smart industrial concept!

#PIWorld ©2018 OSIsoft, LLC

Smart Industrial Concept! Design- and Operational Optimization

Rene Hofmann Scientific Coordinator SIC!

1


2

Worldwide changing energy system Transformation of the energy supply in industrial systems

?

Pic

ture

Sou

rce:

OS

Isof

t

Challenges − Diversification of energy production − Load flexibilization/sanitization − Volatility of renewable energy sources − Consideration of the energy market

Energy-intensive industry − energy supply system − Optimal plant operation → exploit full potential of industrial plants − Need for flexible designs and predictive automation/control concepts


Pic

ture

Sou

rce:

OS

Isof

t

3

Digital Transformation of the Industrial Energy Supply

Cooperative Doctoral School: SIC! [Smart Industrial Concept!]-

Holistic Approach with Digitalization of Industrial Processes and Applications for 2050 and beyond


4

Optimum design of the energy supply

and process demand

Operational optimization (CHP, P2X, TES, HTHP,

etc.)

Power market generation

decentralized/ volatile Sector

coupling

Data handling and treatment

SIC! in a Nutshell Added value through specific

use of data

Development of methods for energy-optimized operation of industrial plants

Optimum system design for future environment

Consideration of mutual interaction industry ↔ energy networks


5

Research Competences

Power market Sector

coupling

Operational optimization

Design optimization and planning

Data handling and

treatment Process Analysis and Integration

Data Driven Modeling

Mathematical Optimization

Sector Coupling, Power Grids,

Markets

Control Development,

MPC

Thermodynamic System Modeling


Pic

ture

Sou

rce:

OS

Isof

t

6

SIC! [Smart Industrial Concept] https://sic.tuwien.ac.at

Power market sector coupling Operational optimization Design optimization and

planning Data handling and

treatment

PhD#1

PhD#2

PhD#4

PhD#8

PhD#6

PhD#7

PhD#3

PhD#5


Experienced industrial partners…

...supported by scientific excellence 7

SIC! united and well balanced approach

CONSULTANT MARKET

OPERATOR

IMPLEMENTERS

R&D


DEMO

8

Industrial Flexibilization Concept – Storage Integration


9

Operational optimization

PhD#3: Operational optimization concepts with integration of storage systems for load flexibilization in the energy-intensive industry

PhD#5: Development of methods to optimally control the supply of energy-intensive industrial processes by integrating waste heat and using components to increase load flexibility

SG

EBTES

Heat Demand

Energy Price

Energy flow

Information flow

Optimization

SG: Steam Generator EB: Electrode Boiler TES: Thermal Energy Storage


10

Intelligent Design/Operational Optimization

Vorführender

Präsentationsnotizen

Unit Commitment Problem (UC): UC Solution Approach: “Mixed Integer Linear Programming”�Simplified MILP optimization model of a ESU for a given prediction horizon Modeling of components Interlinking to the overall system Input: ESU MILP model, forecasts for heat load, electricity prices, etc. Output: When-Which-How the Components of the ESU operated?


11

Energy Supply Units

Supply Demand

Source: R. Hofmann, S. Dusek, M. Koller, H. Walter: "Flexibilisierungspotenzial für Energieanlagen in der Industrie. Intelligentes Demand-Side-Management durch Integration von thermischen Speichern - Teil 1"; BWK, 68 (2016), 9; 6 – 11.

Simple Example

≠ ?

Vorführender


Simple Model: Steam generator (SG) - e.g. biomass plant Electrode boiler (EB) Thermal Energy Storage (TES) Load - e.g. process steam Result of a fictional scenario: Historical electricity prices Constant biomass price Simulation duration: 1 week


12

SG + TES + EB

Supply Demand


Results

=



13


Results

Energy Market-Prices



14


Results

States (SG +TES + EB)



15

Potential depending on individual Process

Result of a test study: o fictitious energy supply system o historical electricity prices o constant biomass price o simulation time: 1 week

0,00%

7,05%

13,87%

20,69%

nur DE TES EK TES + EK0%

5%

10%

15%

20%

25%Saving Potentials Fuel costs / Electricity

costs in %

SG only TES EB TES+EB

Vorführender


4 Scenarios: SG only TES EB TES+EB


DEMO

16

PI System in context of SIC! - Runtime Model


17

Motivation… PI system beneficial for research activities:

single point of truth

preparation of data for data scientists

cross border enablement

“Unified data from different sources to provide data scientists the same view at the same time step…”

Data handling

and treatment


18

Catalog

Glossary/Thesaurus

Taxonomy

Semantic network

Ontology

Information Models

Text/HTML

XML

RDF

RDFS

OWL Semantic Interoperability

Structural Interoperability

Syntactic Interoperability

high semantics

weak semantics

low complexity high complexity

SIC!

Knowledge Representation: Expressiveness

Vorführender


Overview of the expressiveness of possibilities for knowledge representation; the area that we want to achieve with the SIC! knowledge base is recorded in the polygon. Catalog: collection of terms without any additional information about the terms themselves or relations between them Glossary: collection of terms amended by human readable explanations; Taxonomy: model for assigning terms to classes Semantic network: collection of terms and arbitrary relations between them, without formal semantics Ontology: collection of terms and relations between them, including formal semantics Datenbedeutung – Datenpunkte aus Datenbanken – Verständnis was das für ein Datenpunkt sein kann – Maschine benötigt Semantik mittels Ontologie erfassbar (wie diese in Bezieheung stehen) RDF Resouce descrition Famework (Graph basiete Beschreiben) : RDFS = Beschweibungssprache; WOL (web ontology language)


19

SIC! Ontology

Plant Design

Control Design Model

Design

SIC! Runtime

Data Mining

Data Analysis

Model Transformation

Model Tuning

Knowledge Representation: Design/Runtime

Vorführender


Here it is to be expressed that the knowledge base is designed for the design and the draft, but is then instantiated in the "runtime" system by means of model transformation. After evaluation of the data, findings from the runtime system are to flow back into the knowledge base in order to adapt its design to reality.


Sensors

Millions of Smart Devices

PDC/Edge OSIsoft

Cloud Services Classic PI System

20

SIC! Runtime System

Vorführender


So the future really looks more like this picture. On the left we have our pervasive data collection and edge strategy to capture the data no matter where it comes from. And there’s specific problems to solve there like dealing with limited connectivity, being able to run on different operating systems such as Linux, remote updates & management, and others. In the middle, our classic PI System is still the workhorse at the plant level enabling a single view of data across many different types of automation and control systems and enabling visibility into plant floor operations that enables real time decision making. And on the right, OSIsoft Cloud Services is really where we see the customer’s Enterprise system of the future. It’s the system that all edge and plant systems roll up into. It’s the system where you can manage all the deployed systems across your enterprise. It’s the system where you take advantage of Machine Learning across your enterprise data sets. And it’s the system where scale is taken care of, because it’s leveraging the cloud. Finally, OSIsoft Cloud Services is also where you can tap into a broader operational data ecosystem that connects you with best-in-class analytics and your community of service providers, vendors, and business partners that enable you to leverage external expertise to get more value from your operational data


21

Build / Train Models

OSIsoft PI System PI Data Archive ( time series )

PI Asset Framework Hosting Common Asset

Model ( Assets, Event Frames)

Plant

OPC UA

MPC evon

Other Formats

User User


22

Thermal Energy Storage Model

1D Model +

Model validation with real Lab measurement data

Source: M. Koller, R. Hofmann, Mixed Integer Linear Programming Formulation for Sensible Thermal Energy Storages, Proceedings of the 28th ESCAPE, Graz 2018.

Vorführender


Sensible Thermal Energy Storage Storage Medium: Gravel Heat Transfer Fluid: Air Stores heat as internal energy by temperature increase Charging: Hot air from top to bottom Discharging: Cold air from bottom to top Temperature front Moves along the axis during charging and discharging Temperature gradient decreases during a phase Max. power dependent on temperature spread


23

Validation of the 1D model of the fixed bed regenerator Highly dynamic operation Temperature spread

Source: F. Mayrhuber, H. Walter, M. Hameter, 2017. Experimental and numerical investigation on a fixed bed regenerator. 10th International Conference on Sustainable Energy and Environmental Protection.

Model Experiment

Comparison Experiment - Simulation

Vorführender


Validation of the 1D model of the fixed bed regenerator – (const. charge/discharge temperatures / const. mass flow Highly dynamic operation - Simulation-based with different load/unload times Temperature spread depends on storage fill level and max/ min storage load at begin of charge/discharge phase


24

Implementation of OSIsoft PI System PI Data Archive (time series) PI Asset Framework to the fixed bed regenerator at the TU Wien lab.

Connection via evon – XAMControl® automation system

Next Steps…


25

Physical vs. data driven Analysis of model

formulation (for exact description of storage behavior)

Neuronal network techniques Full understanding of the

highly dynamic operation

Experiment

050

100150200250300350400

0 500 1000 1500

Data-Model

?

Model Comparison


26

Acknowledgements To the numerous contributors of this presentation W. Kastner,

PhD-candidates: S. Panuschka, A. Beck, M. Koller, C. Seykora, B. Pesendorfer.

To AIT-TU Wien partnership with the joint professorship of Industrial Energy Systems

To all partners and supporters of the Cooperative Doctoral School SIC!

To OSI partnership and setup suggestions (Frank Batke)


27

This Presentation is based on Publications o Keynote-Lecture: Beck, A. and Hofmann. R: “Extensions for Multi-Period MINLP Superstructure

Formulation for Integration of Thermal Energy Storages in Industrial Processes”, in Proceedings of the 28th European Symposium on Computer Aided Process Engineering, June 10th to 13th, 2018, Graz, Austria. © 2018 Elsevier B.V. http://dx.doi.org/10.1016/B978-0-444-64235-6.50234-5, pp 1335-1340.

o Koller, M. and Hofmann. R: “Mixed Integer Linear Programming Formulation for Sensible Thermal Energy Storages”, in Proceedings of the 28th European Symposium on Computer Aided Process Engineering, June 10th to 13th, 2018, Graz, Austria. © 2018 Elsevier B.V. http://dx.doi.org/10.1016/B978-0-444-64235-6.50163-7, pp 925-930.

o R. Hofmann, S. Dusek, M. Koller, H. Walter: "Flexibilisierungspotenzial für Energieanlagen in der Industrie. Intelligentes Demand-Side-Management durch Integration von thermischen Speichern - Teil 1"; BWK, 68 (2016), 9; 6 – 11.

o F. Mayrhuber, H. Walter, M. Hameter, 2017. Experimental and numerical investigation on a fixed bed regenerator. 10th International Conf. on Sustainable Energy and Environmental Protection.

http://dx.doi.org/10.1016/B978-0-444-64235-6.50234-5

http://dx.doi.org/10.1016/B978-0-444-64235-6.50234-5

http://dx.doi.org/10.1016/B978-0-444-64235-6.50163-7


SIC! [Smart Industrial Concept]

28

Univ.Prof. Dr. René Hofmann Scientific Coordinator SIC! TU Wien – Institute of Energy Systems and Thermodynamics AIT Austrian Institute of Technology [email protected] [email protected]


Questions?

Please wait for the microphone

State your name & company

Please rate this session in the mobile app!

Search “OSIsoft” in

your app store

29


30

smart industrial concept!

Documents