exclusive insights€¦ · 06-11-2019  · customer engagement resiliency water quality monitoring...

12
NAVIGATING THE SMART WATER JOURNEY: From Leadership To Results EXCLUSIVE INSIGHTS Condition Assessment Smart Water Strategy Workforce Management Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management IT/OT Data Science Digital Water This e-book was inspired by SWAN’s 9th Annual Conference, featuring more than 80 speakers covering the latest applications of digital solutions for water and wastewater systems. We’re excited to expand on many of the top insights shared at the conference to help demonstrate the benefits of digitalization, as well as the pitfalls to watch out for, as you progress along your own smart water journey. A Joint Publication www.wateronline.com | www.swan-forum.com | 2019

Upload: others

Post on 04-Jul-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

NAVIGATINGTHE SMART WATER JOURNEY: FromLeadership To Results

EXCLUSIVE INSIGHTS

Condition Assessment

Smart Water Strategy

Workforce Management

Customer Engagement

Resiliency

Water Quality Monitoring

Workforce Management

Collaborative Innovation

Stormwater

Digital Twin

Asset Management

IT/OT

Data Science

Digital Water

This e-book was inspired by SWAN’s 9th Annual Conference, featuring more than 80 speakers covering the latest applications of digital solutions for water and wastewater systems. We’re excited to expand on many of the top insights shared at the conference to help demonstrate the benefi ts of digitalization, as well as the pitfalls to watch out for, as you progress along your own smart water journey.

A Joint Publication

www.wateronline.com | www.swan-forum.com | 2019

Page 2: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

SMART WATER REPORT 9

FOUNDATIONS FOR BUILDING A DIGITAL TWIN FOR WATER UTILITIES

Authors: Gigi Karmous-Edwards, Pilar Conejos, Kumar Mahinthakumar, Sarah Braman, Pascale Vicat-Blanc, Jaime Barba

1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 1.1. WHAT IS A DIGITAL TWIN? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102. KEY OPPORTUNITIES FOR UTILITIES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10 2.1. DATA-DRIVEN INFORMED DECISION MAKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.2. IMPROVING CUSTOMER EXPERIENCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.3. ORGANIZATIONAL REFORM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .123. BIG-PICTURE ARCHITECTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 3.1. TECHNOLOGICAL INFRASTRUCTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 3.1.1. Cloud and Edge Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 3.1.2. Wireless Communications and 5G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 3.1.3. Middleware and APIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 3.1.4. Cybersecurity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2. DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2.1. GIS (Geographical Information Systems) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2.2. SCADA (Supervisory Control and Data Acquisition) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2.3. AMI (Advanced Metering Infrastructure) – Meter Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2.4. CMMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2.5. IoT Sensor Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 3.2.6. LIMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 3.2.7. Weather Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 3.2.8. Access to Accurate, Normalized Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 3.3. ALGORITHMS AND ANALYTICS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 3.3.1. Models: Hydraulic and Chemical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 3.3.2. Leak Detection Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 3.4. VISUALIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.4.1. Reports and Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.4.2. Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.4.3. Web-based SCADA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164. ACTIONABLE STEPS WITH CASE STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 4.1. DEVELOP A VISION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4.2. ASSESS WHERE YOU ARE TODAY – COMPONENTS AND PROCESS & GOALS . . . . . . . . . 16 4.3. DEVELOP AN INNOVATION CULTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 4.4. BUILD AND CALIBRATE A SIMULATION MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 4.5. BRINGING IT ALL TOGETHER – AN EXAMPLE OF A LIVE DIGITAL TWIN . . . . . . . . . . . . . . . . .185. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18

Page 3: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

10 SMART WATER REPORT

1. Introduction Current impacts of water scarcity and pollution are already signifi cant as evidenced by conditions in the Middle East, India, Africa, China, Latin America, and the U.S. As witnessed in the recent headlines, the sixth-largest city in India, Chennai, has literally run out of water. Most of us agree that water stress is increasing all over the world. Climate change, urban population growth, tightened regulations, aging infrastructure, and water scarcity are some of the many global challenges water utilities are faced with. Utilities are now forced to address them in creative and cost-eff ective ways. We know we have to change the way citizens and governments view, value, and manage our water resources. Deploying innovative technologies appears to be essential and urgent for helping to solve these challenges. What if water planners had access to a complete, up-to-date, holistic view of a water system and actionable, informative dashboards at all times? We believe that a powerful software tool that can provide an accurate estimation and awareness of a community’s resource water (ground, surface, wells, desalination, etc.), water production and distribution, wastewater treatment and recycling, as well as what is happening with potential storms nearby, would help societies better manage our precious resource, water. This is the promise of Digital Twin. (In this document we will refer to Digital Twin as DT.)

1.1. What is a Digital Twin? Today, DT technology is used in all industries, ranging from manufacturing and medicine to transportation and now the water sector. In practice, a DT for a water utility is a combination of modeling software that utilizes data from multiple sources and usually across multiple departments and expertise. It is this end-to-end type of operational and business tool that has the potential to change how utilities are managed today. It will unlock value by enabling improved insights to support better decisions, leading to better outcomes in the physical world.

In this article, we aim at defi ning the term and providing solid steps for water utilities of any size to start working toward achieving the holistic-style digital management of utilities.

A new SWAN Digital Twin H2O Work Group was recently created to help accelerate the water sector’s adoption of DT technology. The group brings together global water leaders and stakeholders from utilities, technology companies, engineering fi rms, government, and academia to help identify and solve relevant utility challenges. The ongoing group will deliver best practices and a road map for developing and maintaining a DT based on agreed-upon objectives. There was a high level of enthusiasm and urgency on this important topic in the May 2019 workshop. The second SWAN Digital Twin H2O

Workshop is now planned for Nov. 6, 2019 at Aquatech Amsterdam.

There are many defi nitions of DT. One that was agreed to during the fi rst SWAN DT H2O Workshop in May is: “A Digital Twin can be defi ned as an actively integrated, accurate digital representation of our physical assets, systems, and treatment processes with a constant stream of data pairing from the physical twin for continuous calibration. It will unlock value by enabling improved insights that support better decisions, leading to better outcomes in the physical world.”

2. Key Opportunities for UtilitiesThere is a great deal of untapped value when data from the diff erent specialized systems are used in concert with more holistic management. DT technology brings all operations and business management together in one system through data and application integration. The increased computational power of cloud computing makes it easier to run the models in continuous mode, and with constant “pairing” of near-real-time operational data from the physical system, the models can be calibrated or validated to behave like the physical system or twin.

DT is a disruptive technology that provides a virtual/digital representation of both the elements and the dynamics of a water plant or system. If implemented properly, DT can infl uence the design, build, and operation of the system throughout its lifecycle (design-build-operate) and help optimize operation through informed insights. In other words, it is a dynamic software model (hydraulic model + machine learning) of the physical plant/system based on real-time continuous calibration.

Since DT technology combines data from systems across the diff erent departments as well as integrates the insights of specialized solutions, it becomes a powerful holistic tool for utility management. This means that a more complex system of data and parameters is combined and algorithmically utilized for

Page 4: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

SMART WATER REPORT 11

never-before-seen insights. Although it sounds dramatic, that is the unique strength of a DT. This does not happen automatically; the DT is developed over time, combining data and analytical algorithms to the core model, where it continues to become a more accurate representation of the physical process.

Some of the major applications for DT once developed and calibrated include:

• Leak detection and localization – combining insights from specialized solutions for detecting leaks with many other incorporated operational parameters (i.e., fl ow anomalies, water quality, maintenance, fi eldwork, etc.) results in more accurate localization

• Water quality – tracking chemical components through the entire network on a near-real-time basis

• Energy savings – provide algorithms for optimized pump schedule for low-cost energy usage while taking into consideration all critical parameters, such as water quality, predictive demand schedules, etc.

• Optimization of pressure and fl ow – ability to detect anomalies when they occur and provide insights to control optimum pressure better

• Asset lifecycle management – provides holistic knowledge about assets, including customized maintenance based on actual conditions, fi eldwork, historical knowledge of the brand, environmental conditions, etc.

Aside from the major applications, here are some other, more generic benefi ts to having DT:

• Drive outcomes from critical business insights

• Provide one comprehensive view

• Virtual sensors

• Reduce maintenance costs and unplanned outages by early alerts — alerted at fi rst signs of a problem

• View the dynamic status of the physical system via an integrated and holistic view

• Reduce the data silos and departmental silos

• Provide predictive analysis to avoid future failures

• Proactive operation instead of reactive

• A near-real-time holistic connection between the physical world and the digital world

• Improve effi ciency and increase the optimization of operations

• Ability to run what-if scenarios at any time

• Leverage the Internet of Things (IoT), hydraulic modeling, and machine learning (ML)

• Maximize ROI of investments of assets and tools and extending lifecycle of systems

• Bridge cross-discipline teams together across

the utility

• Lead to better participation from water utilities into smart cities

2.1. Data-driven Informed Decision MakingFundamentally, digital is about using data to make informed and optimized decisions. From design to the asset lifecycle, the DT is useful in many diff erent ways, starting from optimizing the design phase to integrating with the build phase. After that, the DT then takes on a new role to help maximize asset lifecycle while optimizing process effi ciencies.

On the other hand, the DT is a fundamental element in the daily network operation and maintenance, since it provides extensive and in-depth knowledge of the network behavior. It can be further used for training operators, assisting as a decision-support system under emergency conditions, facilitating the planning of the network, optimizing the network operation, improving the detection of anomalies, and providing a means to become more proactive than reactive.

For example, today most ongoing maintenance work and repairs at utilities are based on a limited and mostly historical set of data. However, it is widely accepted that both maintenance and unforeseen events can be resolved more effi ciently if all asset tools, business tools, and work order systems could be accessed through an integrated enterprise-level command-and-control system. This is essentially the role of a DT, which is capable of providing useful insights for formulating decisions based on the complex set of parameters (diverse data sets) and models refl ecting the behavior of the systems.

2.2. Improving Customer Experience One of the utilities’ top priorities is their customers. Improving overall customer experience involves both transparent engagement and the delivery of cost-effi cient, reliable services. Many of the core components of DTs provide signifi cant customer experience improvements in and of themselves. The Town of Cary, NC, is an excellent example of this. They were an early adopter of automatic metering infrastructure (AMI), installing nearly 60,000 of the smart meters for a townwide implementation in 2011. A corresponding customer portal allows citizens to see their usage on an hourly, daily, weekly, and monthly basis, which is a signifi cant improvement from the monthly billing frequency. Citizens can set their desired threshold at which they would receive a text message or email alert. While this provides signifi cant value to those individuals, the town pursued a more proactive approach by developing an algorithm to identify individual meters whose usage was continuous for a 20-hour period on the prior day and comparing it to the average for prior days and months. This allows the town to prioritize whom to contact when workload permits. The resulting high-usage list

Page 5: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

12 SMART WATER REPORT

is reviewed daily by customer service representatives, who can quickly review the data to detect anomalies and reach out to the citizen directly. Daily, this process results in a few dozen phone calls and a handful of work orders initiated to investigate further or, in certain circumstances, shut off the water. By managing high usage within 12 hours of its occurrence, the total value of adjustments made to monthly bills decreased over $1 million in the fi rst year of implementing this process. The town also provides in-home water audits upon request, wherein a staff member will bring the citizen’s AMI history to discuss with the citizen and investigate potential causes for the high usage. Providing more information and data transparency with customers has resulted in improved customer experience and helped to develop greater trust.

2.3. Organizational ReformDT requires cross-functional teams with diff erent perspectives to come together and collaborate toward solutions. Indeed, taking a holistic view of utility management and operation involves a great deal more than just technology. Leveraging the most benefi t from technological solutions means understanding the role of people, process, and policy within an organization, which some term as digital transformation. Successful digital transformation of water utilities has often involved a digital strategy that includes a three-pronged approach as can be seen in the following graphic. The fi rst prong is the development of a well-thought-out road map (continually evolving as technology changes) of the utility’s digital infrastructure. The second prong involves a utilitywide eff ort by all staff members to revisit each process, both front and back offi ce processes, and look for ways to become more effi cient. The third and most important prong is the development of an innovative culture. The creation of an innovative culture is the most signifi cant indicator of a successful digital transformation. An innovative culture involves strong leadership with a

commitment to innovation that is customer-centric, as well as the participation of every single staff member. Ideas come from all stakeholders, and priorities are set to determine which ones to follow through with and which ideas are placed on the back burner. This helps the organization migrate from a risk-averse to an agile and adaptable organization. An innovative culture is also one that creates an environment for cross-pollination, interdepartmental collaboration, strong customer focus and mission, and fi nally, rewarding risks. This type of collaboration works best when there is data transparency and applications are integrated across the utility. Utilities that have had success in establishing an innovative culture often stated that they provided utilitywide culture training. This type of training will be very benefi cial to the water sector as a whole.

Regardless of utility type (government, authority, private industry), silos naturally exist and must be intentionally broken down for this type of technology to be developed and have value. The technical development of a DT itself at the very least requires a mutual understanding of what the data represents and what insights we want to extract. Close coordination among the operations, business, and technical teams is fundamental to understanding the data.

Regardless of the extent of implementation, the conversation by necessity shifts from just data to one of considering values, risk, and costs associated with each level of response. To have such a discussion requires various stakeholder views, including fi nance, customer perspective, asset management, operations, maintenance, engineering, and of course, technology. Using only a single department’s perspective can result in a variety of ineffi ciencies ranging from a tool that is data-heavy and implementation-light to a process that is implementation-heavy but with minimal return on investment. Interdepartmental coordination forces a balance that maximizes effi ciency and ultimately prioritizes the ratepayer’s currency of trust. These types of conversations and consensus won’t occur without breaking the silo barriers.

While making the jump to an innovative culture can seem like a far reach, it comes hand in hand with the

Page 6: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

SMART WATER REPORT 13

innovative technology needed to address the issues of aging workforce and infrastructure. Technology is necessary for systematically capturing institutional knowledge as well as identifying and prioritizing asset management needs. Exploring open-source technologies outside of DTs, such as Microsoft Power BI, cloud-based CMMS systems, and open-data portals is a great way to foster a culture of transparency and help attract millennials. An open culture cannot be cultivated without open technology, but pursuing both simultaneously can be challenging.

3. Big-Picture ArchitectureWater utilities are already using an array of disruptive technologies like sensors, cloud, and analytics to run their operations and engage with their customers. In the past few years, we have seen a signifi cant uptake of digital technologies in the water sector. We detail below the key technologies that make DT possible: infrastructure, data, algorithms, and visualization.

3.1. Technological Infrastructure

3.1.1. Cloud and Edge Computing

Although there are still many applications running on premise-based servers, cloud-based applications are being adopted at a much faster pace in recent years. Cloud computing provides computing resources and data storage on-demand, allowing utilities to consume the resource they need, when needed, and pay for what they use. Utilities have been hesitating to move their data and critical applications to the cloud for security and privacy concerns. However, the trust in the cloud providers has increased and the benefi t of the approach has overcome the barriers. Edge computing is an extension of the cloud computing paradigm. The principle here is to compute the data closer to the data source. This saves bandwidth as well as enables lower cost and lower latencies. For some customers, this is also an increase in safety and privacy.

3.1.2. Wireless Communications and 5G

Wireless communications leverage various radio technologies, which diff er according to the frequency, the modulation scheme used, and the distance they can cover (range). LPWANs with low-power, long-range, and low-cost communication, such as Sigfox, LoRaWAN, or NB-IoT networks, are suitable for IoT applications that only need to transmit tiny amounts of data in the long range. 5G is the fi fth-generation cellular network technology. The fi rst essential deployments (more than 30,000 base stations) of this technology started in April 2019. In 5G, millimeter waves, with shorter range than microwaves, will be used for very short distances (centimeters). Massive MIMO (multiple-input multiple-output) will be used to allow multiple bitstreams of data that can be transmitted simultaneously in parallel, allowing a high data rate. 5G can support a million devices per square kilometer, which is 10 times more than 4G. 5G combined with IoT and artifi cial intelligence is foreseen as a means to accelerate the technological development of DTs.

3.1.3. Middleware and APIs

Water utilities are now dealing with large volumes of data that comprise both structured (easily searchable types) and unstructured (video, satellite images, social media, etc.) data coming from disparate sources. Accessing data from legacy systems is still a challenge. Middleware essentially translates the data from the source format into one that can be consumed easily by an analytical platform. This requires data normalization, which includes ascribing a common name and metadata associated with each data string as well as protocols for cleaning of inaccurate or corrupt records. An application programming interface (API) can then consume the normalized and prepared data. APIs provide a programmatic way for retrieving data.

By utilizing a middleware solution, the data can be consumed by various applications while keeping only one point of maintenance. As an example, the Town of Cary uses a few types of middleware packages, including a Feature Manipulation Engine (FME) for spatialized data. The tool also provides a way for the data viewed in an open data portal or ArcGIS Online. The town also utilizes an integration platform, Dell Boomi, for leveraging data from SCADA as well as customer meter data. This data is then available for consumption by the town’s CMMS for use in work order development. As a result of having this fundamental tool, the town is well-positioned to pursue integrated technologies such as DTs.

Another example used by Global Omnium is GoAigua’s Nexus platform, which is designed to normalize how sensor data is acquired, stored, managed, and shared across the organization in real time. The platform integrates information coming from diff erent vendors and equipment, including on-fi eld components, IoT devices, proprietary, and third-party data. Global

Page 7: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

14 SMART WATER REPORT

Omnium uses GoAigua’s solutions to integrate the organization’s data to provide a holistic view of the integral water cycle.

3.1.4. Cybersecurity

Cybersecurity aims at protecting internet-connected hardware equipment, software, and data from cyberattacks, data breaches, and identity theft. Cybersecurity is used to complement physical security to protect enterprises against attacks and other unauthorized access. Cyberattacks can take several forms, including ransomware (locking a system by encryption and demanding a payment to decrypt and unlock), malware (fi les or programs that harm a computer, such as a virus or spyware), social engineering (tricking users to gain sensitive information), and phishing (fraudulent emails to steal sensitive data).

Security has to be coordinated across the whole communication information system and throughout the organization. All business and operational applications require cybersecurity, including operations, network-connected devices, end-user interactions, and disaster recovery.

Global Omnium is all too aware of this situation and has developed specifi c protocols to prevent incidents as well as response plans. Also, security countermeasures have been introduced to mitigate the associated risks that follow the international standards and guidance. Global Omnium uses the technology provided by GoAigua, which includes key measures and tools to protect both data and analytical processes from attacks, thefts, or other malicious activities that could seriously damage systems.

3.2. Data

Understanding the goal and application of the DT will inform the developer which data is necessary and how frequent the data will have to be retrieved. The complexity and value return can vary signifi cantly from simplistic systems with just a few data streams to ones that incorporate machine learning and sophisticated algorithms utilizing multiple variables. This section outlines the typical data sets needed for a mature DT.

It is imperative that the data utilized is reliable; this sometimes is diffi cult since most of the infrastructure is buried, and sometimes it is challenging to obtain critical pieces of information. In any case, it is necessary to have a culture of quality information registration. Also, having automatic error detection and correction tools (inconsistent properties, failures on the topology of the network or connection) will facilitate the feeding of the DT.

3.2.1. GIS (Geographical Information Systems)

Most utilities today use a geographical information

system (GIS), where all the information on assets is georeferenced. Asset information usually includes pipe network and properties, characteristics, and location of water treatment plants, water storage, regulating elements, sensors, and sometimes the user’s location. The geospatial data along with physical properties of the utility’s assets provide the foundation for the DT model.

3.2.2. SCADA (Supervisory Control and Data Acquisition)

The DT must reproduce the behavior of the real system at every time stamp, so it is necessary to know its operation status. Most utilities have a SCADA (supervisory control and data acquisition) system in place. With their many sensors, they provide real-time data including pressure, fl ow, water quality, and tank level, as well as regulating elements such as valves, pumps, etc. — all of which are critical for data-pairing of the DT.

3.2.3. AMI (Advanced Metering Infrastructure) – Meter Data

Utilities have increased their deployment of smart AMI (advanced metering infrastructure) meters in recent years. The reason for this is that utilities have gained a great deal of value from the hourly consumption data. Water distribution utilities have leveraged AMI data in many ways, such as using analytics to help reduce nonrevenue water (NRW), improving the customer experience by sharing usage data and insights, and performing an analysis to determine loss of revenue due to data leaks. AMI data also helps in the calibration of the DT.

3.2.4. CMMS

Most utilities utilize a CMMS (Computerized Maintenance Management System) to manage workfl ow related to an asset. Since the goal of the DT is to inform and prioritize the optimization of the system, it is vital to exchange information with the CMMS.

For example, in Global Omnium, the main assets such as pumps, regulating valves, and tanks are continually sending their status from the connected sensors, along with the information exchanges via the CMMS. However, some elements — pipes and manual valves, for example — are not monitored. Status information on those elements is exchanged mainly from the CMMS via maintenance people and work orders.

3.2.5. IoT Sensor Data

The Internet of Things (IoT) interconnects computer resources with devices such as sensors via a communications network. In recent years, sensor technology has made signifi cant improvements, resulting in novel sensing capabilities, reduction in the cost of manufacturing, and an increase in the battery life, which is critical to water utilities. These changes

Page 8: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

SMART WATER REPORT 15

make it easier for water utilities to deploy more sensors throughout the utility. However, the main cost to utilities lies in the maintenance rather than the initial purchase of the sensors. Sensor data reliability is critical. As such, the data retrieved must be reliable. Otherwise, erroneous data can translate to faulty insights that can have detrimental implications. Although some novel sensors now include automatic cleaning devices (e.g., using pressurized air or mechanical brushes or wipers), additional development is required to increase the reliability of sensors and reduce maintenance eff orts. Near-real-time sensor data can be used as input parameters to the DT or as data points to help the validity and calibration.

3.2.6. LIMS

Laboratory information management systems (LIMS) can provide water quality data to a DT for updating and calibration of the chemical process model. Since most LIMS are manually updated after analyzing samples collected from the fi eld, there will always be a lag between the DT and the LIMS. This lag could be minimized to less than a day if there is direct integration of LIMS and the DT. Another possibility is to integrate online water quality sensor data along with laboratory-measured data directly into the LIMS. In this case, appropriate time stamps need to be provided for each data point so that they can be appropriately used in the calibration of the DT.

3.2.7. Weather Data

Climate and weather data will play an important role in the evolution of a DT. For example, climatic data such as temperature and soil moisture may be correlated to pipe breakages and leaks. This information can be infused in real time to improve pipe condition assessment of a DT of a water network. Similarly, climate infl uences water demand, giving the ability to forecast both short-term and long-term water demands that can be used readily in a DT in the absence of automatic metering data. Knowing when storms are forecasted will help in preparation and the avoidance of failures.

3.2.8. Access to Accurate, Normalized Data

Critical to DT is the requirement for useful data sets to be secure and agile to access, normalized, and reliable. Data is often collected in various isolated systems, thus making it diffi cult to obtain and maintain for coherence and quality. The units, acquisition, and storage periods are generally diff erent, which makes it diffi cult to normalize. Data normalization is one of the biggest challenges holding utilities back from developing a DT. In most cases, it means some manual intervention initially to set up the rules and algorithms for normalizing data across the utility. After this initial step, it becomes more automated. The quality of the input data to the DT is critical to assure proper behavior.

3.3. Algorithms and Analytics3.3.1. Models: Hydraulic and Chemical

Water utilities around the world have been using hydraulic models for engineering and expansion since the 1950s. A hydraulic model is a physics-based mathematical model of a fl uid fl ow system for water. Typically, the hydraulic model simulation is run in batch mode with input parameters based on historical data. The output of the simulation provides the engineering teams with optimal design choices for determining pipe sizes, developing master plans, and evaluating system expansions. Over the past decades, the models have become more sophisticated by integrating data from the utility’s GIS system, SCADA, and sometimes IoT devices.

In some cases, utilities are starting to use the model simulations for operational uses. What is diff erent with the DT, as discussed earlier, is running the model in continuous mode rather than batch mode and with real-time data pairing from the physical twin. It is necessary to augment the hydraulic model with other models such as a chemical process model, as well as machine learning algorithms for a more accurate holistic model of the water system behavior. For example, the DT of a water distribution system should model the fate and transport of disinfectants and their byproducts as well as potential contaminants through the water pipelines. Recent work by North Carolina State University (NCSU) has shown that a DT of a water network based on hydraulic and chemical process modeling can replace extensive sampling of free chlorine (Ricca et al. 2019) by modeling its movement.

3.3.2. Leak Detection Solutions

Historically, fi nding and fi xing leaks has been challenging, as even a substantial leak can potentially show no manifest signs (Ponce et al. 2014). Large leaks may intuitively seem to be the most signifi cant contributors to water loss, but these leaks often manifest in visible above-ground ways, often with a large enough impact that they are found and fi xed quickly. Small leaks can have a profound impact over time on water loss because they often remain unobservable from the surface. Also, these small background leaks can lead to catastrophic pipe bursts over time. Traditional leak detection methods, such as acoustic surveys, require signifi cant resources and specialized training. The success of an acoustic survey is infl uenced heavily by the pipe material and the magnitude of the leak. Other methods include ground-penetrating radar, infrared imaging, thermal imaging, and gas injection, among many others (Hamilton and Charalambous 2013). With the increasing availability of routinely measured operational data such as pressure, fl ow, and quality, DT-driven methods that work with an entire network and seek to minimize the diff erence between simulated and observed data have recently gained attention. In a DT, leak detection can be envisioned as part of the real-time calibration of the water network since leaks are essentially a manifestation of the physical condition

Page 9: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

16 SMART WATER REPORT

of the pipe network. This calibration can be performed using optimization or machine-learning approaches that are embedded in the analytics driving the DT. For real-time applicability, these methods should be fast and computationally tractable. For this reason, heuristic optimization methods such as genetic algorithms and particle swarm methods that have often been used in numerous published studies will have limited applicability. Alternatively, optimization methods based on successive linear approximation developed at NCSU (Berglund et al. 2017, Mahinthakumar et al. 2018) have shown that online pressure measurements can be used with a real-time hydraulic model to detect background leaks accurately and in real time. Such methods should be explored toward leakage detection in a DT.

3.4. Visualization

The enormous potential of the DT is based on the capacity to present a large amount of information collected in it in an organized and understandable way, adapted to diverse users and requirements. The DT permits the development of very diff erent display environments for a diverse set of users, operational staff , fi eld staff , and business staff , to name a few.

3.4.1. Reports and Charts

The tables and graphs join all the required data together and transform it in a way that can be easily read and interpreted. In short, the visualization of the data should be a powerful and fl exible query tool, adaptable to the required use.

3.4.2. Simulations

The DT is an interactive environment where users can perform simulations from a historical time perspective, in near-real time, and in future time, such as what-if scenarios. These capabilities help staff members better understand what is happening across the utility at any time frame. Utilizing a DT for test scenarios can help illuminate the impact of specifi c actions before one takes them. The geographical representation of the DT simulation is especially powerful with its ability to continually show the state of the whole system.

3.4.3.Web-based SCADA

An example of data visualization is the Town of Cary’s SCADA Ignition portal, which makes SCADA historian data web-available in a user-friendly form for various sectors. Traditionally, each treatment plant has SCADA visualization on-site, but this is not accessible to fi eld staff or engineers working off -site. This lightweight interface provides quick and helpful access to various data, including collection and distribution data. Below is a view of the Town of Cary’s SCADA Ignition portal for wastewater pump station data.

4. Actionable Steps with Case Studies4.1. Develop a VisionBecause DT aff ects so many departments, a unifi ed vision for the desired outcomes is necessary to build consensus and evaluate success. Prioritizing the drivers, which can range from leak management to operational effi ciency, to improve water quality will help develop the business case. This vision needs to be cross-departmental and aimed at fostering an innovative culture. Only once there is a shared consensus on the big-picture vision for adaptive management can a digital strategy be developed.

4.2. Assess Where You Are Today — Components and Process & GoalsFor utilities that are just beginning to consider DTs, it’s important to understand the core components and start working toward those fi rst. Fortunately, each of these core components, identifi ed below, has its benefi ts.

• The Base: GIS & Hydraulic Model – Is your GIS reliable? Is your model calibrated? The level of accuracy needed for master planning is less than what would be required to make operational decisions. A well-calibrated model, even if it’s skeletal with only major trunk lines, is critical to the success of a DT.

• The Data: SCADA & Meter Data – Most utilities GoAigua visualization of Global Omnium DT in Valencia, Spain and its metropolitan-area water distribution network

Page 10: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

SMART WATER REPORT 17

are utilizing SCADA, but the level of customer meter data can vary. Technically, a DT can be leveraged even without customer data and using pressure or fl ow meter data for smaller district metered areas (DMAs). AMI data provides a level of granularity on distribution consumption that, when paired with specifi c machine-learning algorithms, can provide critical insights for greater effi ciency and early failure detection. Such insights include the discovery of anomalies, distribution leaks, customer leaks, as well as data leaks (faulty meters, wrong-size meters, missing meters, missing bills, etc.). Either way, consider how the data can be accessed and the frequency of its availability (this determines how “live” your system will be, but may need to be balanced with battery life).

• The Platform: Middleware & API – Some type of middleware will be needed to access SCADA and meter data. Custom portals such as SCADA Ignition are a good start for building the appetite for more complex DTs, but pulling multiple sources into one API will require more coordination. While data exports can be leveraged to an extent, manually incorporating data will quickly prove insuffi cient with the amount of sensor and meter data involved in a mature model. Also, the premise of DTs is an active integration, so custom portals per data stream will only get you so far.

After reviewing the Key Opportunities (Section 2) and assessing the core components, you can then refi ne your goal and data profi le required to achieve it. Cross-departmental conversations are needed even at this step to understand the feasibility of the goal and the level of eff ort needed to get there.

For example, we have a goal to improve the speed of response to reported leaks. Today, most leaks are reported by a customer calling in faster than the required data is collected and analyzed. In order to have data analysis happen sooner than a customer could detect the leak, the frequency of data collection must be faster than what is being done today. The key takeaway is that understanding the data requirements to achieving your goals is critical during the planning stages and cost vs. return analysis. Additionally, cumulative value must be weighed against the maintenance involved for both physical aspects (sensors, IT network) as well as support needed to program and manage these technologies (model updates, network patches) and more.

4.3. Develop an Innovation CultureOpen technology requires an open culture to both develop and sustain. The cross-departmental questions of “How could we use a DT?”, “How do we create a DT?”, and “What are the long-term costs and benefi ts?” need to be brainstormed together.

The Town of Cary hired a new town manager, Sean Stegall, in 2016 to help the town transition from a phase of building to one of maintaining. The premise was to

fundamentally shift from a traditional siloed approach to become more like a startup, with agility and innovation fostered between departments. To realize this vision, a deliberate and ongoing discussion began about the value of adaptive leadership. Simultaneous to this culture shift, several open technologies were brought online and rolled out internally and somewhat organically. The technologies range from out-of-the-box software (like Offi ce 365) to more complex data management (like ESRI) and even an open source and open design database tool (Salesforce), which the town is confi guring, using it as both a CMMS and work order system as well as a project management tool. This vision for both open culture and open data has been a great breeding ground for fostering young talent and lays the groundwork for innovative technologies like DT.

4.4.Build and Calibrate a Simulation Model

The hydraulic model is the basis for the DT decision making because it allows operators to simulate the system response under any request. In this way, the hydraulic model must be reliable and alive, able to accurately reproduce the network behavior at any time and under any circumstance. Also, the simulations must be carried out quickly, so the development of the hydraulic model is the fi rst challenge to overcome to develop the DT of the system.

Nowadays, all the necessary information to develop a hydraulic model is available in diff erent systems (GIS, SCADA, AMI, CMMS, etc.). So to migrate from a batch hydraulic model to a more continuous model, we have to connect the hydraulic model with all the information sources to always have an updated model. The information collected by SCADA is the most variable, so the fi rst step, once we have a base hydraulic model, is to connect it with the SCADA system (Bou et al, 2006).

Furthermore, the DT must be live and reliable; therefore, the hydraulic model has to be calibrated. There is much research about hydraulic model calibration techniques, emphasizing the adjustment variables to choose, the optimal location of the measurement points, and the objective function to minimize in each case with the restrictions to be imposed, as well as in the numerical techniques to minimize the error (Savic et al. 2009, Martínez et al. 2017). When building a hydraulic model, the information provided by SCADA is used for establishing the network operation and for model calibration, so one set of data is used to establish the real system operation (control rules) while the other set of data is used for calibration or validation of the model.

However, if the goal is to obtain a real DT, the hydraulic model must always be updated with the key information sources discussed above. GoAigua and Global Omnium have made a big eff ort in recent years to have a hydraulic model that is constantly data-paired with the right data sources, making it a live

Page 11: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

18 SMART WATER REPORT

DT. Additionally, it is important to establish boundary conditions that may aff ect the operation or behavior of the system (user demand, climatology, raw water quality, availability of resources, etc.). While the data streams will become the source of this information, it is helpful to have anticipated ranges to improve model calibration.

4.5. Bringing It All Together – An Example of a Live Digital TwinGlobal Omnium’s digital transformation started 12 years ago and focused on both its processes and infrastructures. Global Omnium made a bold investment in equipping the entire network with sensors at nearly every asset. There was a great deal of emphasis on modifying the internal processes as well. This technological transformation allowed the utility to obtain a great amount of information and key data, represented in real-time environments. These early changes saved the company 7 million euros (US$7.7 million) annually. The money saved was then reinvested back into technology and integration. First up was installing smart meters across the City of Valencia, 700,000 meters within a fi xed network.

Five years later, in 2014, the company decided to develop a scientifi c data unit, incorporating disruptive technologies such as machine learning, artifi cial intelligence, Big Data, and advanced algorithms. This experience allowed Global Omnium to develop ad hoc solutions that stitch together the separate systems (silos) and adapt to the diff erent stages of the integral water cycle. Examples of this include fi eld validation, leakage detection, work order management, virtual offi ce, customer service, and DT, among others. These solutions are now also products of the newly formed spinoff from Global Omnium called GoAigua.

Today, the GoAigua platform integrates information coming from diff erent vendors and equipment, including on-fi eld components, IoT devices, and proprietary and third-party data. As an example, the DT

of Valencia and its metropolitan area is very accurately calibrated, with an error less than 2 percent in pressures and 4 percent in fl ows, which also precisely reproduces all the tank levels.

5. ConclusionWhen done correctly, the DT can be an eff ective and powerful holistic tool for utilities. It has the power to bring together diverse perspectives, data sets, and solutions under one umbrella application. This, in turn, can lay the groundwork for increasing interdepartmental transparency and collaboration on broader business and operational goals. It is evident that for utilities to be successful in implementing DT, they must simultaneously undergo digital transformation, which begins and ends with people, and that change management is vital. The reason for this is that a process of this kind must overcome the inertias rooted in the organization’s structures and identify the key players to ensure that new technologies are not only implemented, but they are operationalized and executed successfully.

It is clear that the global water sector has embraced this technology; however, it is understood that it will not be an easy journey for most utilities, regardless of size and structure. There exist many challenges ahead for utilities, both technological and cultural, when implementing DT technology. For this reason, the SWAN organization launched the SWAN Digital Twin H2O Work Group, creating an environment for global leaders to develop best practices and share aff ordable solutions and lessons learned from successful case studies in a collaborative environment. Regardless of the challenges ahead, all utilities should be working toward the development of DT technology to help address the global water crisis.

General approach proposed.

Page 12: EXCLUSIVE INSIGHTS€¦ · 06-11-2019  · Customer Engagement Resiliency Water Quality Monitoring Workforce Management Collaborative Innovation Stormwater Digital Twin Asset Management

SMART WATER REPORT 19

is a water resource engineer for the Town of Cary’s Utility Department. She came to the Town of Cary in 2016 with eight years of experience in engineering consulting. Her primary role is long-range water resource planning, including trend and data analysis.

SARAH BRAMAN

president of Karmous-Edwards Consulting, has spent the majority of her career on state-of-the-art digital technologies, including IoT, Big Data, data analytics, control and automation, and, most recently, Digital Twin technology. As an academic researcher, she authored over 23 peer-reviewed publications, including a Wiley Book on Grid Networks, and has recently published several papers on the digital transformation of the water sector. She was the founder and CEO of a product company with a successful exit and now consults on digital strategy and transformation with current focus on the Digital Twin in the global water sector for utilities and technology companies. She also advises

digital technology companies in the water and agriculture sectors on product road maps, digital strategy, IoT strategy, competitive analysis, business growth, market segmentation, M&A, and potential partnerships. Karmous-Edwards recently founded and is a cochair of the SWAN Digital Twin H2O Work Group to help advance the use of the much-needed holistic utility management using Digital Twins. She earned her B.S. in chemical engineering and an M.S. in electrical engineering.

GIGI KARMOUS-EDWARDS,

is the CEO of GoAigua and chief digital and disruptive offi cer of Global Omnium, CEO of Nexus Integra, and a member of the advisory board of Adesal Telecom. Barba is an expert in the digital transformation of the water industry, with more than 18 years of experience in managerial roles, currently leading organizations and water utilities in their digitalization journey. He is passionate about technology and people, and working to transform the water sector.

JAIME BARBA

is a professor in the Department of Civil, Construction, and Environmental Engineering at North Carolina State University. He has worked in the area of water distribution systems analysis for nearly two decades. His areas of interest include real-time hydraulic and water quality modeling, leakage detection and pipe condition assessment, resilience analysis, and optimization algorithms.

KUMAR MAHINTHAKUMAR

is responsible for network control and operation at Global Omnium. She is a Ph.D. industrial engineer with more than 15 years of experience in the fi eld of hydraulic engineering and water management. Pilar is an assistant professor in the Department of Hydraulic Engineering at the Universitat Politècnica de València, Spain, as well as the coauthor of several articles and papers in congresses.

PILAR CONEJOS

is a graduate of the National Institute of Applied Sciences (Lyon, France) in software engineering and computer science. Vicat-Blanc obtained a Ph.D. in robot programming in 1988 and research habilitation in 2002. She pursued an academic career as associate professor and senior scientist in computer science up to 2010. Her research interests are networking, cloud computing, and IoT. She has published more than 150 articles, supervised 12 Ph.D. students, and managed a dozen international projects. In 2010, she created the cloud-computing startups Lyatiss in France and CloudWeaver Inc. in the U.S., selling CloudWeaver to F5 Networks in 2015. In 2018, after eight years in Silicon Valley, Vicat-Blanc

returned to France, where she collaborates with the computer science lab at Sorbonne University and conducts research at INRIA. She is also founder and CTO of Stackitect, a new startup, to help the IoT teams and executives with strategic, technical, and operational decisions. StackiLab, which combines complex systems theory, agent-based simulation, and Digital Twin technology, is the fi rst IoT architecture intelligence platform. Vicat-Blanc received the Legion of Honor medal in 2010, the Ministry of Research’s Women and Business Award in 2011, and the Innovation Award from the Academy of Sciences in 2013.

PASCALE VICAT-BLANC