA

OPREA S.V., PETRESCU D.E., BOLBORICI D., STĂNESCU O.R.

National Power Grid Company - Transelectrica, Bucharest [email protected]

Abstract – The European Union’s (EU) vision implies tremendous changes in the power systems. These should contribute to sustainable development and protection of the environment by enabling the EU to achieve its targets of a 20% reduction of greenhouse gas emissions, 20% increase in energy efficiency and 20% of renewable energy in final energy consumption by 2020. Encouraging of the renewable energy sources (RES) by the Romanian Government through different incentives lead since 2007 to a large volume of projects. Nowadays, the installed wind power in Romania is about 1140 MW, most of them (1103 MW) being concentrated in the south-eastern part of the country, called Dobrogea. Taking into account the different specificity of the WPP operation in different countries, based on recorded data in Romania, a couple of analyses have been performed in order to understand and try to better predict it for medium and long term network planning. These analyses were focused on full load hours, average loading of WPP, particularity of operation at night and depending on seasonal conditions, variability of WPP output for short time periods and implications on daily load curve.

Keywords: WPP integration, full load hours, short time variability of WPP, grid reinforcements 1. INTRODUCTION

The EU's energy and climate policy objectives consist in completing the internal market in energy, guaranteeing security of supply, notably for gas and oil, reducing greenhouse gas emissions by 20%, increasing the share of renewable energy in the final energy consumption to 20% and achieving a 20% increase in energy efficiency by 2020.

Romania as one of the State Members has to fulfill its obligations related to EU’s targets in terms of RES integration. The incentive support scheme for RES has been enacted by Law no. 220/2008 for establishing the promoting scheme for energy produced out of RES, Law no. 139/2010 (modifying Law 220/2008) and a series of four governmental orders dated November 2011. As a result of the supporting scheme mainly based on green certificates, since 2007, the National Grid Company – Transelectrica, received a large number of applications for connection. Most of them are located in Dobrogea, Moldova and Banat areas as in Fig. 1.

Fig. 1. Areas with large projects in Romania

This concentration of interest from the private

investors coincides with the wind potential map as in Fig. 2 [1].

Fig. 2. Wind potential in Romania

Starting from 2010, installed power increased from

13 MW to 400 MW by the end of the year. In 2011, the installed power was almost double (700 MW) compared with the previous year. The first project and most developed one is connected in the new substation 400/110 kV Tariverde.

101


ISSN 2067-5534 © 2012 JSE

Mountain

m/s W/m2 Off-shore m/s W/m2

Coast m/s W/m2

Plain m/s W/m2

Hills m/s W/m2

>11,0 >1800 >9,0 >800 >8,5 >700 >7,5 >500 >6,0 >250

10,0- 1200- 11,5 1800

8,0- 300- 9,0 800

7,0- 400- 8,0 700

6,5- 300- 7,5 500

5,0- 150- 6,0 250

8,5- 700- 10,0 1200

7,0- 400- 8,0 600

6,0- 250- 7,0 400

5,5- 200- 8,5 300

4,5- 100- 5,0; 150

7,0- 400- 8,5 700

5,5- 200- 7,0 400

5,0- 150- 6,0 250

4,5- 100- 5,5 200

3,0 50- 4,5 100

<7,0 <400 <5,5 <200 <5,0 <150 <4,5 <100 <3,5 <50

These figures are given in Table 1. Due to the fact

that the installing process is very dynamic, the total figures are approximate.

Table 1. Installed power in 2010 and 2011

Interval 2010-2011

Pi Tariverde [MW]

Pi Total [MW]

01.01.2012 387.5 1140

November 2011 300 780

August 2011 300 563

June 2011 300 518

March 2011 300 518

February 2011 300 424

January 2011 300 424

Total 2011 About 700

October 2010 264 322

September 2010 222.5 n.a.

August 2010 166 n.a.

January 2010 0 13

Total 2010 About 400

2. FULL LOAD HOURS

Based on the recorded data between August 2010 and August 2011 and between December 2010 and December 2011, two duration curves describe the behaviour of wind generation in these relatively short time intervals as in Fig. 3.

Fig. 3. WPP output in August 2010-2011 and December 2010-2011

Due to the dynamic process of WPP installation, installed power was roughly estimated at different time intervals. Therefore it is difficult to draw very precise conclusions, but roughly in 50% of the time in the first interval, WPP output was less than 55 MW, while in the second interval considered, 50% of the time their output was less than 75 MW. These figures represent less than 10% of the installed power in both intervals (considering 563 MW installed power in the first interval and 800 MW installed power by the beginning of December 2011).

10% of the time WPP output was more than 50% of the installed power in both time intervals. For different percentages of the time, the results are given in Table 2.

Table 2. Full load hours

First interval

Second interval

% of the time MW

% of the Pi MW

% of the Pi

10% >258 45.83 >303 37.88

30% >117 20.78 >147 18.38

50% >55 9.77 >75 9.38

70% >21 3.73 >31 3.88

3. WPP OUTPUT VARIATIONS

Based on the same recorded data in the interval

August 2010 – August 2011, from ten to ten minutes (consisting of almost 52000 data), the maximum variation of WPP output, from moment t to moment t+1, was +180/-214 MW as in Fig. 4.

102


ISSN 2067-5534 © 2012 JSE

Fig. 4. WPP output variations

It can be easily noticed that in most of the cases

variations are concentrated in -50 and +50 interval. Only few values are out of this range.

Tabel 3. WPP output – large variations

Variations more than dif >50 dif >100 dif>150 dif>200

No. of variations 128 22 3 0

% 0.247 0.042 0.006 0.000

Variations less than dif<-50 dif<-100 dif<-150 dif<-200

No. of variations 136 18 5 1

% 0.262 0.035 0.010 0.002

Very short term variations of more than -50/+50 MW in ten minutes were recorded in about 0.25% of the total number of 51914 records.

Variations in between +10/-10 MW and +20/-20 MW are much more frequent (8% of the total number of records for variations between +10 and -10 MW, 2,5% of the total number of records for variations between +20 and -20 MW). In other words, small amplitude of variation is more frequent as in Table 4.

Tabel 4. WPP output - Smaller variations

Variations more or less than dif>10 dif<-10 dif>20 dif<-20

No. of variations 4339 4179 1332 1274

% 8.358 8.050 2.566 2.454

However we need to be aware of the fact that the

time interval is not enough to complete this analysis. The obtained results are based on a relatively short time interval, but in the future, more data will be used and the results will be more precise.

The trend of variations in case of the biggest farm, WPP Fantanele+Cogealac is very similar. We expect many variations around +/-20 MW, but cases of variation around +/-100 MW can be found.

Fig. 5. WPP Fantanele+Cogealac output variations

From other WPP operation in Europe, it could be noticed that if the WPP was spread over a large area, these variations tend to diminish (Fig. 6), due to the fact that wind conditions are slightly different [2].

Fig. 6. WPP output for different size of the plant 4. HOURLY AVERAGE OUTPUT

WPP Fantanele+Cogealac is the biggest WPP in

Romania (located in Dobrogea area). Based on the hourly recorded data provided by OMEPA – Metering Operator, it was found that at the off-peak WPP Fantanele+Cogealac produces more than at peak consumption. Each hourly mean value for WPP Tariverde and total WPP in Dobrogea is given in Tabel 5.

Tabel 5. Average hourly output WPP Fantanele+Cogealac and WPP total in Dobrogea in 2011

Average hour Average Pg [MW] - Tariverde

Average Pg [MW] - Dobrogea

Average hour 1 75.4 123.76
























103


ISSN 2067-5534 © 2012 JSE

In both cases (WPP Fantanele+Cogealac and total WPP in Dobrogea), the lowest value was recorded at 11 a.m., while the highest value was recorded at 5 a.m. for WPP Fantanele+Cogealac and 6 a.m. for total WPP in Dobrogea. Comparing mean WPP Fantanele+Cogealac output values with a randomly chosen daily load curve [3], it is obvious that wind blows more at night and the daily generation curve of wind farmas increases stress on system operation as in Fig. 7.

Fig. 7. Hourly average output WPP Fantanele+Cogealac and daily load curve

If we consider all WPP from Dobrogea area, the trend is similar, only the average is higher as in Fig. 8.

Hourly average WPP Total 2011

0

20

40

60

80

100

120

140

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Hours

MW

Fig. 8. Hourly average output of WPP in Dobrogea

area

This trend of the wind behaviour is confirmed on monthly basis as in Fig 9. It can be noticed that summer months were less windy in 2011. In February and March mean output values were the highest and more constant compared with the rest of the year.

Hourly Average Output in 2011, WPP Tariverde

0.0

20.0

40.0

60.0

80.0

100.0

1 3 5 7 9 11 13 15 17 19 21 23

Hour

Pg

[M

W]

Hourly Average of Pg - June 2011

0

20

40

60

80

100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - January 2011

0

10

20

30

40

50

60

70

80

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - April 2011

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - May 2011

0

10

20

30

40

50

60

70

80

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - February 2011

0

20

40

60

80

100

120

140

160

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - March 2011

0

20

40

60

80

100

120

140

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

104


ISSN 2067-5534 © 2012 JSE

Fig. 9. Monthly Pg hour by hour

Daily variation between generation of WPP Fantanele+Cogealac at night minimum load hour and next day maximum load hour was also anaysed.

Duration curve of this variations (fig.10) show that in 10% of cases the increase of generation from night to day is more than 30% of the installed power.

Duration curve Pg_NightLow-Pg_DayMorning Peak [MW]

-250

-200

-150

-100

-50

0

50

100

150

200

250

300

0 10 20 30 40 50 60 70 80 90 100

%

Duration Curve Pg_Night Low-Pg_DayEvening Peak [MW]

-300

-200

-100

0

100

200

300

0 10 20 30 40 50 60 70 80 90 100

%

Fig.10. daily variation between generation of WPP Fantanele+Cogealac at night minimum load hour and next day (morning/evening peak load hour)

This situation is indicative on the necessary

supplementary flexibility of the rest of the generation park needed to accommodate increased wind generation in the system.

It should be further studied, when enough data shall be available, the degree of correlation between generation in one farm and total generation in a neighboring area and in the whole system.

5. CONCLUSIONS Taking into consideration the amount of connection

requests, the grids must be urgently extended and upgraded, including through electricity highways, to foster market integration and maintain the existing levels of system's security, but especially to transport and balance electricity generated from renewable sources, which is expected to more than double in the period 2007-2020. At the same time, reaching the EU's 2020 energy efficiency and renewable energy targets will not be possible without WPP integration.

It is very important to know the particularities of WPP operation in a specific area in Romania. Based on the recorded data over a relatively short time interval, the results of a few analyses have been shown in this paper. Main conclusions are:

- 10% of the time WPP output was more than 50% of the installed power;

- the maximum variation of WPP output, in ten minutes, was +180/-214 MW

Hourly Average of Pg - December 2011

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - July 2011

0

10

20

30

40

50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - August 2011

0102030405060708090

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - September 2011

0

10

20

30

40

50

60

70

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - October 2011

0

20

40

60

80

100

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

Hourly Average of Pg - November 2011

0

10

20

30

40

50

60

70

80

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

ora

MW

105


ISSN 2067-5534 © 2012 JSE

- in most of the cases variations are concentrated in -50 and +50 interval and only few values are out of this range;

- variations in between +10/-10 MW and +20/-20 MW are much more frequent;

- wind blows more at night increasing stress on the system operation;

- in 10% of cases the increase of generation of awind farm from night to day is more than 30% of the installed power

- summer months were less windy in 2011; - in February and March mean output values were

the highest and more constant compared with the rest of the year.

More data will be used and the results will be more precise. REFERENCES [1]. Turcu, I. – Evaluarea potenţialului eolian în zone de

interes din România pe bază de măsurători în concordanţă cu procedurile actuale pe plan European, stabilirea soluţiilor de valorificare şi indicatori de performanţă ICEMENERG, Bucureşti, 2007

[2]. Oprea, S. – Aspecte privind accesul deschis la reţelele electrice – Integrarea surselor regenerabile de energie – Teză de doctorat, Bucureşti, 2009

[3]. http://www.transelectrica.ro/4OperareSEN/functionare.php

106

JOURNAL OF SUSTAINABLE ENERGY, VOL. III, NO. 2, JUNE, 2012

ISSN 2067-5534 © 2012 JSE

GROUND SOURCE HEAT PUMPS FOR MEDITERRANEAN CLIMATE: PRESENT STATUS AND EXPECTATIONS FROM

THE GROUND-MED PROJECT

MENDRINOS D., KARYTSAS C. Centre for Renewable Energy Sources and Saving,

19th km Marathonos ave., 19009 Pikermi Attikis, Greece [email protected]

Abstract - Thanks to the European support under the FP7 program, the Ground-Med partners, after three years of project activities, have developed a new generation of energy efficient ground source pump systems, including heat pumps, fan-coil units and air handling units, which have been installed and provide heating and cooling at the eight demonstration buildings of the project, located in South European EU member States. First estimates, based on COP laboratory measurements and monitoring data from project demo sites, indicate that the project objective of SPF>5.0 can be exceeded, even when considering at the calculations the electricity consumption of both the heat pump compressor and the external (BHE) pump. Apart from using energy efficient equipment, key parameters towards achieving maximum SPF, are the BHE and system design, system operating conditions and control methodology used. Keywords: geothermal heat pumps, GSHP, BHE, heating and cooling. 1. INTRODUCTION

With more than 1 million installed units totaling 11 GWth in European Union and annual sales of 125000 units rising 10% annually, ground source heat pumps (GSHP) are an established and mature technology gaining market share in Europe. They have been providing efficient heating and domestic hot water in central and North Europe for the past few decades. During the last five years, GSHPs have also penetrated the markets of South Europe providing both heating and cooling.

GSHPs are one of the most energy efficient technologies for heating, cooling and domestic hot water provision, with measured seasonal performance factor (SPF) improved by 40% compared to air source heat pumps [1]. Due to their high efficiency, they are considered as renewable energy technology according to the directive 2009/28/EC of the European Parliament and European Council of 23.04.2009.

2. THE GROUND-MED PROJECT

The main objective of project TREN/FP7EN/ 218895/GROUND-MED “Advanced Ground Source Heat Pump Systems for Heating and Cooling in Mediterranean Climate” is to develop and demonstrate a new generation of ground source heat pumps systems with measured seasonal performance factor (SPF) exceeding 5,0. SPF is defined as the ratio of the useful energy delivered to the building during a year in terms of heating, cooling and domestic hot water, divided by the electricity consumption during the same period. SPF equals to the average COP throughout the year. In cases where heating can be separated form cooling, a smaller period can be considered as well for the SPF calculation, such as the heating season or the cooling season.

GROUND-MED project started on 1 January 2009 and has duration of 5 years until 31 December 2013. Its budget amounts at around 7.25 million euro, including technology development, demonstration, monitoring and dissemination activities. It is supported by the European Union through its FP7 program with up to approximately 4.3 million euro.

GROUND-MED project is implemented by a consortium of 24 organizations across Europe, including universities and research institutes, heat pump manufacturers and industrial associations, GSHP systems installers and energy consultants and works contractors. They are: Centre for Renewable Energy Sources and Saving -

CRES, (Greece) as project coordinator, Compagnie Industrielle d'Applications Thermiques –

CIAT (France), HIREF spa (Italy), OCHSNER Wärmepumpen GmbH (Austria), Institute of Systems and Robotics - University of

Coimbra (Portugal), University of Oradea (Romania), Gejzir Consulting (Slovenia), Ecoserveis (Spain), Edrasis – Ch. Psallidas SA (Greece), Eneren srl (Italy), University College Dublin (Ireland),

107


ISSN 2067-5534 © 2012 JSE

Università degli Studi di Padova (Italy), Universidad Politécnica de Valencia (Spain), Commissariat à l'Énergie Atomique - CEA (France), Groupement pour la Recherche sur les Echangeurs

Thermiques - GRETh (France), Instituto Politecnico de Setubal (Portugal), KTH – Royal Institute of Technology (Sweden), Geoteam GmbH (Austria), Groenholland BV (Netherlands), Besel SA (Spain), Centre Technique des Industries Aérauliques et

Thermique – CETIAT (France), Fachinformationszentrum Karlsruhe GmbH – FIZ

(Germany), European Heat Pump Association - EHPA, European Geothermal Energy Council - EGEC.

More information about the project, as well as on GSHP applications and news, are provided at the project web site, at the internet address: http://www.groundmed.eu/.

3. TECHNOLOGY DEVELOPMENT

a) Heat pump boundary conditions GROUND-MED project regards GSHPs as an

integrated system comprising the field of borehole heat exchangers (BHE), the water source heat pump, the external (outdoor) heat transmission circuit with its pump (circulator), the internal (in-building) heat transmission circuit with its pumps and the heat supply system, which depending on the site are fan-coils, in-wall pipes, radiators and/or air-handling units. Technology development concerns not only new energy efficient equipment, but system design and operation as well, aiming in maximizing overall energy efficiency in terms of SPF.

1

2

3

4

5

6

7

8

30 35 40 45 50

Condensing Temperature, ºC

CO

P

Evaporation T = +5 ºC

Evaporation T = −3 ºC

GROUNDHIT

GROUND-MED

Typical GSHPwith fan-coils

R410A refrigerant4K superheat & subcooling90% motor efficiency 80% compressor isentropic efficiencyResults by NIST/CYCLE_D heat pump cycle simulator

Fig. 1 - SPF improvement in GROUND-MED heat

pump systems

The GROUND-MED concept towards maximizing the seasonal performance of a GSHP is presented in Fig. 1, which shows the heat pump COP as a function of the condensing temperature for two evaporation temperatures, one equal to -3ºC (dark red line) and another at +5ºC (pink line). Both lines have been calculated with the aid of the NIST/CYCLE_D heat pump cycle simulator, using the assumptions listed in

Fig. 1, which correspond to state of the art equipment as described below.

R410A refrigerant is the new trend in heat pump market, resulting in high energy efficiency, not only because of its heat transfer properties, but because the corresponding heat pump components are new; they have been designed and made available to the market recently, when energy efficiency is one of the main market driving forces. In addition, its wide adoption by the heat pump industry results in economies of scale and hence improved competitiveness. The latter is necessary in order to achieve maximum replication of the developed technology and increased project impact.

Superheat and subcooling of 4K are typical for efficient heat pump cycles, according to our 8 years experience in heat pump technology development; cycle optimization however, may further fine-tune these values.

Motor efficiency of 90% corresponds to the most efficient motors for compressors of low capacity, while motor efficiency of 95% can be reached in systems of high capacity, with maximum value of 97% for horizontal coaxial centrifugal compressors with magnetic cushion bearings.

Compressor isentropic efficiency of 80% corresponds to the best market available compressors operating at optimum pressure ratio.

Evaporation temperature of -3ºC corresponds to typical BHEs design with water and antifreeze as heat transfer fluid. It is encountered in the majority of GSHP installations as a compromise between energy efficiency and costs. In GROUND-MED project, larger BHE fields are used in terms of total length, as BHEs are designed for +5ºC evaporation temperature and water as heat transfer fluid. This improvement alone results in the heat pump to operate along the pink line of Fig. 1, instead of the red one, with a corresponding improvement in the heat pump (compressor) COP in the range 20-35% depending on the condensing temperature. The same principle applies in cooling mode, where the longer BHE will result in lower operating temperature in the BHE and hence in improved heat pump COP.

As shown in Fig. 1, the condensing temperature is of major importance to the determination of heat pump COP. Traditional design of fan-coil systems uses 45ºC as temperature of the water supplied by the heat pump to the fan-coil. This value was defined in the past, when energy costs were negligible and system reliability and users comfort were the main driving forces, whilst energy efficiency was of no importance at all. Resulting heat pump operation range is shown in the low right part of the graph of Fig. 1, which corresponds to condensing temperature to a constant value in the range of 47-50ºC.

In the GROUNDHIT project, which was the GROUND-MED predecessor, the temperature supplied by the heat pump to the fan-coils was set to 40 ºC, which resulted in 25% higher COPs than standard systems.

GROUND-MED systems are designed in such a way, so that the temperature supplied by the heat pump to the fan-coils in heating mode varies depending on the load conditions, starting from 40ºC at peak load, and going down to 30ºC at low load conditions. The same principle applies during the cooling season as well, e.g.

108


ISSN 2067-5534 © 2012 JSE

by cooling with 10ºC fluid at peak load conditions, increasing this temperature as the load drops and reaching a pre-defined maximum value, say 20ºC at minimum load requirements. This requires a temperature compensation function to be introduced in the heat pump controller and results in a further 25% COP improvement. In this function, the heating or cooling temperature delivered by the heat pump depends on the ambient temperature, which is the main parameter defining the heating or cooling load of the building. An example of user defined heating and cooling compensation functions is shown in Fig. 2.

0

10

20

30

40

50

-10 0 10 20 30 40

ambient temperature, oC

wa

ter

tem

pe

ratu

re d

eliv

ere

d b

y t

he

he

at

pu

mp

, oC

heating

cooling

Fig. 2 - Sample temperature compensation functions

for heating and cooling System control is done through the heat pump

controller supplied by the heat pump manufacturers. In addition to temperature compensation, a smart control models and strategies are under development for testing and evaluation at different demonstration sites.

b) Heat pump prototypes

Eight heat pump prototypes have been developed for the project, each one tailor made for the needs of the corresponding demonstration site. Three heat pump prototypes have been developed by CIAT, two by HIREF and three by OCHSNER WP. All heat pumps are characterised by superior energy efficiency, using the methodology and experience already gained by previous technology development projects, such as the GROUNDHIT. It includes oversized evaporators and condensers, electronic expansion valves, refrigerant distribution device at the evaporator and compressors of high isentropic efficiency. Also COP optimization in terms of calculation and optimization of heat exchangers, refrigerant charge, optimization of refrigerant lines, definition and dimensioning of heat pump components.

The main difference of the GROUND-MED heat pumps is that the focus of the technology development is now placed on the SPF optimisation rather than the COP. COP reflects system performance under specific conditions, which may correspond to its operating parameters during limited time periods only. On the other hand, SPF is the parameter that reflects the actual energy efficiency of the installation throughout its operation during both heating and cooling seasons.

In order to optimize the heat pump SPF, two aspects are of major importance. The first one is the capacity control of the heat pump and the second is the flow directions at both the evaporator and the condenser.

The heat pump capacity control methodologies considered and studied with laboratory experiments were the following: Constant speed compressors with on-off capacity

control. Inverter compressors with variable speed capacity

control. Twin constant speed compressors in tandem. Two constant speed compressors of different capacity

in tandem. One constant speed and one variable speed compressor

in tandem. Due to the high volumetric heat capacity of the water

and the corresponding thermal inertia and thermal storage at the water distribution circuits, and in contrast to air-to-air heat pumps, on-off compressors perform quite satisfactory in ground source systems in terms of SPF. In addition, the compressors of the highest isentropic efficiency available in the market are constant speed scroll type ones. Three heat pump prototypes developed by OCHSNER WP use on-off control and R407C as refrigerant.

Inverter compressors have the advantage of achieving improved temperature approach between the refrigerant and the water circuits at low capacity [2], as well as reduced in number and smooth compressor start-ups. Both result in higher heat pump SPF, while the latter also results in improved reliability and longer compressor life. However, in contrast to air-to-air systems, due to the superior heat transfer properties of the water the temperature approach difference is mitigated in ground source heat pumps, limiting that way the inverter control benefits.

On the other hand, additional electricity losses occur at the inverter resulting in lower SPF by 10% at peak load conditions. Moreover, both inverter and compressor motor efficiency tend to drop as the machine operates away from its nominal point (usually at peak load conditions). This drop in efficiency can become very profound at extreme frequencies. Furthermore, there is limited market availability of efficient inverter compressors. In order to evaluate this option, one heat pump prototype developed by HIREF and installed at its factory in Tribano, Italy, uses an inverter compressor and R410A as refrigerant.

Compressors in tandem exploit the market availability of constant speed compressors of high isentropic efficiency, while achieving excellent temperature approach at both the evaporator and the condenser, as 80% of the time only one compressor runs using the entire heat exchange surface of the heat pump, which has twice the compressor’s nominal capacity. Laboratory experiments done for the project indicated that the option of twin constant speed compressors in tandem results in the maximum possible heat pump SPF. The three CIAT heat pump prototypes and one HIREF heat pump prototype use twin constant speed compressors in tandem with R410A as refrigerant.

The temperature approach between the refrigerant and the water circuits at the heat pump evaporator and condenser is minimized (and hence SPF is maximized), when these fluids are in counter-flow. Market available ground source heat pumps are usually refrigerant

109


ISSN 2067-5534 © 2012 JSE

reversible; when optimized for heating, their heat exchangers operate in counter-flow in heating mode, and in parallel flow in cooling mode. The reverse occurs if they are optimised for cooling. In order to maximize SPF, counter flow conditions are required in all modes. In GROUND-MED systems, when switching from heating operation to cooling and vice versa, this is achieved by also reversing the external flow together with the refrigerant flow, using either four 3-way valves, or two 4-way valves.

In case of water side reversible heat pumps, the hydraulic system is designed in such a way, so that counter flow conditions occur at the heat exchangers during all times, again using the four 3-way valves. A schematic presentation of a water side reversible heat pump is shown in Fig. 3 [3].

Fig. 3 - Water side reversible heat pump: schematic presentation of switching external flow with four 3-

way valves

c) Fan coil units The entire fan coil unit has been re-designed in order

to minimize its electricity requirements, using brushless permanent magnet motors, low weight impellers and variable speed control. The new fan coils delivered by CIAT should be able to reduce electricity consumption by 75% compared to their predecessors, e.g. from 40 Watts to 10 Watts per unit.

In order to operate at the low temperature range required by the project, and as traditional fan-coils are designed for 45ºC water supply temperature, a new generation of fan-coils was developed, which ensures that user comfort is also achieved together with the energy efficiency improvement. Comfort is also aided by adjusting the fan-coil position and fan speed in order to exploit the “coanda effect” and maintain a small vertical air flow at the walls and a small horizontal air flow at the ceiling and floor. These fan-coils are able to provide users comfort at the same water supply temperature with the in-wall and floor heating systems, where its standard value is 35ºC.

d) Air handling unit prototype

Air handling units condition incoming air to the building, proving heating, cooling and de-humidification. During the summer, de-humidification is achieved by cooling the air in order to remove excess humidity and re-heating it to the desired temperature. In standard air-handling units an electrical resistor is used for re-heating the air, even during the summer. In GROUND-MED

project, CIAT develops a new air handling unit prototype that uses condensing heat for de-humidification, improving that way the unit efficiency in terms of primary energy. The unit is also equipped with an inverter driven fan.

e) Thermal storage nodules

Thermal storage is a technology gaining acceptance for use with heat pumps, as it allows the reduction of peak loads and use of heat or cold produced during off-peak hours of the day (e.g. during the night), when electricity tariffs are lower. Apart from reducing heating and cooling operation costs, thermal storage results in improving primary energy efficiency, as it results in utilising otherwise lost electricity produced during off-peak hours, when electricity demand is less than the supply from coal plants, nuclear plants and wind farms.

Standard cold storage material is ice, which exploits the latent heat of water during its phase change from solid to liquid and vice-versa at 0ºC. In the GROUND-MED project new cold storage nodules are under development by CIAT. They use a material which changes phase at 8ºC, resulting in higher evaporation temperature at the heat pump, than in the case of ice, improving that way its SPF, as already explained in this paper.

In addition, at the Coimbra demonstration site, thermal storage will be also installed, but in order to store heat. A salt hydrate will be used, sodium thiosulfate pentahydrate in particular, which changes phase at ~46ºC, temperature sufficient to allow heating with 40ºC water at peak load conditions [3].

4. DEMONSTRATION SITES All eight project demonstration systems are now in

operation, demonstrating the developed technology. Their main features are summarised in the next paragraphs [4].

CIAT commercial agency office building at Septèmes les Vallons, near Marseille, France, has 338 m² surface, heating requirements 25 kW and cooling needs 15 kW. It has 6 BHEs x 100 m deep each, with double-U PE pipes of 40 mm diameter and bentonitic grouting. The heat pump prototype has tandem compressors, R410A fluid, yielding SCOP=6.2 and SEER=7.3, as calculated from laboratory COP measurements according to the standard PrEN14825-2008. The indoor system comprises 14 CIAT fan-coils and one air handling unit prototypes. Cold storage prototype nodules will be also installed within 2012. Maximum system SPF is achieved by variable speed pumps and by 6 three-way valves, which allow counter-current flow at the evaporator and condenser in both heating and cooling modes, together with free cooling operation when possible.

The Regional Administration building of Coimbra, Portugal, is a renovated old milling factory. Its top floor is heated and cooled by another CIAT heat pump prototype, which covers heating needs of 34 kW and cooling needs of 48 kW. 7 BHEs x 125 m deep each have been constructed, using 32 mm double-U PE pipes and 3-8 mm graded sand grouting. The heat pump has tandem

110


ISSN 2067-5534 © 2012 JSE

compressors, R410A fluid, yielding SCOP=5.65 and SEER=6.19, as calculated from laboratory COP measurements according to PrEN14825-2008. The indoor heating and cooling system comprises CIAT Coadis-2 roof fan-coil units of high efficiency and variable speed, as well as an 100 kWh PCM tank for heat storage.

The Municipal cultural centre of Benedikt in Slovenia, includes a large auditorium and offices, and has heating needs of 20.3 kW and cooling requirements of 19.2 kW. It has 3 double-U BHEs with 40 mm PE pipes of depths 166 m, 126 m and 97 m. An OCHSNER heat pump prototype of COP=6.1 at W10/35 has been commissioned. The system is able to provide free cooling as well.

At HIREF factory in Tribano near Padova, Italy, the installed heat pump prototype covers not only the heating (13.2 kW) and cooling (14.0 kW) needs of the factory technical office, but the sanitary hot water demand of the factory as well. It has 4 BHEs x 80 m deep each and bentonitic grouting. Both single and double U pipes are used for evaluation purposes. It is the only system that uses a heat pump with an inverter compressor. The heat pump is both refrigerant and water reversible with the aid of two 4-way valves. Its heating and cooling system includes variable speed pumps, and inverter driven brushless motor fan coils.

“La Fabrica del Sol” is an information centre for renewable energies, renovated from an old gas factory. It comprises a workshop, offices and social rooms. The ground floor area of 375 m² is heated and cooled by a 61 kW CIAT heat pump prototype, supplied by 7 BHEs x 110 m deep each, of single-U 40mm PE pipes. The heat pump prototype includes tandem compressors and temperature compensation function, yielding SCOP1=5.9 and SEER1=6.7, as calculated from laboratory COP measurements according to standard PrEN14825-2008. The indoor system includes an air handling unit and variable speed pumps of energy class A.

The faculty of visual arts of the University of Oradea in Romania is heated and cooled by a Ground-Med heat pump prototype of OCHSNER manufacture, which covers 38 kW heating and 31 kW cooling peak load. It has 10 BHEs x 130 m depth each in two lines, with single-U, 40mm PE pipe and coarse sand grouting material with a bentonitic plug above the water table. It is the only system that uses antifreeze at the BHE, in particular 10% monoethylene glycol water. The heat pump is reversible at both refrigerant and water sides, and has a COP of 5.64. It provides heating, active cooling, and free cooling through ~1 cm pipes embedded in walls, coupled to 4 fan-coil units.

At the University Polytechnic of Valencia, the second HIREF Ground-Med heat pump prototype provides heating (17 kW) and cooling (15 kW) to 250 m² of offices, computer and service rooms. It has 6 BHEs x 50 m deep each, with single-U 25 mm PE pipes using a variety of grouts. The heat pump prototype includes two Copeland compressors in tandem, R410A fluid and is both refrigerant and water side reversible. The heating and cooling system includes roof fan-coils and inverter pumps. Its data acquisition system has been recording system performance for almost one year now. A lot of

experimentation took place in this system during the past 12 months yielding SPF values depending on the defined system operating parameters, and proving that the project objective of SPF> 5 is feasible, even when including the electricity consumption at the water pumps in the calculations.

At the Edrasis headquarters building near Athens international airport, the heat pump prototype of OCHSNER manufacture covers 55 kW base load heating and cooling, supplying already installed fan-coils. It has 12 BHEs x 110 m deep each as heat source/sink, with 32 mm PE pipes, and graded sand or gravel grouting with an upper 30 m bentonitic plug above the water table. The heat pump is water side reversible with COP of 5.92 at nominal conditions.

5. SYSTEM ENERGY PERFORMANCE The project monitoring system includes local data

acquisition equipment installed at each demonstration site. It includes National Instruments hardware and transducers with Labview interface and FTP communication to a central data management server, located at the Coimbra University.

The thermal energy taken from or supplied to the BHE and building water circuits by the heat pump is measured by heat meters from Brunata or other competent supplier. Electricity consumption at individual system components is measured by Carlo Gavazzi or Itron power meters. Data communication with the interface is done using the RS485 protocol (Modbus).

In order to evaluate the developed technology, the system design methodology and the control algorithms tested, four distinct seasonal performance factors are calculated from the acquired data. These SPFs equal to the ratio of the useful heat and cold delivered to the building divided by the electricity consumption in different system components as explained below. SPF1: heat pump compressor only. SPF2: heat pump compressor and external pump at the

BHE. SPF3: heat pump compressor, external pump at the

BHE and all internal pumps. SPF4: heat pump compressor, external pump at the

BHE, all internal pumps and all fan-coils and air-handling units.

SPF1 is useful for the evaluation of the heat pump itself. SPF2 should be used when comparing ground source heat pumps with other heating and cooling systems, as the external pump is unique to GSHPs. SPF3 is useful for the evaluation of system hydraulic design and pumps efficiency. SPF4 is the measure of overall system efficiency, and should be used for the evaluation of the indoor system components, as well as of the control methodology adopted.

Apart from SPF values which are calculated for the entire season, for technology evaluation purposes, similar daily average COP values are also calculated, e.g. COP1, COP2, COP3 and COP4.

As the project data management system is expected to be fully operational in June 2012, only limited monitoring data were available during the preparation of

111


ISSN 2067-5534 © 2012 JSE

this paper, mainly from the Valencia demonstration site. There, the GROUND-MED heat pump prototype replaced the one already developed for the GEOCOOL European project, which operated for five years using the same BHE and fan-coils.

Comparisons between sample measurements in terms of daily average COPs of the GROUND-MED heat pump [5] and monitoring data from the GEOCOOL project [6] are presented in Fig. 4 for heating and in Fig. 5 for cooling. Although the GROUND-MED data correspond to selected days, they are indicative of what the technology can achieve in terms of energy efficiency, provided that the heat pump operates according to the boundary conditions described above.

7.09

5.81

4.444.09

4.303.70

3.13 3.00

0

1

2

3

4

5

6

7

8

CO

P, d

aily

ave

rag

e

COP1 compressor only

COP2 compressor &

BHE pump

COP3 compressor &

pumps of BHE &building

COP4 compressor &

pumps & fan-coils

GeoCool, Feb 2010

Ground-Med, 22 Dec 2010

Fig. 4 - Daily average COP of the GROUND-MED

heat pump in heating mode compared to its predecessor system (GEOCOOL) at the Valencia

demonstration site

7.56

6.40

4.88 4.775.18

4.38

3.61 3.54

0

1

2

3

4

5

6

7

8

CO

P, d

aily

ave

rag

e


COP2 compressor &

BHE pump

COP3 compressor &


COP4 compressor &pumps & fan-

coils

GeoCool, Sep 2010

Ground-Med, 16 Sep 2011

Fig. 5 - Daily average COP of the GROUND-MED

heat pump in cooling mode compared to its predecessor system (GEOCOOL) at the Valencia

demonstration site

As can be derived from Fig. 4 and Fig. 5, project objectives can be achieved in terms of SPF1 and SPF2, but additional effort is necessary for SPF3 and SPF4 in order to achieve the target value of 5.0. In addition, during the GEOCOOL project, the electricity consumption at the pumps corresponded to the 26-29% of the overall electricity consumption of the system, value that was increased to 34-35% in the GROUND-MED installation, as the heat pump efficiency increased, resulting in lower power consumption at the compressor. Due to the low occupancy of the conditioned offices, only one or two fan-coils are in operation, resulting in their corresponding electricity consumption to account for 2-4% of the total in GEOCOOL and 2-8% in GROUND-MED.

This implies that the pumps and fans correspond to a large fraction of system electricity use and special attention should be paid to them as well. Electricity consumption at the pumps can be minimized by selecting energy class A inverter driven pumps only and regulating the flow according to the load demand from the building.

Preliminary monitoring data are also available from the Coimbra and Barcelona demonstration sites. In Coimbra, the daily average COP1 was calculated from measurements as 6.03, during the first day of its operation for heating [3].

In Barcelona, only instant COP values have been calculated from thermal and electrical power measurements, again in heating mode [7], which are presented in Fig. 6. There, the pumps are energy class A inverter driven ones according to project requirements, resulting in their electricity consumption to account for 7% of total electricity consumption of the system in full load with two compressors running and in 13% in part load with one compressor on. At this site the electricity consumption at the air handling unit is of major importance, which amounts as 13% of total consumption at full load, rising to 25% of total consumption at half load.

6.82

6.045.67

4.275.04

4.78 4.644.03

0

1

2

3

4

5

6

7

CO

P, i

nst

an

t


COP2 compressor &

BHE pump

COP3 compressor &


COP4 compressor &pumps & airhandling unit

Two compressors

One compressor

Fig. 6 - Instant COP of the GROUND-MED heat

pump in heating mode at the Barcelona demonstration site

6. CONCLUSIONS

The GROUND-MED partners have developed a new generation of integrated ground source heat pump systems, comprising borehole heat exchangers of innovative design, water source heat pumps of extraordinary efficiency, pumps and fans of low electricity consumption, as well as smart system controls with built-in temperature compensation function. This technology is demonstrated in eight buildings of South Europe.

BHE main features used in the majority of the project sites include grout of coarse sand (fine gravel) within the water table and bentonitic above, more total borehole meters, and water as heat transfer fluid.

Laboratory measurements and first monitoring data from demonstration buildings indicate that the overall project objective to achieve measured seasonal performance factors SPF higher than 5.0 will be exceeded for SPF1 and SPF2, and may become feasible even for SPF3, when considering the electricity consumption in the external and all internal pumps in the calculations.

112


ISSN 2067-5534 © 2012 JSE

As seasonal performance is the key objective of the project, instead of the heat pump COP at pre-defined conditions, capacity control has been an issue of major importance. Studies, experiments and measurements done during the project proved that twin compressors in tandem outperform the single on-off and the inverter compressor. SPF is further improved by either water reversible, or both refrigerant and water reversible heat pumps, so that both the evaporator and the condenser always operate in counter flow.

As electricity consumption at the pumps and fans corresponds to a significant part of total system electricity consumption, advanced fan-coils and air handling units have been developed, while energy class A inverter driven pumps and fans are also used. Systems are designed in order to provide low temperature heating, high temperature cooling, while also maintaining high users comfort.

Due to the already developed technology, the system design and control methodology adopted and the first promising results from the demonstration sites, we expect the GROUND-MED project to become exemplar for future ground source heat pump systems, effectively contributing towards improving energy efficiency and renewable energy use.

ACKNOWLEDGEMENTS This paper has been prepared towards the

dissemination activities of the GROUND-MED project. The project is supported by the FP7 framework

programme of the European Commission, which is gratefully acknowledged.

REFERENCES

[1] Mendrinos D. - Ground source heat pumps: latest technology and advancement through European projects, in Energy Bulletin No 1 [8], 2010, international sustainable Energy development centre under the auspices of UNESCO (ISEDC), Moscow, Russia, pg 70-76.

[2] Mendrinos D., Corberan J.M., Da Riva E., Del Col D., Montagud C. - Ground Source Heat Pumps for Heating and Cooling in the Mediterranean Countries, in Proc., international conference «Sources/Sinks alternative to the outside Air for Heat Pump and Air-Conditioning Techniques (Alternative Sources - AS)», Padova, Italy 5-7 April 2011, pg 113-122.

[3] Institute of Systems and Robotics - Ground-Med deliverable D6.3: University of Coimbra demo system in Portugal, 2011, pg 11, 31.

[4] Mendrinos D., Karytsas C. – Ground-Med demo sites, in Proc., Ground Med ”the International European project conference”, Marseille, France 6-7 October 2011, pg 40-41.

[5] Universidad Politécnica de Valencia - Ground-Med deliverable D6.5: University Polytechnic of Valencia demo system in Spain, 2011, pg 30.

[6] Montagud C., Corberan J.M., Montero A., Urchueguia J.F. - Analysis of the energy performance of a ground source heat pump system after five years of operation, in Energy and Buildings 43, 2011, pg 3618–3626.

[7] Alsius A.E. - Ground-Med deliverable D6.6: ECOSERVEIS sun factory demo system in Barcelona, Spain, 2011, pg 29.

113


ISSN 2067-5534 © 2012 JSE

MODELING OF WIND POWER PLANTS GENERATORS IN TRANSIENT STABILITY ANALYSIS

STANESCU O., BOLBORICI D., OPREA S.

CNTEE Transelectrica SA, Bucharest [email protected]

Abstract - The fast development of wind power generation brings new requirements for wind turbine integration into the network. After clearance of an external short-circuit fault, grid-connected wind turbines should restore their normal operation. This article presents theoretical issues about transient stability and about wind power plants generators modelling. It describes the types of wind generators used today: double fed induction generator and full scale convertor. In programs developed by different vendors (Eurostag 4.5 and PSS/E 32) have been shown a few simple examples. The transient stability was monitored. Keywords: transient stability, dynamic modelling, wind power plants. 1. THEORETICAL ASPECTS ABOUT TRANSIENT STABILITY

According to the literature [1], [8], [9], the stability of the power systems is defined by their ability to remain in a steady state after the appearance of a small perturbation and return to an acceptable steady state after the occurrence of a large disturbance. A disturbance is defined as a sudden change or as a sequence of abrupt changes of one or more parameters of the electrical system.

Large disturbances (strong or severe) - are those

disturbances that do not allow a linearization of the system of equations, that model the power system operating mode. In this case is used a system of nonlinear equations to model dynamic phenomena. Such large disturbances can be considered: three-phase short circuits, disconnections of generators, consumers or parts of the transmission network.

Small disturbances – are those disturbances that

allow linearization of the system of equations, which model the regime of the system around the initial operating point. Such disturbances occur frequently in the power system, such as small variations in the consumed or generated power.

2. MAIN TYPES OF CONSTRUCTIVE SOLUTIONS

In terms of velocity, wind generators are of two

types: with fixed speed and with variable speed. The ones with fixed speed use squirrel cage induction generator and the ones with variable speed can be with asynchronous or synchronous generator [2]. 2.1. Fixed speed wind generator

Initially, the wind generators were operating at fixed

speed. Regardless of wind speed, the rotor speed of the generator is fixed and determined by the frequency of the network, by the speed multiplier ratio and by the generator type. They are designed to achieve maximum efficiency for a given wind speed. To increase power production, some wind generators were equipped with two types of coils: one for low wind speeds (typically 8-pole) and one for medium and high wind speeds (usually 4-6 poles). Obviously these tricks were hindering the plant and led to increase of the overall equipment price.

The asynchronous generator is equipped with condensers to produce reactive power, soft starters for proper start-up and step-up transformer (0, 6 / 20 kV). The soft starter is used because the starting current of the asynchronous machine is very large (about 6 to 8 In), which can lead to voltage variations in a weak network.

Fig. 1 - The diagram of a fixed speed wind generator

2.2. Variable speed double fed induction wind generator

This machine has the stator directly connected to the

grid and the rotor connected to the grid through a bidirectional converter AC / DC / AC (figure 2). Dual fed means that the stator voltage comes from the network

114


ISSN 2067-5534 © 2012 JSE

and the rotor voltage is induced by the converter. This allows operation over a relatively wide range of variable speed.

Fig. 2 - The diagram of a double fed induction wind

generator 2.3. Synchronous wind generator

It contains a synchronous generator and a

bidirectional frequency converter AC / DC / AC, through which it can connect to the network. The synchronous generator is more expensive and mechanically complex than the double fed induction generator. It is also used in applications where all nominal power produced crosses the converter, which is smaller than the one at double fed asynchronous generators. Synchronous generator has the great advantage that it doesn’t need a magnetizing reactive current from the grid. The magnetic field can be created by permanent magnets (figure 3) or classical excitation coil – with wound rotor.

Fig. 3 - The diagram of a synchronous wind generator

Synchronous generator is more expensive at 1.6

times the double fed induction generator. Main advantages of variable speed wind generators

in relation to the fixed speed are:

- Improved dynamic behaviour which results in reduced mechanical application of the shaft and reduces power fluctuations;

- Better quality of electricity; - Increasing the power produced.

Disadvantages compared with fixed speed wind generators are: - More complicated electrical system; - Use of multiple components; - Losses of power in the incorporating power

electronics items; - High cost.

3. MODELLING OF WIND POWER PLANTS 3.1. The dynamic modeling of wind power plants (WPP) in Eurostag

The Eurostag standard model library contains seven

different models of wind turbines [4]: 1. Model of wind turbine coupled with an

asynchronous generator. This model represents a fixed speed wind turbine. The turbine is connected to a squirrel-cage asynchronous machine.

2. Model of wind turbine with pitch control coupled with an asynchronous generator. This model represents a fixed speed wind turbine. The turbine is connected to a squirrel-cage asynchronous machine whose speed is controlled with pitch control.

3. Model of wind turbine equipped with a

doubly-fed induction generator (DFIG). This model represents a variable speed wind turbine. The turbine is connected to an asynchronous generator whose stator is directly connected to the grid while rotor is connected to the grid via power electronics. The DC link of such power electronics is not represented. The machine disconnects and reconnects to the grid in case of fault.

4. Model of wind turbine equipped with a

doubly-fed induction generator (+ crow bar +chopper). This model is based on the DFIG model mentioned here above. The DC link is here represented with a crowbar and a chopper. This allows the machine to remain connected to the grid in case of fault.

5. Model of variable-speed wind turbine with

direct driven synchronous generator. This model represents a variable speed wind turbine. The turbine is connected to a synchronous generator with permanent magnets excitation. The generator is connected to the grid through a full load frequency converter.

6. Universal model of wind farm. This model represents a wind power plant (of doubly-fed induction generators or direct driven

115


ISSN 2067-5534 © 2012 JSE

synchronous generators). Global injection and dynamic behaviour are considered. It is suitable for power system studies of a grid where wind farms are connected or for wind farm connection studies. There is no single representation of a wind turbine and it does not detail the internal grid of the wind farm.

7. Universal model of wind farm with power/frequency control. This model is based on the Wind farm model mentioned here above. The farm is here equipped with a power/frequency control for primary reserve participation.

3.1.1. Example using the double fed induction generator (DFIG) model

We consider the following simple system (figure 4)

consisting of: three generators (one equivalent synchronous machine (PN=55000MW) at node NHV3, a classical generator at node NGEN and one double fed induction generator DFIG), 3 substations (380/150 kV, 380kV/24kV and 150/0.69/0.69 kV), a consumer at node NLOAD and three electrical lines (380kV).[3]

The double fed induction generator (DFIG) driven by wind (wind turbine) is represented by an induction machine (WT_STAT) with the rotor connected to a converter (WT_GSC).

WT_STAT is a DFIG connected to node NST and represents an aggregated model of 10 wind turbines of each 2MW.

Fig. 4 - The analyzed system for the double fed induction generator model (DFIG)

The behavior of the DFIG WT_STAT is controlled

by three dynamic models. The structure and the links between the macroblocks of the DFIG model are shown in the diagram hereunder.

Fig. 5 - Structure and the links between the

macroblocks of the DFIG model

The macroblock WINDTURB calculates the mechanical torque CM and the optimal rotor speed based on the actual wind speed @VENT. The reference rotor speed NREF is transmitted to the macroblock REGDFIG.

The macroblock REGDFIG calculates the rotor voltages U2R and U2I to control the rotor currents and calculates the rotor active power R2.

The macroblock RECONNE manages the operation of the induction machine while the stator is disconnected in case of disturbance on the network. The rotor voltage regulation is then carried out in the macroblock RECONNE and the calculated values are used in REGDFIG. The stator opening and reclosing is managed by an automaton (DFIG stator protection).

The macroblock INTERRO controls the converter WT_GSC and models the grid side converter. The value of the active power generated by the rotor P2 is transmitted to INTERRO via a measurement block. So WT_STAT and WT_GSC are coupled.

In case of a voltage drop, due to a short-circuit for example, an automatic device (DFIG stator protection) is able to disconnect temporarily the stator from the network, and to reconnect it when the situation has become normal again. While the stator is disconnected from the network, the functioning of the DFIG is managed by the macroblock RECONNE. The rotor voltage regulation is made in RECONNE and the calculated values are used in REGDFIG.

Recently Transmission System Operators have imposed more restrictive grid connection conditions to wind farm in Europe. Wind turbines are now required to remain connected to the grid during voltage dips of certain depths and durations. Also active power production has to be restored as soon as the grid voltage drop is allowed, provided the wind turbine reconnects as soon as the voltage is restored. In the case that the voltage drops below the thresholds set by the TSO or the voltage is not restored within the specified time, the wind turbine will disconnect from the grid, reduce its speed (or

116


ISSN 2067-5534 © 2012 JSE

even initiate a complete stop) and a slower reconnection will occur.

The default values for the stator disconnection thresholds are: -terminal voltage below: -0.75 p.u. for at least 0.08s, -0.85 p.u. for at least 0.4s, -0.9 p.u. for at least 60s. -the output current above: 2 p.u. for at least 0.3s.

There were considered the following events to be simulated: -from t=20s to t=30s: an increase of the wind speed; -when t=50s: short-circuit at 50% of the line NHV1-NHV2-1; -when t=100s: decrease by 50% of the load at NLOAD; -when t=120s: shut down of the wind turbine; -when t=140s: start-up of the wind turbine. The machine is connected to the network and its speed is set to the “rotor speed for synchronisation” specified for the DFIG in the dynamic file(.dta). The pitch angle is set to BETADEM specified by the macroblock parameter and the wind speed is set to VENTDEM. BETADEM is a pitch angle value for which the output power is zero for a wind speed of VENTDEM. -when t=160s: enabling the pitch angle regulator and the wind turbine begins to inject power into the network.

0 20 40 60 80 100 120 140 160 180 200

0.9

1.0

1.1

1.2

s

[sim5] MACHINE : WT_STAT SPEED Unit : p.u.[sim5] 1

a) Speed of the machine WT_STAT

0 20 40 60 80 100 120 140 160 180 200

0.4

0.6

0.8

1.0

s

[sim5] VOLTAGE AT NODE : NST Unit : p.u.

b) Voltage at node NST

0 20 40 60 80 100 120 140 160 180 200

-0

2

4

6

8

10

12

s

MW

[sim5] MACHINE : WT_GSC ACTIVE POWER Unit : MW[sim5] MACHINE : WT_STAT ACTIVE POWER Unit : MW

c) Active power at WT_GSC and WT_STAT

0 20 40 60 80 100 120 140 160 180 200

-0

2

4

6

8

10

12

14

s

[sim5] P_STAT+P_GSC

d) Sum of Active power at WT_GSC and

WT_STAT

Figure 6 - The simulation results 3.2. The dynamic modelling of wind power plants (WPP) in PSS/E

The program PSS/E 32 (Power System Simulator for

Engineers), supplied by Siemens, has four standard models for wind power plants (WPP) [6], [7]:

- Type 1: Wind turbine with directly grid connected induction generator with fixed rotor resistance (typically squirrel-cage);

- Type 2: Wind turbine with directly connected induction generator with variable rotor resistance;

- Type 3: Wind turbine with double-fed asynchronous generators (directly connected stator and rotor connected through power converter);

- Type 4: Wind turbines connected fully through a power converter.

Table 1 presents the main models in PSS/E 32 library for wind power plants.

Table 1. The main modules for wind power plants in PSS/E 32 Generic model

WT1 (type 1)

WT2 (type 2)

WT3 (type 3)

WT4 (type 4)

Generator WT1G WT2G WT3G WT4G Controller WT2E WT3E WT4E Turbine WT12T WT12T WT3T Pitch system

WT3P

Pseudo governor

WT12A WT12A

We will continue with transient stability case studies

for type 3 of wind generators.

3.2.1. Example for type 3 wind generator (DFIG)

The figure below illustrates the connection between the four modules in the type 3 of wind generator [5].

117


ISSN 2067-5534 © 2012 JSE

Fig. 7 - The interaction between the four modules of

type 3 wind generator

The most important module is the one that shapes the electrical part of the generator (Figure 8). In this module, parameter VARFLG can cause constant reactive power control (VARFLG = 0), reactive power control (VARFLG = 1) and power factor control (VARFLG = -1). The parameter VLTFLG is operating the control voltage at the terminal (if VLTFLG is different from 0).

Fig. 8 - Electric control module (WT3E)

We consider the following simple system (figure 9)

consisting of: two generators (one classical and one double fed asynchronous wind generator), two substations (154/34 kV and 34/0.6 kV), a consumer and three electrical lines. The active power of the wind power plant is considered the equivalent of 67 turbines of 1.5 MW. Parameter VARFLG was considered equal to -1 (control power factor)[7].

Fig. 9 - The analyzed system for type 3 of wind

generator It is considered a fault (three phase short circuit) on line 100-101. Short circuit is removed in stage I of the basic protection (Zl = 150 ms). The results are shown in the figures below.

a) Voltage at node 101 and 301

b) Active and reactive power at node 301

118


ISSN 2067-5534 © 2012 JSE

c) Rotor speed at node 301

d) Pitch control system speed at node 301

e)The exitation current at the classical generator Fig. 10 - The simulation results

4. CONCLUSIONS

This article presented theoretical issues about transient stability and wind power plants generators modelling. It described the types of wind generators used today: double fed induction generator and full scale convertor. It presented a few simple examples in programs developed by different vendors (Eurostag 4.5 and PSS/E 32) and the transient stability was studied, when applying an external short-circuit fault in the system.

It is necessary to compare these models with models developed by generic manufacturers of wind generators, to correct the dynamic behaviour associated to the wind power plant in the models used in transient stability analysis.

Eurostag gives the facility to model a large variety of wind power plants types and in detail. REFERENCES [1]. Eremia, M. – Dinamica sistemelor electroenergetice,

Editura PrinTech, Bucure�ti, 2005 [2]. Oprea, S. – Aspecte privind accesul deschis la reţelele

electrice – Integrarea surselor regenerabile de energie – Teză de doctorat, Bucureşti, 2009

[3]. Tractebel Engineering GDF Suez & RTE – Eurostag Tutorial, Brussels, 2010

[4]. Tractebel Engineering GDF Suez & RTE – User’s Manual Part 1, Eurostag Standard Model Library, Brussels, 2010

[5]. SIEMENS –PSS/E Model Library 32.0.5, Schenectady, New York, 2010

[6]. Ruhle, O. – Wind Park Modelling, Simulation, Connection and Integration into Transmission Systems, Bucharest, 2009

[7]. Gençoğlu, C. –Dynamical Modelling of Type 3&4 Wind Turbines with PSS/E, Istanbul, 2011

[8]. Kundur, P. – Power System Stability and Control, McGraw – Hill, Inc., New York, 1994

[9]. Pavella, M., Murthy, P.G. – Transient Stability of Power Systems. Theory and Practice. John Wiley & Sons, New York, 1994

119


ISSN 2067-5534 © 2012 JSE

NITROGEN COMPOUNDS REMOVAL FROM MUNICIPAL WASTE WATER

CRET P., LOLEA M., HODISAN S.

University of Oradea, Universitatii no.1, Oradea 410087, Romania, [email protected]

Abstract. The paper is a short review on nitrogen compounds removal, as ammonia, nitrate, nitrite, from municipal waste water, containing primary information of nitrogen chemistry, consideration on the physical-chemistry methods of removal and also on the most popular wastewater nitrification systems[1]. Keywords: municipal waste water, nitrification-denitrification treatments. 1. INTRODUCTION

The use of water becomes in time more impractical but the water resources. Human activities, from cities, industry or irrigation systems are discharging effluent which becomes the supply of water for other users. Not only intensive agriculture [2,3], but also the nuclear industry[4] are the main pollution source for the drinking water. The admission levels for nitrate in CE is 50 mg.L-1 (adults) -15 mg.L-1 (children) [3,5,6], high levels of nitrate concentration leading to serious medical problems.

Chemicals as metallic salts or complex organic compounds are increasing frequency in the waterways. They enter the drinking water and eventually end up in the wastewater. Primary and secondary waste treatment processes could be not very effectively in removing these chemicals. For years, dilution and purification of the effluent in the receiving stream was considered acceptable. Because these streams have more pollutant loadings, natural processes are not enough in these days. Often is necessary some more treatment than primary and secondary wastewater. In the last years, some physical, chemical, and biological processes come into light in the wastewater technology. Currently wastewater technologies which have been used for advanced wastewater treatment are the following: filtration, adsorption, chemical oxidation, reverse osmosis, nitrate removal by denitrification, phosphorus removal.

Ammonia removal could be done by the biological nitrification of wastewater and after that the denitrification, which involves the conversion of nitrate nitrogen ions (NO3

-) to gaseous nitrogen (N), as shown on the left side of nitrogen cycle (figure 1).

Fig. 1 - Nitrogen cycle[7]

2. EXPERIMENTAL AND DISCUSSION

The removal of ammonia from wastewater treatment becomes a a very important operation in special for the lands where agriculture is done intensively. One method of ammonia removal is the biological nitrification of wastewaters, a process where ammonia is converted to nitrate using aerobic autotrophic bacteria in the treatment process. The process of nitrification is a two-step process for removing ammonia from wastewater and this is done by the utilization of two types of autotrophic bacteria that oxidize ammonia to nitrite (nitrosomonas) and then oxidize nitrite to nitrate (nitrobacter). Biological nitrification systems are projected to convert the entire amount of ammonia into nitrate[1,7].

The two types of autotrophic bacteria need proper biomass concentrations, in a specific environmental conditions (temperature, pH, alkalinity, etc.), enough time for the treatment process, and an increased amount of air, more that requires, for the treatment of biochemical oxygen demand only. A different factor that should be take in consideration in projecting of wastewater treatment plants, that assure biological nitrification is the low alkalinity. Adding sodium hydroxide or other chemicals in order to increase the alkalinity may be needed.

The treatment processes which are recomended for biological nitrification at wastewater treatment plants are :

conventional activated sludge system(figure 2) extended aeration treatment systems(figure 3).

120


ISSN 2067-5534 © 2012 JSE

sequencing batch reactor(figure 4) fixed film (figure 5) membrane bioreactor(figure 6) lagoon systems (figure 7)

The conventional activated treatment process (figure 2) has the advantage that the proven treatment process is able of treating many types of wastewater and is easier to operate, comparative with other treatment processes.

Fig. 2 - Wastewater conventional system[1].

Conventional activated sludge treatment processes

that were projected for biochemical oxygen demand removal only could be changed to assure biological nitrification, too. Constructing conventional activated sludge treatment processes have the disadvantage of being are very expensive. Aeration basins and clarifiers are usually built of concrete and demand expensive mechanical equipment (blowers, pumps, clarifier mechanisms, etc). Conventional treatment processes are also more sensizitive to bulking sludge from filamentous organisms.

Extended aeration treatment processes (figure 3) are similar to conventional activated sludge treatment processes and involve: aeration basins, clarifiers, return activated sludge, and waste activated sludge processes. The most important difference is the longer hydraulic and solids residence times in the process. The hydraulic residence time is typically around 24 hours and the sludge residence time is over 20 days at design flow rates and organic loadings [1]. Having sufficient air, the nitrification will take place faster in extended aeration processes.

Fig. 3 - Conversion of ammonia in a diffused aeration system [1].

Because of these longer times for hydraulic and solids residence, the extended aeration treatment processes can assure the better quality effluent for any kind of wastewater. Extended aeration processes are easier to operate than conventional activated sludge treatment process. With a sufficient amount of oxygen, extended aeration treatment processes can assure raised levels of biological nitrification. The main disadvantage to extended aeration treatment processes consist in the dimension of the facilities that are required to guarantee longer times for the hydraulic and solids residence. The cost of building these types of processes is bigger because the aeration basins and clarifiers are mainly made of concrete and because of the costs of the mechanical equipment.

Sequencing batch reactors (figure 4) are using extended aeration activated sludge treatment process, the difference being that the aeration and clarification processes are taking place in the same reactor basin, having the next steps: fill, react/aeration, settle, and decant. Wasting usually occurs during the react/aeration step[1]. Having sufficient air, the hydraulic and solids residence times could be changed in such a manner to activate the nitrification in the reactor basin.

Fig. 4 - Photo of sequencing batch reactors[8].

The most important advantage of the treatment

process is the lower dimension of the treatment system. Using the combination of the aeration and clarification steps into one basin, the processes can be controlled, measuring the time for each step to assure the required quality of the treated effluent. This type of treatment process, to be performance requires professional operations personnel, with long practice in working and maintenance of these devices. The majority of the municipal systems also need multiple reactor basins and equalization tanks.

Biochemical oxygen demand removal and biological nitrification could be done by fixed film treatment[1] process: trickling filter/activated sludge treatment process, rotating biological contactors, or moving bed bioreactors. In the place of the microorganisms that treat the wastewater suspended in the liquid, the microorganisms are placed to fixed media and treat the wastewater as it flows through the reactor. The trickling filter/activated sludge treatment process also includes plastic media for the microorganisms to develop on packed inside a tower where wastewater is used for treatment. The trickling filter is followed by a conventional activated sludge process. Fans, blowers, clarifiers and pumps are

121


ISSN 2067-5534 © 2012 JSE

necessary. Rotating biological contactors are made of a series of closely packed plastic circular disks that are partially submerged and rotated through the wastewater to be treated. Microorganisms develop on the disks and aeration is accomplished as the disks are exposed to the air during rotation.

Fig. 5 - Fixed film treatment system.[9].

Moving bed bioreactors include plastic media that is immersed in the wastewater in a separate basin with screens to keep the media in the basin. They are built in the manner of conventional aeration basins for biochemical oxygen demand removal and are projected specifically for biological nitrification only. Utilization of the trickling filters/activated sludge treatment process have the advantage of both processes. Trickling filters are more energy efficient and the activated sludge process holds off scaling material from creating lower effluent quality. These types of systems can also reduce the footprint necessary for the conventional activated sludge treatment processes. The main disadvantages to fixed film treatment technologies are the increased high solids retention time requirements, pumping energy required, the potential for rotten egg odors, and the potential for snails and filter flies. The moving bed bioreactors treatment process also requires higher levels of dissolved oxygen.(up to 7 mg/L)[1].

The membrane bioreactor treatment process[1] have three mainly components: 1) anoxic basins, 2) pre-aeration basins, and 3) the membrane bioreactor basins. Rough wastewater have to be screened through a fine screen prior to the anoxic basin. From the anoxic basin, mixed liquor goes into the pre-aeration basins and then into the membrane bioreactor basins. The membranes are placed in the membrane bioreactor basins where wastewater is passed through the membranes and permeate pumps deliver the effluent to the disinfection process prior to discharge[1]. The membranes eliminate the necessity for secondary clarification, Pumping is required as in the same way as in conventional treatment processes. The membrane bioreactor treatment process produces a high quality of the effluent, without any additional operations, will assure both biological nitrification but also nitrogen removal, fitting into a less area, but there are outgoing costs connected to the operation devices and also with purchasing replacement membranes.

Fig. 6 - Membrane bioreactor treatment system[10].

The costs of operation and maintenance of these

systems are higher because they need more power and more operator attention. Lagoon treatment systems [1] are not projected to provide more than biochemical oxygen demand and total suspended solids removal. Biological nitrification takes place, having enough long hydraulic and solids residence times, the proper temperature and sufficient oxygen. Hydraulic residence times have to be extended to at least five to seven days in the aeration process, higher temperatures must be maintained, and enough oxygen have to be assured. A mixed liquor recycle system could be involved to keep a high enough biomass to promote the growth of nitrifying bacteria.

Fig. 7 - Lagoon treatment system [11]

Lagoon treatment systems have the primary advantage

of having building low costs and they are easier to operate and maintain comparative to mechanical wastewater treatment systems. Basins are built mainly by excavation , very little concrete is required. In this case, expensive mechanical equipment (pumps, clarifier mechanisms) are not required. Unfortunately, it is harder to control the parameters that influence effluent quality such as wastewater temperature, wasting, return rate, and oxygen levels in lagoon treatment systems [1].

The effluent quality may fluctuate, that means there are needed facilities projected to be be more versatile in design and operation. The facilities use very large areas, these kind of system is recommended only for small treatment systems.

After ammonia was converted in nitrate/nitrite, the denitrification treatment will be applied following the cycle below:

122


ISSN 2067-5534 © 2012 JSE

Fig. 8 - Nitrification-Denitrification Cycle [7].

Biological denitrification [7] is realized in anaerobic

conditions by heterotrophic bacteria that use nitrate during the fermentation of organic carbon materials. Contrary to nitrification, in which only Nitrosomonas and Nitrobacter bacteria are necessary, a relatively large range of bacteria could done the denitrification. These include Pseudomonas, Micrococcus, Achromobacter, and Bacillus. These groups accomplish nitrate reduction through the process of nitrate dissimulation[1]. In nitrate dissimulation, nitrate or nitrite replaces oxygen in the respiratory processes of the organism under anoxic conditions. Due to the ability of these organisms to “eat” either the oxygen bound in nitrate or free oxygen, these organisms are named facultative heterotrophic bacteria. Denitrifiers are able of an assimilation process where nitrate/nitrite is converted to ammonia. Ammonia is then utilized for the bacterial cell’s nitrogen requirements. If ammonia is already present, assimilation of nitrate need not occur to satisfy cell requirements. Electrons pass from the carbon source (the electron donor) to nitrate or nitrite (the electron acceptor) to promote the conversion to nitrogen gas. This involves the nitrifiers’ “electron transport system” and releases energy from the carbon source for use in organism growth. This electron transport system is similar to that used for respiration by organisms oxidizing organic matter aerobically, except for one enzyme. Because of this close relationship, many facultative bacteria can shift between using oxygen or nitrate (or nitrite) rapidly and without difficulty[7].

In wastewater treatment, organic carbon is the pollutant to be removed, and oxygen must be added. In denitrification, it is nitrate that is removed, and a carbon source must be available. If an insufficient amount of organic carbon is available for denitrification, sufficient carbon (such as methanol) is added to accomplish the nitrate removal.

For the nitrates removal from the potable water, various methods have been proposed: chemical reduction [12,13],physical–chemical processes [14-16], biological

methods [17,18], electrochemical reduction of nitrates on different electrodes: Pt [19,20], Pd[21, 22], Cu[23-28], Ag[27], Ni [29,30], Rh[31], Sn[32], Pb [33], binary alloys [34], CuSn [35] CuZn [36], PdRh1.5/Ti [37], metallic electrodes modified by upd deposition [38-41] or opd deposition [20,26,42–35].

Transforming nitrate/nitrate in nitrogen gas, could also be a challenge using electrochemical treatment methods, based on nitrogen cicle, but taking in consideration the reactions with electron transfer.

3. CONCLUSIONS

Removal of nitrogen compounds from municipal wastewater, but also resulting mainly from agriculture, animal farms or nuclear industry is following the scheme :

ammonia→nitrate→nitrite→nitrogen gas

As it could be seen there are many types of conventional activated sludge treatment processes for the nitrification process as: complete-mix, plug flow, and step feed treatment. They all have the same basic layout of an aeration basin and secondary clarifier with return and waste activated sludge pumps. The conversion of ammonia occurs in the aeration basins. Because the duration of the required time for nitrification, more than for biochemical oxygen demand removal, there are not recomanded high-rate and contact stabilization activated sludge treatment processes. All the presented method has advantages or disadvantages, and one option is not the best solution for all systems and a good consulting engineer can assist with evaluating all of the options before recommending the best solution for each system. The gaseous product is primarily nitrogen gas, but some nitrous oxide or nitric oxide may also result during denitrification. REFERENCES:

[1]. www.waterworld.com [2]. The Nitrate Directive of European Council

91/676/EHS. [3]. N.F. Gray, DrinkingWater Quality: Problems and

Solutions, John Wiley and Sons Ltd., Chichester, 1994, p. 21.

[4]. J.O’M. Bockris, J. Kim, J. Appl. Electrochem. 27 (1997) 623.

[5]. EEC Council Recomandations, 1987. [6]. WHO, Guidelines for Drinking Water Quality, vol. 2

(and Addendum) (World Health Organisation, 1996, 1998).

[7]. cursuri pdf-www.waterworldce.com [8]. www.esi.info [9]. www.aquatechsys.com [10]. www.constructionweekonline.com [11]. www. wmfhillinc.com [12]. F. Gauthard, F. Epron, J. Barbier, J. Catal. 220 (2003)

182. [13]. K. Inazu, M. Kitahara, K. Aika, Catal. Today 93–95

(2004) 263. [14]. L. Panyor, C. Fabiani, Desalination 104 (1996) 165. [15]. K.N. Mani, J. Membr. Sci. 38 (1991) 117.

123


ISSN 2067-5534 © 2012 JSE

[16]. J.J. Shoeman, A. Steyn, Desalination 155 (2003) 15. [17]. V. Mateju, S. Cizinska, J. Krejei, T. Ianoch,

EnzymeMicrobiol. Technol. 14 (1992) 170. [18]. C.-P. Huang, H.-W.Wang, P.-C. Chiu,Wat. Res. 32

(1998) 2257. [19]. G. Horanyi, E.M. Rizmayer, J. Electroanal. Chem. 188

(1985) 265. [20]. O.A. Petrii si, T.Ya. Safonova, J. Electroanal. Chem.

331 (1992) 897. [21]. J.F.E. Gootzen, L. Lefferts, J.A.R. Veen, Appl. Catal.

188 (1999) 127. [22]. G.E. Dima, A.C.A. Vooysde, M.T.M. Keper, J.

Electroanal. Chem. 554 (2003) 15. [23]. G.E. Badea, T. Badea, Sci. Technol. Environ. Prot. 5

(1998) 42. [24]. G.E. Badea, T. Badea, U.P.B. Sci, Ser. B: Chem.

Mater. Sci. 63 (2001) 493. [25]. K. Bouzek,M. Paidar, A. Sadilkova, H. Bergmann, J.

Appl. Electrochem. 31 (2001) 1185. [26]. M. Paidar, I. Rouˇsar, K. Bouzek, J. Appl.

Electrochem. 29 (1999) 611. [27]. S. Cattarin, J. Appl. Electrochem. 22 (1992) 1077. [28]. G.E.Badea, Electrochimica Acta , 54(2009), 996-

1001. [29]. Li Hu Lin, D. Robertson, J. Chambers, D. Hobbs, J.

Electrochem. Soc. 135 (1998) 1154. [30]. J.O’M. Bockris, J. Kim, J. Electrochem. Soc. 143

(1997) 3801. [31]. Ph. M. Tucker, M.J.Waite, B.E. Hayden, J. Appl.

Electrochem. 34 (2004) 781.

[32]. I. Katsounaros, D. Ipsakis, C. Polatides, G. Kyriacou, Electrochim. Acta 52 (2006) 1329.

[33]. J.D. Genders, D. Hartsough, D.T. Hobbs, J. Appl. Electrochem. 26 (1996) 1.

[34]. T.Ya. Safonova, O.A. Petrii, J. Electroanal. Chem. 448 (1998) 211.

[35]. C. Polatides, G. Kyriacou, J. Appl. Electrochem. 35 (2005) 421.

[36]. Z. Mácová, K. Bouzek, J. Appl. Electrochem. 35 (2005) 1203.

[37]. H. Cheng, K. Scott, P.A. Christensen, J. Appl. Electrochem. 35 (2005) 551.

[38]. Ma Ling, Li Hu Lin, Cai Cheng-Liang, Electrochim. Acta 38 (1993) 2773.

[39]. H. Huang, M. Zhao, X. Xing, I.T. Bae, D.A. Scherson, J. Electroanal. Chem. 293 (1990) 179.

[40]. G.E. Badea, T. Badea, Rev. Roum. Chim. 48 (2003) 843.

[41]. G.E.Badea, T.Badea, Rev.Roum.Chim., 48(11), pg.843-848, 2003.

[42]. M.Wasberg, G. Horanyi, Electrochim. Acta 40 (1995) 615.

[43]. M.Wasberg, G. Horanyi, Electrochim. Acta 42 (1997) 261.

[44]. G.E. Badea, T. Badea, U.P.B. Sci. Ser. B: Chem. Mater. Sci. 65 (2003) 11.

[45]. K. Shimizu, R. Goto, S. Piao, R. Kayama, K. Nakata, Y. Yoshinga, J. Electroanal.

[46]. Chem. 601 (2007) 161.

124


ISSN 2067-5534 © 2012 JSE

AN INNOVATIVE MODEL FOR HIGH ACCURACY WIND

POWER PREDICTION

ALBU R.D., POPENŢIU-VLĂDICESCU F., University of Oradea, Universităţii no.1, Oradea,

[email protected]

Abstract: This study is another version of the work described in our paper accepted at the PSAM11 & ESREL 2012 international conference where a non-linear ensemble system is used to develop a new model for predicting wind speed in short-term time scale. Regardless of the recent advancements in the re-search of prediction models, it was observed that different models have different capabilities and also no single model is appropriate under all situations. The idea behind EPS (ensemble prediction systems) is to take advantage of the features of each subsystem to confine diverse patterns that exist in the dataset. The low prediction errors demonstrate the increased prediction accuracy. Keywords: ensemble, wind-power, short-term, prediction, forecast, recurrent neural networks, wind energy. 1. INTRODUCTION

Wind power prediction is of vast importance for the safety and stabilization of grids (Moyano et al. 2009). The most intricate problem now is to improve the prediction accuracy. BP (Backward Propagation) neural networks have been used extensively in wind power prediction but results have shown slow convergence rate (Gaofeng et al. 2008).

In the technical literature, we can find two major approaches to forecast wind power:

1. Physical methods: Require many physical considerations to gain the best prediction precision. The input variables will be physical or meteorology information. They present advantages in long-term prediction

2. Statistical methods: Aspire at finding a relationship between the on-line measured power data. They will use the historical data of the wind farm. They present advantages in short-term prediction.

Ensembles are prediction techniques used to produce

a representative sample of a dynamic system possible future state (Raj Kiran et al. 2011). Sometimes the EPS may use different models for different members, or different formulations of a model.

The multiple simulations show two main sources of uncertainty in prediction models

Errors introduced by chaos or sensitive dependence on the initial conditions.

Errors introduced because of imperfections in the model.

This article is structured as follows: we focus on the

proposed EPS and its architecture. The next section contains numerical results from a real world case study, particularly our EPS prediction results. We tested the proposed EPS using data sets collected from the ANM (National Meteorology Administration) website. In the last section, we raise some interesting conclusions. 2. THE PROPOSED ENSEMBLE PREDICTION SYSTEM

The time scales we use in this short-term prediction

solution are in the order of some days for the forecast horizon and from minutes to hours for the time-step. For the purpose of time series prediction, an ensemble can be considered to be a general nonlinear mapping between a subset of the past time series and the future time series values. The proposed EPS is presented in Fig. 1. In the following sections we will briefly describe each prediction model. For more details about their architecture and function see the referred papers.

Figure 1. The EPS

2.1. The RQVM The RQVM (Recurrent Quadratic Volterra Model) is

similar to the one proposed by Duehee Lee in 2011. His paper shows a way to use the recurrent quadratic Volterra system to predict the wind power. The RQVM is a second-order polynomial equation that uses the output data as feedback recursively, after passing a time-delay

125


ISSN 2067-5534 © 2012 JSE

filter. The Volterra system is extracted from the weights of the Recurrent Neural Network. In order to build Volterra kernels from the combination of weights, the activation function is approximated to the high order polynomial function using the Lagrangian interpolation. The memory of the Volterra system is identified using the PACF Partial Autocorrelation Function. The Volterra system can analyze order and memory and it captures the output power patterns that can be used for short-term prediction. For simplicity, it is assumed that the system is causal and homogeneous. The truncated recurrent Volterra system V (P, M) with the finite order of P and the finite memory of M is defined as.

x (n) = h0 +

M

k

M

kppp

M

k

M

k

M

k

p

teknxknxkkh

knxknxkkh

knxkh

1 111

211 1

212

1111

1

1 2

1

)()()...(),...,(...

)()(),(

)()(

n = M+1, M+2, ... (1) Where h0 is the constant, which is zero in this case, and hi (k1, ... ki) is the set of Pth order Volterra kernel coefficients. Consecutively, to reduce the computational complexity, the kernels should be assumed as: hp (k1, ... kp) = 0 if k1 > ... > kp (2)

The kernels hi are assumed to be a symmetric function with respect to all permutations of the indices ki, . . . ,kP. As a result, one kernel per each permutation is enough to describe the Volterra system. The other kernels become zero. The input signal is Xn-1 = {x (n − 1), x (n − 2), . . . ,x (n − M)}, and the output signal is the x(n). All signals and kernels are real numbers. Since it is difficult to categorize errors in the truncated Volterra model as natural system errors or higher order term errors, the author assumed higher order terms are absorbed into the Volterra system errors. The number of Volterra kernels of each order can be calculated by using the Combination with a Repetition. The equation is as follows:

)!(!

!

MPPM

(3)

The order is limited to the second order, and the memory is limited to 20. The RNN in this model has three layers: one input layer, one hidden layer, and one output layer. The input layer receives as many input neurons as the number of memory, and input neurons are in the tap-delayed form. One hidden layer has an arbitrary number of neurons which is heuristically decided by the number of Volterra kernels. Hidden layers receive”net input” which is the sum of the input values that are multiplied by their corresponding weights. Hidden neurons use the

tangent hyperbolic function tanh as the activation function. The output layer has only one neuron and uses a linear activation function. 2.2. ANN Prediction Model Based on MTS Algorithm

As the second model in our ensemble system we used a model analogous with the one described by Shuang Han et al. 2011. They have put forward a wind power prediction model build with a BP neural network optimized by Tabu search algorithm with memory function. The basic principle of ANN (Artificial Neural Network) based on MTS (Multiple Tabu Search) algorithm is to optimize neural network’s connection weights using TS algorithm which has memory function. Achieve the global optimal solution using the global search capability of TS algorithm and thus avoid getting into local minimal.

Suppose the error function of some BP network is: f = f (Wh, Wo, Өh, Өo) where:

Wh, are connection weights between input layer and hidden layer.

Wo, are connection weights between hidden layer and output layer.

Өh, is the threshold value of hidden layer neurons. Өo is the output layer neurons.

The optimization for network is the process of solving min (f (Wh, Wo, Өh, Өo)). For the expediency of representation, symbol Δ is used to denote vector (Wh, Wo, Өh, Өo).

The following is the procedure of optimizing neural network with TS algorithm:

1. Initialize Δ; endow every component of Δ with a little random number denoted Δinitial.

2. Δbest denotes the optimal solution and Δnow denotes the current solution. Δbest = Δinitial = Δnow.

The vector Δnow is storied in the Tabu table. 3. Produce a neighbourhood solution Δnew of Δinitial

and calculate f (Δnew) and f (Δbest). 4. If f (Δnew) has not varied for many times

continuously n stop the algorithm and output the result, else go on with the next step.

5. If f (Δnew) < f (Δbest) then Δbest = Δnew. The vector Δnew enters Tabu table and the memory point in table backward in turn. If f (Δnew) >= f (Δbest) we need to judge if Δnew is within some memory point’s given neighbourhood. If it is a neighbourhood solution vector Δnew is reconstructed. If it isn’t then Δnow = Δnew and update the Tabu table at the same time.

6. Produce a neighbourhood solution Δnew of Δnow

and go to step (4). 7. The optimized weight vectors and threshold

value vectors are obtained when the training finished.

126


ISSN 2067-5534 © 2012 JSE

The inputs of the BP network are: wind speed, wind direction, air temperature, air pressure, relative humidity. Less neurons in the hidden layer result in worse prediction precision. On the contrary, over fitting may occur with too many hidden neurons and this result in poor prediction precision and long training time. The common method is to adjust it in training process. The optimal number is corresponding to the minimal error. 2.3 Artificial neural network–Markov chain model

In the model proposed by S.A. Pourmousavi Kani et al. 2011, artificial neural network (ANN) and Markov chain (MC) are used to develop a new ANN–MC model for predicting wind speed in very short-term time scale. In this study, the short-term patterns in wind speed data are captured by ANN and the long-term patterns are considered utilizing the MC approach and four neighbourhood indices. The proposed model consists of two ANNs. The first one, ANN-1, is used for short-term wind speed prediction. ANN-1 is a multi-layered perceptron (MLP) that consists of one input layer, one hidden layer, and one output layer that has only one neuron. This step is called primary prediction by the author. This network has 10 inputs fed with actual wind speeds corresponding to times t to t-10. The training data consists of 30 sets with 10 measured wind speeds in each set. After the primary prediction, TPs (transition probabilities) for predicted values, other four indices and primary prediction outputs are fed as input variables to a second ANN (ANN-2). Finally, the constructed model is used for different time horizon predictions. Regarding the Markov process, the probability of the given condition in the given moment may be deduced from information about the previous conditions. The order of the chain gives the number of time steps in the past influencing the probability distribution of the present state, which can be greater than one. Many natural processes are considered as Markov processes. Actually, the TPM (transition probabilities matrix) is a tool for describing the MC activities. Each element of the matrix represents probability of moving from a specific condition to a next state. TPM is formed by 600 preceding wind speed data and the calculated matrix is used for primary predicted values. Initially they calculate the Markov state for primary predicted values, the outputs of ANN-1, for one step ahead. Then, according to TPM, the probability of predicted value in the next step is calculated. This process is executed for all primary predictions. For longer prediction horizon, transition probabilities for two or three steps ahead are necessary. In these cases, the above TPM is multiplied according to the number of time steps in the future. It is observed that there is a logical relation between the states for predicted values in comparison with the states of actual values.

The relations between the primary prediction results and the coefficients obtained from MC are difficult to be

established. Since ANNs can encode complex and non-linear associations, the ANN-2 is used to discover the relations between the primary prediction values and the obtained probabilities. The inputs of the ANN-2 are: the transition probability of the predicted values’ state, transition probabilities toward two next states (FNIs) and transition probabilities toward two backward states (BNIs). The main purpose of ANN-2 is to achieve higher accuracy of prediction in contrast with primary predicted values. Since the ANN-2 has six input variables and one output variable, number of neurons in each layer should be in the range of variables. The best structure for the ANN-2 with the least MAPE (Mean absolute percentage error) is determined by the author as 3, 0 and 1 neurons for input, hidden and output layers, respectively, with 10 training vectors and the learning rate of 0.01–0.05. 2.4 The arbitrator

Figure 2. The arbitrator

The ensemble uses the output obtained from the

individual constituents as inputs to it and the data is processed according to the design of the arbitrator. As the arbitrator in our approach we use the following RNN shown in (Fig. 2).

The RNN has two context layers: the Elman context layer and the Jordan context layer, both with some differences from the original Elman and Jordan recurrent neural networks. The Elman context layer differs from the original Elman RNN because the two context neurons obtain inputs from the output of the hidden layer after a delay of one time unit. In Elman context layer the number of neurons must match the number of neurons in hidden layer. A Jordna RNN has a number of neuron in context layer that matches the number of neurons in output layer. Another difference is that in a Jordan network the output is used to feed the context layer through a non-weighted connection and the context layer is going to feed to the hidden layer just as in Elman networks.

127


ISSN 2067-5534 © 2012 JSE

3. PREDICTION RESULTS

We tested the proposed EPS using data sets collected from the ANM (National Meteorology Administration) website.

a)

b)

c)

Figure 3. Prediction results a) a winter day b) a summer day c) a fallen day

We randomly selected three days, one from each

season, and predict the wind speed at each hour for these days. Our EPS results are illustrated in (Fig. 3). The solid line represents the measured wind speed and the dotted line indicates the predicted values. The wind speeds were measured in Tulcea and collected from the ANM web site. To predict W (d, h), the wind speed at hour h of day d, we train the RNN with only last two values: W (d-1, h), W (d-2, h); it has been reported that every data point in a time series is only strongly dependent on the immediate past two values. The training is complete when we provide as inputs all wind speed values, for a number of n epochs. One epoch is finished when the entire training set is exposed to the RNN. The number of epochs is the number of steps of the training process, it is a dynamic value; we set it high and let it stop according to the validation set. The initial learning rate is 0.001, results in good coarse training quickly. For better performance, we

used a schedule of 0.0005 for two epochs, followed by 0.0002 for the next three, 0.0001 for the next three, 0.00005 for the next four, and 0.00001 thereafter. The learning rate is decreased by 79.4% of its value after every epoch. In order to implant fixed points into recurrent systems, the backpropagation technique is used. In fixed-point learning, the first action is the forward propagation of the activations. This procedure repeated for a certain number of times will induce the relaxation period. This has to be repeated until the network attains its own dynamic. After the net become stable, an error can be computed at the output. Then, the error is propagated backwards through the network. The error at each output can be multiplied by the relaxed activation for updating the weights. We have to select the relaxation time both in the forward and backpropagation phases.

In (Fig. 4) is presented a comparison between the prediction accuracy of the ensemble subsystems and the EPS, at different time-horizons. As we can see in the, we obtained very good prediction results proven by a very low average error rate. The secret is the joint usage of Neuro Solutions features and our innovative EPS architecture.

The prediction error of a model is classically defined as the difference between the measured and the predicted value. A horizon dependent model error e (t + k|t) is given by:

e (t + k|t) = v (t+k) – vp (t+k!t) (4) Where v (t + k) is the measured wind speed at time t

+ k, and vp (t+k!t) is the wind speed predicted for time t + k and computed at time t. The evaluation criterion we used is the MAPE defined as:

MAPE (k) =

N

tX

ktvtkte

N 1

)100)(

)|((

1 (5)

Also, k and N represent the prediction horizon and

number of prediction respectively. In this study for each hour we have done 30 predictions, so N=30.

128


47 ISSN 2067-5534 © 2012 JSE

Figure 4. MAPE at different time-horizons for each ensemble subsystem

Table 1. Best prediction errors and MAPE

Hour

Winter day 22.01.2011

Summer day 22.06.2011

Fallen day 22.10.2011

0 +0.01 / 1.35 -0.01/ 1.2 +0.01/ 1.12 1 -0.23 / 11.5 +0.4 / 403 +0.4 /403 2 -0.02 / 1.06 -0.13 / 13.13 +0.02 / 2.2 3 +0.25 / 12.5 -0.003 / 0.35 +0.1 / 98 4 +0.79 / 39.5 +0.17 / 167 +0.02 / 0.2 5 -0.65 / 21.6 -0.35 / 305 +0.02 / 0.22 6 -0.24 / 12 -0.15 / 147 +0.01 /1.55 7 +0.06 / 2.96 -0.13 / 132 +0.1 / 107 8 -0.02 / 0.66 +0.1 / 9.69 +0.2 / 203 9 +0.34 / 17 -0.3 / 15 -0.01 / 0.55 10 -0.19 / 6.33 +0.002 / 0.1 -0.01 /0.55 11 +0.01 / 0.33 +0.3 / 31 +0.5 / 508 12 +0.19 / 6.35 -0.32 / 11.28 -0.6 /22 13 +0.12 / 4,04 +0.85 / 42.6 +0.7 / 36 14 -0.02 / 0.77 +0.32 / 18 +0.4 / 19.9 15 -0.01 / 0.33 -0.01 / 0.52 -0.1 / 5.6 16 +0.13 / 4.43 +0.42 / 43 +0.65 /66 17 +0.8 /28.65 -0.001 / 0.06 +0.18 /9 18 +0.03 / 0.75 +0.62 / 32 +0.21 / 10.52 19 -0.1 / 3.35 +0.79 / 80.5 +0.32 / 33 20 -0.21 / 5.52 +0.001 / 0.05 -0.03 / 1.53 21 +0.05 / 1.66 -0.23 / 23.2 +0.2 /20.1 22 -0.21 / 7.22 +0.03 / 1.5 +0.16 / 8 23 +0.01 / 0.39 +0.34 / 339 +0.21 /230

129


ISSN 2067-5534 © 2012 JSE

Table 2. Average MAPE for each ensemble subsystem

4. CONCLUSIONS AND FUTURE WORK

A repeatedly scenario appears when new prediction results are presented: the new model is argued to have higher predictive accuracy than do other prediction models. This situation shares the issue that predictive accuracies are being calculated and compared in different test sets. The prediction model may have inherent accuracy, but the accuracy as measured will vary across test sets. This makes it impossible to define the accuracy of a prediction model independent of the test set to which it is applied. Basically, no real-world prediction model can foresee perfectly. There are four major reasons for this. 1. The prediction model may not have one or more

required variables: we simply do not know everything necessary to predict with perfect accuracy.

2. Measurement error can reduce accuracy: a poor data collection phase introduces noise in the data set and damage predictive accuracy.

3. The prediction model may not fit as well as it could: a predictor might be forced to have a linear outcome when a nonlinear result would have fit the data better and would permit more precise predictions.

4. The effective sample size may be unsatisfactory to approximate the prediction model coefficients as accurately as possible.

Consequently, the new prediction models must be compared to existing models on the same data sets if we want to judge whether progress has been made or not. ACKNOWLEDGEMENTS This work was partially supported by the strategic grant POSDRU/CPP107/DMI1.5/S/80272, Project ID80272 (2010), co-financed by the European Social Fund-

Investing in People, within the Sectorial Operational Program Human Resources Development 2007-2013. REFERENCES [1]. Henrik Madsena, Răzvan-Daniel Albu, Ioan Felea, Florin

Popenţiu-Vlădicescu, (2012) “A New Ensemble Model for Short Term Wind Power Prediction”, accepted at PSAM11 & ESREL 2012 international conference.

[2]. Barbounis T. and Theocharis J., (2007) Locally recurrent neural networks for wind speed prediction using spatial correlation, Information Sciences, vol. 177, no. 24, pp. 5775 – 5797.

[3]. Costa A., Crespo A., Navarro N., Lizcano G., Madsen H. and Feitosa E., (2008) A review on the young history of the wind power short-term prediction, Renewable and Sustainable Energy Reviews, vol. 12, no. 6, pp. 1725-1744, June 2008.

[4]. Duehee L., (2011), Short-Term Prediction of Wind Farm Output Using the Recurrent Quadratic Volterra Model, IEEE Trans. Reliability, vol. 58, no. 1, pp. 106–117.

[5]. Gaofeng F., Weisheng W., Chun L., Huizhu D. (2008) Wind power prediction based on Artificial Neural Network, Proceedings of the CSEE, vol. 28, no. 34, pp. 118-123, Dec. 2008.

[6]. Gebraeel N., & Pan, J., (2009), Prognostic degradation models for computing and updating residual life distributions in a time-varying environment, IEEE Trans. Reliability, vol. 57, no. 4, pp. 539–550.

[7]. Montgomery D. C. and Murat K. (2008), Introduction to Time Series Analysis and Forecasting, D. J. B., N. A. C., and G. M. F., Eds. Wiley series in probability and statistics.

[8]. Moyano C, F., Pecas L. J. A. (2009) An optimization approach for wind turbine commitment and dispatch in a wind park. Electric Power Syst Res, 79(1):71–9.

[9]. Negnevitsky M., Mandal P., and Srivastava A., (2009) An overview of forecasting problems and techniques in power systems, in Power Energy Society General Meeting, PES ’09. IEEE, 2009, pp. 1 –4.

Model Average MAPE Winter day

Average MAPE Summer day

Average MAPE Fallen day

RQVM 8.73 76.18 74.57 ANN-MTS 8.91 75.92 74.98 ANN-MC 8.12 76.09 74.68 EPS 7.92 75.64 74.46

130


ISSN 2067-5534 © 2012 JSE

INFLUENCE FACTORS ON ENERGY INTENSITY AGRICULTURAL PRODUCTS

Rotari V.

State Agrarian University of Moldova [email protected]

Abstract - Energy intensity of final products regardless of their destination becomes a more significant indicator in production. This indicator takes on significance a special role to produce agricultural products. Given article is devoted to the determination and group decision makers influencing the creation and formation energy intensity of agricultural products.

Keywords - Energy intensity, influence factors, agricultural products

1. INTRODUCTION

Since the Republic of Moldova is a country where agriculture is one of the basic branches creating the national budget and energy resources are imported almost 96%, then cheltuielelor energy analysis to final agricultural production is a very current problem.

The notion of energy intensity of the agricultural product is meant the ratio between the gross and final product cost energy spent in producing this product. It can be concluded that the energy intensity of agricultural products is a function determined, which depends on a number of random factors

Determination of energy intensity in production of agricultural products and minimizing their ability to create competition in agricultural products on the market for selling both internal and external. 2. DEPLOYMENT PROBLEM

Develop a methodology for determining the energy

efficiency of enterprises production of agricultural products and minimize the energy intensity

Increasing the energy efficiency of enterprises producing agricultural products (reducing energy intensity of final product) requires each enterprise manager to develop and respect the decisions necessary for the rational use of energy resources.

At the first stage of making the analysis of energy consumption in order to increase energy efficiency necessary to perform auditory energy.

Auditory energy needed to be conducted by experts external to the auditory to be quite objective.:

a) the auditory making it necessary to include the following steps) determining and diversion of energy flows based on the technological processes;

b) the composition scheme of use of electricity, which includes the total consumption of energy and their differentiation based on the technological processes;

c) analyzing the energy balance in the company; d) determining the energy share in total financial enterprise;

d) determining places where technology can be reduced energy consumption;

e) determining technology, which can give the effect of lowering energy consumption.

Analysis and differentiation of energy sources used which includes as components>

a) determining the composition of energy sources in relation to the amount of energy used in the enterprise;

b) determining the energy losses; c) analyze energy usage charts (chart by day, season, year); d) analyze the probability of reducing the energy

consumption and using renewable energy sources, which may include:

- Saving energy sources in heating technology processes;

- Economy (reducing energy consumption and increase power factor) of electricity by reactive power compensation;

- Energy saving lighting sources; - Optimizing the use of equipment technological

regimes technological processes. - Conduct energy management Energy management is about a system to manage

energy consumption, based on typical performance measurement and verification of results obtained with the standard, which can be:

- Energy management of any company is a management tool with the consumption of energy sources, which makes it possible to determine and specify the quality of energy consumption from that undertaking.

Simultaneously power management allows to perform a behavior of the energy efficiency of every enterprise, compared with similar companies and technical measurements and technological development to increase the effectiveness of electricity (reducing energy intensity of final product).

In the process of making the enterprise energy management, particular attention will be given to the series of issues related to energy consumption as follows:

- Composition of subunits Charter energy component of the undertaking;

- Accumulation of static data on energy consumption; - Endowment plan composition control equipment

and control energy consumption;

131


ISSN 2067-5534 © 2012 JSE

- Analytical calculation of the effectiveness of preventive use of energy resources;

- Drafting measurements with destination Reduce energy intensity;

- Assessment and measurement of destination location to reduce energy intensity of final products;

- Calculation of energy cost in prime cost final product of the enterprise;

In making the measurements mentioned above often may need to perform some calculations that consider operational energy consumption and compare the specific consumption and the total for a period of time to appreciate the changes that occurred in technological processes. 3. ASSESSING OPPORTUNITIES FOR SAVING ENERGY IN THE COMPANY

Power saving mode to any enterprise in manufacturing is a function carrying a determined character and is a feature that bears a character is determined by several factors and indefinite.

This phenomenon depends on how the best use of technological equipment used. All technological processes in production plants and improvement of wine can be derived:

- Direct processing technology of wine products; - Additional technologies that are specifically

intended to provide direct technology. For any industrial enterprise the power savings and

lower energy intensity in the center of the final process of undertaking. Thorough analysis of the problems of electricity consumption and saving mode makes it possible to solve these problems and allow to reduce the energy intensity of final product. In the process of saving electricity, it is impossible to using disconnection processes.

Each position must be reached based on the following conditions:

- Volume production technology should be maximized;

- The right choice of technological equipment, to achieve technologies that have increased energy intensity;

- Low level of losses in power systems and power distribution;

- Installation of reactive power compensation; - Analysis of electric drives and motors replace

electric motors that have less power and operating mode can be replaced with a similar scheme but optimal. Necessary expenses in this case the economy will recover electricity used.

Equivalent replacement of electric motors is recommended if his load is not more than 50%.

In the replacement of electric motors is recommended to take into account the particular construction of electric motors. Thus the basic recommendations of the energy savings from use of electric drives are - Appropriate choice of electric motors based on the task will be running the machine press work;

- Electrical drives routing schemes needed to be automated under load and time;

- If the machine work has several operating modes is necessary that electric motors have speeds frequency variation.

Another significant consumer electricity to primary processors of grapes is lighting. Lighting to deviate in the following types:

- Production lighting, lighting technology, lighting communal street lighting.

Total electricity consumption for lighting is about (12-16)% of total electricity consumption.

For these conditions energy saving lighting processes can be quite significant.

To reduce electricity consumption in lighting technology companies are recommended contemporary (modern) as follows:

- Use lamps contemporary lighting increased flow; - Use of equipment and fixtures - Use zoning schemes in various rooms lighting

technology; - Use ballast resistors with a low level of loss of

electricity consumption; - Use of automatic process control systems for

lighting, which is back for one year. It is worth mentioning that modern systems used in lighting technology can reduce energy consumption from 20 to 30% of the total. Primary processing enterprises grapes can divert energy used in the following way. 4. CONCLUSION

The analysis of statistical material (from several areas of the Republic of Moldova where vines are grown) and technical, we can state that were argued and determined energy costs destined for planting and processing grapes as required by current technology for each year until the first harvest technology.

REFERENCES

[1]. Guide of energy sources consumer, Chişinau, Lumina;

2001; 137 p. [2]. Metodologia determinării preşurilor la produsele petroliere.

ANRE; Energia, 2007; nr.5 p. 8-18. [3]. Regulament pentru furnizarea şi utilizarea energiei

electrice. ANRE; Energia,2005; nr. 6; p.10-33. [4]. Metodologia determinării, aprobării şi aplicării tarifelor la

producerea energiei electrice , energiei termice şi apei de adaos.; ANRE; Energia, 2004;nr. 1; p. 8 -25.

[5].***Metodologia determinării, aprobării şi aplicării tarifelor la energia electrică livrată de întreprinderile de distribuţie „ RED NORD” şi „ RED NORD-VEST” S.A; ANRE; Energia, 2004; nr. 2 p. 8-25.

[6]. ***Instrucţia pentru determinarea consumului specific (perderilor tehnice) a energiei electrice în reţelele electrice. ANRE; Energia, 2005; nr. 5 p. 8-18.

[7]. ***Regulament privind racordarea mini – centralelor electrice la reţeaua electrică şi funcţionarea lor în regim autonom sau în paralel cu sistemul electroenergetic. ANRE; Energia, 2005; nr. 2; p. 11-16.

[8]. ***Metodologia determinării, aprobării şi aplicării tarifelor pentru serviciile de transport şi distribuţie a energiei electrice. ANRE; Energia, 2004; nr. 1; p. 13 -21.

132


ISSN 2067-5534 © 2012 JSE

SIMULATION AND ENERGY EFFICIENCY EVALUATION OF A LOW-ENERGY BUILDING

VLAD. G.E., IONESCU C., NECULA H.

*University POLITEHNICA of Bucharest, 313 Spl. Independen�ei, Bucharest [email protected]

Abstract – The paper presents the results of the simulations made for the one of the low energy houses built in the campus of University Politehnica of Bucharest. The aim of the simulations was to emphasize the role of the solutions applied in order to achieve “the passive house” standard. Some of these solutions can be easily adopted by old buildings. They refer to a special architecture, materials with improved thermal properties, systems requiring renewable sources of energy and heat recovery of the exhausted technologies. Keywords: thermal load, heat recovery, TRNSYS simulation, passive house. 1. INTRODUCTION

The residential sector is an important energy consumer in all the countries. For this reason people became more interested in finding solutions to reduce both thermal and electrical consumptions of the buildings. In many countries the energy efficiency of buildings is promoted by implementing national programs. After the energy crises in the 70s, due to the price of the fuels and the depletion of energy sources, the low energy building became a standard for the new buildings. There are many solutions which can be applied to both new and existing buildings. A special architecture, a certain orientation, a low ratio of surface to volume are adopted by new buildings [1]. Materials with low thermal resistance and low permeability are used by both types of buildings. There are a lot of studies that showed the impact of the envelope materials on the building energy consumption [2]. Technologies based on renewable sources of energy, capable to maintain the parameters characteristic to buildings environment at desired value, are among these options.

The employment of the renewable energy for space heating leads on one hand to the reduction of primary energy consumption and on the other hand to the reduction of the greenhouse gases emissions.

In the European Union countries, space heating is responsible for more than 50% from the total energy consumption and about 10% are provided by renewable energy sources [3]. In the USA and Asia the situation concerning the energy consumption of buildings is similar [4,5]. In the 90s, an institute from Darmstadt, Germany developed a new concept called “passive house”. The passive house is a building created for the

central European climate with certain standards referring to the energy consumption: 15 kWhm-2year-1 thermal energy consumption and 120 kWhm-2year-1 total energy consumption [6]. There are also other requirements referring to the overall heat transfer coefficients of walls, windows, to the number of air changes per hour. Since the first passive house built in 1991, more than 27000 houses were built in Europe [7].

In order to improve the energy performance of the buildings, in May 2010 the European Directive on Energy Performance of Buildings was adopted [8]. The EU Member States agreed to apply a “Net Zero Energy Program”. A general definition says that a net zero-energy building (ZEB) is a building that over a year does not use more energy than it generates. But this is not the only definition of this building. There are four definitions for the ZEB depending on different points of view: the aim of the building project, the intentions of the investors, the energy expenses and the impact on the environment [9]. A complete definition should emphasize the importance of using renewable sources of energy but also the energy efficiency of buildings. According to this directive, by December 31st 2020, all new buildings must be nearly zero-energy buildings and after December 31st 2018, new buildings occupied and owned by public authorities must be nearly ZEB.

The newest building concept is “the positive energy house”. The positive energy house is a building that generates energy more than its consumption. The users and the management of this building have a very important influence in achieving the standard.

In Romania, the majority of buildings have a mean thermal energy consumption of 300 kW.m-2.year-1. This value is double compared to the values specific for European countries buildings. The economical potential to improve the energy efficiency was estimated to 30%-50% for residential field and 13-19% for public field [3].

In developed countries, economical and environmental reasons stimulated the low energy buildings construction. Taking into account these premises, a new project was developed by the University “Politehnica” of Bucharest. The aim of this project was to promote the energy efficiency of buildings in Romania. 2. PROJECT DESCRIPTION

The building comprising two houses is located in the

campus of the University Politehnica of Bucharest and it

133


ISSN 2067-5534 © 2012 JSE

has a total surface of approximately 2x140 m2. Each house includes a hall, a living room, a kitchen, a small bathroom and a technical room on the ground floor and two bedrooms, two bathrooms, an office and a hall on the first floor. To take advantage of the solar energy, the building has a special architecture and a southern orientation (Fig. 1). The building has very large windows on the southern side and small windows on the north, used only for lighting.

(a)

(b)

Fig. 1 - Politehnica Houses: (a) southern side; (b) northern side

One of the houses has an eastern orientation (called

“East House”) while the other has a western orientation (called “West House”). The study presented in this paper was carried out for the East House. In order to realize an optimization of the space, the living and the kitchen have a southern orientation, and the bedrooms are oriented to east. Each bedroom and the office have only one eastern window. Materials such as mineral wool and polystyrene, which are among the best insulation materials, were used to insulate the envelope of the house. Due to these materials with very good properties (table 1), the heat losses are extremely low compared with those of a standard house.

Table 1. Properties of the insulation materials

Wall type Insulation material Thickness

[mm]

Thermal conductivity [W.m-1.K-1]

Roof mineral wool 400 0.034

Outside walls mineral wool 300 0.034

mineral wool 150 0.034 Floor slab

XPS polystyrene 180 0.04

The house has triple glazing windows with glass covered with a low E layer and bio cleaner, very low heat transfer coefficient and a very low solar heat gain factor (U=0.7 W.m-2.K-1, g=0.5). The space between the glass layers is filled with argon (90%) and air (10%).

The Politehnica House is very well insulated and has a very tight envelope. In this case, due to the high level of insulation, there is no air leakage and the outside air does not enter the house uncontrollably. A mechanical ventilation heat recovery system (MVHR) is absolutely necessary to supply the fresh air and to remove the exhausted air.

To reduce energy consumption for heating, the air is preheated in two steps (Fig. 2): first in the ground heat exchanger and second in the heat recovery unit which can save even more than 91% from the energy of exhausted air.

Fig. 2 - Scheme of the EAHX-MVHR system The role of the earth to air heat exchanger is to preheat

the required ventilation air in the heating season and to cool it in the summer. The fresh air is drawn from outside and then flows through the ground heat exchanger.

In the cooling season, the condensate is discharged through a tower. The condensation tower is placed at the lowest point of the system and is fitted with an airtight cover at ground level to prevent false air infiltration into the ventilation system.

The material chosen for the geothermal collector is a high density polypropylene. This is called Awaduct Thermo and is provided by Rehau. In order to prevent the bacteria growth on the inner surface of the collector, very small particles of silver are incorporated during a special process. As it is known, the silver particles have an antimicrobial effect.

The chosen pipe has an outer diameter of 200 mm and a wall thickness of 7.8 mm. The tube has an enhanced thermal conductivity (0.28 W.m-1.K-1) to ensure a better heat transfer between the air and the ground.

After passing through the geothermal collector, buried at 2 m below ground surface, the fresh air enters the heat recovery unit. A volume flow (165 m3.h-1) is drawn inside the geothermal system by a fan. This air flow assures an air change rate of about 0.5 h-1 that guarantees a proper ventilation of the house.

In order to find the optimum value of the EAHX length, a thermo-economic criterion was applied. The algorithm is presented in detail in [10]. An optimal pipe length of 38 m corresponds to the highest thermal efficiency of the EAHX system.

The MVHR unit installed in the experimental house consist of two fans and an air to air heat exchanger. One

EAHX

fresh air inlet

preheated air

air extracted from building

exhausted air MVHR unit

134


ISSN 2067-5534 © 2012 JSE

fan supplies new air to the living, bedroom and working rooms and in similar way the other one extracts the polluted air from kitchen and bathrooms. The two air streams flow in a cross-current direction inside the plate heat exchanger. After the MVHR unit, the fresh air temperature increases due to the energy received from the exhausted air. To reach the desired temperature, the air is electrically heated after the MVHR unit. The fresh air and exhausted air volume flows are equal to prevent the discomfort generated by pressure differences.

2. TRNSYS PROJECT OF POLITEHNICA HOUSE

The first step in accomplishing the simulations was to divide the building into 10 thermal zones (Fig. 3). As it can be seen, a thermal zone is represented by one room or a group of rooms. Each thermal zone has a certain volume and a certain thermal air capacity.

(a)

(b)

Fig. 3 - Thermal zones of the house (ground floor/ first floor)

1) living + kitchen + bathroom; 2) technical room 1; 3) entrance hall; 4) bathroom 1; 5) bedroom 1; 6) office; 7) bedroom 2; 8) bathroom 2+ technical room 2; 9) hall;

10) space between roof and ceiling The simulations were carried out with TRNSYS

software that is a complete and extensible environment for the transient simulation of systems including buildings. The simulation of the entire house requires a higher number of parameters. Every thermal zone consisting in one room or a group of rooms has a certain air volume. A minimum value of 20 ºC for the inside air temperature is set for the entire year. Due to the very

good tightness of the house envelope, the air leakage/infiltration is considered zero.

3. SIMULATIONS RESULTS

The first simulation was carried out with a simple flux ventilation system. The fresh ventilation air was the outside air temperature. The first step (Fig. 4 - case 1) was to add walls with half of the actual insulation thickness and regular windows with high overall heat transfer coefficient and high solar heat gain (U=2.8 W.m-

2.K-1, g=0.76). In order to find out the influence of the insulation on the energy consumption of the building, the envelope insulation was changed to different values. As expected, the energy consumption decreases with insulation thickness increasing. In our simulation, the thermal load decrease is about 7% for insulation thickness increased with 50%.

Fig. 4 - The influence of the insulation thickness on the thermal load

1) 50% of insulation thickness; 2) 75% of insulation thickness; 3) 100% of insulation thickness

The value of the heat flux, P (W.m-2), corresponding

to the case of the real insulation thickness (100%) is about 31.1 Wm-2 (Fig. 5).

Fig. 5 - The influence of the insulation on the heat flux Another important design factor is the type of the

window chosen for the house. During the simulations, the geometrical characteristics were kept constant and only the thermal properties of windows were changed. For this purpose, three types of windows existing on the market were considered. The third type of the window is the one installed in the house. As it can be seen in Fig. 6, the

135


ISSN 2067-5534 © 2012 JSE

thermal properties of the windows have a significant impact on the thermal energy consumption. Compared to the first case, the third one has the energy consumption lower with more than 60%.

Fig. 6 - The influence of the windows properties on the thermal load

1) U=2.8 W.m-2.K-1, g=0.76; 2) U=1.4 W.m-2.K-1, g=0.59; 3) U=0.7 W.m-2.K-1, g=0.5

The influence of the type of the windows installed is

also reflected in the heat flux lost to outside (Fig. 7).

Fig. 7 - The influence of the windows properties on the heat flux

1) U=2.8 W.m-2.K-1, g=0.76; 2) U=1.4 W.m-2.K-1, g=0.59; 3) U=0.7 W.m-2.K-1, g=0.5

After installing the high efficiency windows, the thermal load of the experimental house is about 36 kWh.m-2

.year-1. This value can be further reduced. All the new types of buildings adopt technologies

such as the air-to-air heat recovery. Both Politehnica houses have implemented double-flux mechanical ventilation system based on heat recovery (MVHR). The next step in performing the simulations was to add the MVHR unit. The most important part of the MVHR unit is the air to air heat exchanger. The heat exchanger is made of corrugated plastic plates that separate the two air streams (fresh air / exhausted air).

In order to obtain more energy savings and also to meet the hygiene and comfort conditions, the MVHR unit is connected to an earth to air heat exchanger (EAHX). The fresh air is taken from the outside of the building and is introduced inside the EAHX where it is preheated in the winter and cooled in the summer. After this, the air is further heated in the MVHR unit.

After the introduction of the MVHR unit, a significant reduction of the thermal load can be observed (first case in Fig. 8). The decreasing of energy consumption is even more important after coupling MVHR unit to the EAHX.

Fig. 8 - The influence of the EAHX and MVHR unit on the thermal load

According to the results of the simulation presented

in Fig. 8 the thermal load of studied house is about 11.4 kWh.m-2.year-1, due to the ventilation system which comprises the geothermal component (EAHX) and the heat recovery unit. The value obtained is in accordance with the one of the requirements of the passive house standard, according to which the energy demand for space heating must not exceed 15 kWh.m-2.year-1 [4].

The combination made between MVHR unit and EAHX system has also an important influence in reduction of the heat flux. Fig. 9 shows a reduction of over 20% after coupling EAHX to the MVHR unit, compared with MVHR unit operation only.

Fig. 9 - The influence of the EAHX and MVHR unit on the heat flux

As the cumulative thermal load shows (Fig. 10), the

heating season starts in November and ends in April. Fig. 10 also illustrates thermal demand variation (in W.m-2 with a time step of 1 h) with peak load around 8.5 W.m-2 achieved in December and January at the lowest temperatures.

The house was built for a family of four occupants. The rate of the heat gain from the occupants is about 100W. Schedules were created for the most important thermal zones (living, bedroom 1, bedroom 2, office).

136


ISSN 2067-5534 © 2012 JSE

Fig. 10 - Variation of thermal load and cumulative thermal load

Two different schedules for living occupancy are

represented in Fig. 11: one for the week days (a) and the second one for the weekend (b).

(a)

(b)

Fig. 11 - Living schedule (a) week days; (b) weekend

There are eight different schedules for the four

thermal zones. The heat rate of the household appliances, the lighting system and other equipment are taken into consideration. Table 2 contains the values of these heat rates [11].

Table 2. Heat rates

Schedules were also created for the electric

equipment from table 2 excluding the refrigerator assumed to run all day.

In Fig. 12 and 13, the influence of the heat rates produced by humans and all other electrical devices on the thermal loads and heat flux are presented.

Fig. 12 - The influence of the internal heat rates on the thermal loads

1) human heat rates, 2) all types of internal heat rates

Fig. 13 - The influence of the internal heat rates on the heat flux

1) human heat rates, 2) all types of internal heat rates During winter, the electrical heater placed after the

MVHR is turned on. In order to achieve the control of this auxiliary heater, a thermostat is added to the scheme of the system and it is connected to the house.

In order to assure the thermal comfort in all rooms of the house, the fresh air temperature after the electric heater is settled to 30 ºC.

Comparing the power consumption of the house used as laboratory to the house inhabited by a family of four persons, there is a noticeable difference (Fig. 14).

Fig. 14 - Electrical energy consumption of the house

Thermal zone

Type Heat rate

[kJ/h]

TV 540

household appliances 720 living+ kitchen

refrigerator 60

office PC 140

137


ISSN 2067-5534 © 2012 JSE

1) house used as a laboratory; 2) house used as a home for a four persons family

The lower energy consumption of the inhabited house is due to the internal heat gains provided by the electric equipment and humans. 4. CONCLUSION

Inside the campus of University Politehnica of

Bucharest, two low energy houses which are intended for passive house certification were built. To reduce the primary energy consumption for heating and cooling the ventilation air, one of these houses is equipped with an earth to air heat exchanger (EAHX) coupled with a double flux mechanical ventilation heat recovery system (MVHR).

The orientation plays an important role in the heating process of the house. During the winter when the sun is lower in the sky, this allows the building to capture the free heat coming from the sun. In the summer, when the sun is higher on the sky, the building is able to reject the solar heat. The walls and windows orientation of each thermal zone was set from the beginning and it was not changed during the simulations.

The simulations of the thermal behaviour of the house were carried out step by step and every time a new element was added to the house in order to find out its influence on the thermal load. The simulations show the energy benefits achieved with the insulation materials adopted by Politehnica House.

These simulations also show that the impact of the windows is even greater than the one of the insulation. Not only the heating demand but also the cooling demand can be easily manipulated with the help of the windows properties (one of the properties is the shading factor).

After introducing the EAHX and the MVHR unit, a very important reduction of the annual heating energy consumption was recorded. The obtained value satisfies one of the passive house concept requirements which, theoretically, can qualify the house to obtain the certification.

Acknowledgements

The work has been co-funded by the Sectoral Operational Programme Human Resources Development 2007-2013 of the Romanian Ministry of Labour, Family and Social Protection through the Financial Agreement POSDRU/88/1.5/S/61178. REFERENCES [1]. Su, B., Challenges – Opportunities and Solutions in

Structural Engineering and Construction (Chap. 140 - Energy consumption related to winter housing thermal performance), CRC Press, 2009

[2]. Kim, J.J., J.W., Moon – Impact of insulation on building energy consumption, Eleventh International IBPSA Conference, Glasgow, Scotland, July 27-30, 2009

[3]. Mladin, E.C., Georgescu, M., Dutianu, D. – Eficienţa energiei în clădiri şi acquis-ul comunitar, Masa Rotunda "Eficienta energetica, prioritate nationala si factor de integrare" organizata de ENERO in colaborare cu CNR-CME (Comitetul National Roman pentru Consiliul Mondial al Energiei), Bucharest, August 28th 2003

[4]. U.S. Dpartment of Energy – Energy Efficiency and Renewable Energy, Energy Efficiency Trends in Residential and Commercial Buildings, 2008

[5]. *** http://www.e2b-ei.eu [6]. *** http://www.passiv.de [7]. *** http://www.pass-net.net [8]. *** Directive 2010/31/EU of the European Parliament and

of the Council of 19 May 2010 on the energy performance of buildings (recast), Official Journal of the European Union

[9]. P. Torcellini, S. Pless, M. Deru – Zero Energy Buildings: A critical look at the definition preprint, National Renewable Energy Laboratory, June 2006

[10]. Vlad, G.E., Ionescu, C., Necula, H., Badea, A. – Thermoeconomic design of an earth to air heat exchanger used to preheat ventilation air in low energy buildings, International Conference on Energy, Environment, Entrepreneurship, Innovation, Lanzarote, Spain, 27-29 May 2011

[11]. G. Krauss, B. Lips, J. Virgone, E. Blanco - Modelisation sous TRNSYS d'une maison a energie positive, International Building Performance Simulation Association, IBPSA France, Nov 2006

138

table of content no2 - oradea - descriptionenergy-cie.ro/archives/2012/jse_vol_3_no_2.pdf ·...

Documents