Skip to content
Case study

Digital Twins for Building Flexibility into Power Plant Operations

29 April 2019

A research group in the US has developed a digital twin for optimising power plant operations in numerous ways.

Digital Twins for Building Flexibility into Power Plant Operations
fossil-fired thermal power plants

Enhancing the operational flexibility of fossil-fired thermal power plants is a key challenge facing today's power generation industry given the increasing penetration of intermittent renewable energy sources into the electric grid (IEA, 2018). An important digitalization technology for meeting this challenge is the digital twin — a powerful tool for analyzing and testing approaches for mitigating the negative impacts of flexible plant operations.

The foundation of our nation’s electric grid is coal- and gas-fired power plants that were originally designed for base-load electricity generation. Although designed for continuous operation, these plants are increasingly forced into flexible operating modes, including frequent cycling and sustained low-load operation.

It’s not necessarily a good fit. Industry experience to date shows that flexible operations can lead to large thermal, pressure, and chemical stresses that damage equipment, compromise plant life expectancy and performance, escalate plant operations and maintenance (O&M) expenditures, and increase environmental emissions (NETL, 2015).

Digital twins

Digital twins are virtualizations of physical plant assets and represent the structure and behavior of those assets in real life. Engineers and operators can interact with digital twins using 2-D human-machine interfaces or 3-D virtual- and augmented-reality environments based on computer aided design (CAD) models.

A digital twin based on high-fidelity real-time dynamic models can be used across the plant life cycle to bridge design, operations, maintenance, and optimization (AVEVA, 2017; GSE, 2016). The real-time status of the digital twin can be updated constantly using operating data from sensors in the plant. Historical plant data can also be integrated into the digital twin for use in condition monitoring and predictive maintenance applications.

Power plant digital twin for research on flexible operations

The U.S. Department of Energy’s National Energy Technology Laboratory (NETL) and its research and development (R&D) partners developed a power plant digital twin and deployed it at NETL and the Advanced Virtual Energy Simulation Training And Research (AVESTAR®) Center at West Virginia University (WVU, 2019).

The team’s digital twin simulates an integrated gasification combined cycle (IGCC) power plant with carbon dioxide (CO2) capture — an attractive technology option for coal-fired power plants requiring efficient CO2 capture and storage (NETL, 2019). The digital twin combines high-fidelity real-time dynamic simulators for “gasification-based chemical processing and cleanup” and “combined-cycle power generation” together with 3D CAD-based virtual-reality technology (Zitney et al., 2017).

Now in use, the IGCC digital twin serves as a virtual test bed to help address the critical R&D challenges facing the energy industry on its drive toward operational excellence (Zitney, 2019). R&D focus areas of interest include optimal sensor network design, advanced process control, and operational strategies for improving power plant performance, flexibility, and health monitoring.

Optimal sensor network design

One important advantage offered by digital twins is their cost-effective use in optimizing a plant’s sensor network. Digital twins can be used to test novel algorithms for automatically and systematically calculating the optimum placement, number, and type of sensors. This is a key digitalization research challenge, because the data collected from sensors is vital in enhancing power plant performance, improving reliability and availability, and enabling effective condition monitoring and fault diagnosis for key plant components under flexible operation (NETL, 2019).

At NETL and WVU, researchers have developed a nonlinear dynamic model-based multi-objective sensor placement algorithm for plants with estimator-based control systems (Paul et al., 2017, 2015; Jones et al., 2014a, 2014b). Using the digital twin, the research team developed an optimal sensor network design that can maximize CO2 capture efficiency during IGCC load-following operation.

The research team is also using the IGCC digital twin to analyze sensor networks for monitoring the spatial profile of process states and estimating fault severity levels at the component-level. The gasifier represents the heart of the IGCC system, and the digital twin is being used to optimize sensor networks that will minimize degradation of the refractory lining under the high-temperature and high-pressure conditions used to convert coal into synthesis gas for fuelling a combustion turbine as part of the combined cycle to produce power (Mobed et al., 2016).

Advanced process control for power plants

Considerable research challenges and opportunities exist in the development and application of advanced process control strategies for power plants with carbon capture. Digital twins can be valuable tools for studying control strategies for generating and controlling supervisory set points to optimize power plant performance.

For example, transient analysis using the IGCC digital twin shows that the air separation unit providing high-purity oxygen for the gasification process poses a bottleneck by limiting the overall plant ramp rate during flexible operation. The digital twin-based analysis also shows that a multiple-model predictive control strategy for air separation unit operation that switches between a collection of linear model predictive controllers along the power demand load range would provide a better response to plant-wide IGCC load-following compared to conventional control approaches (Zitney et al., 2017).

The IGCC digital twin is also assisting in another effort to improve flexible power plant operations. NETL and WVU researchers are developing a biologically inspired, distributed multi-agent control strategy with an artificial neural network-based adaptation. Using the IGCC digital twin as a computational test bed, these biomimetic adaptive control approaches are providing faster and more accurate setpoint tracking for maintaining the target carbon capture rate during IGCC transients (Mirlekar et al., 2018).

Industrial applications of digital twins

The digital twin has also shown value in applications beyond IGCC power plants. Under the auspices of a cooperative research and development agreement established with the National Rural Electric Cooperative Association, NETL researchers modified the power generation section of the IGCC digital twin to analyze flexible operations of an industrial natural gas combined cycle (NGCC) with duct burners in the heat recovery steam generator (Liese and Zitney, 2017).

The NGCC digital twin revealed the potential for saturation conditions in the final high-pressure superheater as the attemperator tried to control temperature at the superheater outlet during loading and unloading of the gas turbine. Subsequent NGCC plant operational data confirmed the dynamic simulation results. Multiple simulations performed during loading and unloading of the gas turbine identified flexible operational strategies that prevented saturation and increased the approach to saturation temperature. The solutions included changes to the attemperator temperature control setpoints and strategic control of the steam turbine inlet pressure control valve.

Conclusions and future applications of digital twin technology

As the era of digital transformation accelerates, the energy industries are pursuing opportunities for improving flexible power plant operations. The goals are increasing plant efficiency, reliability, and profitability, while reducing plant downtime and O&M costs.

Digital twins are key enabling technologies for industrial application and R&D on flexible power plant operations, including optimal sensor network design, advanced process control, condition monitoring, and fault diagnosis.

Future applications of these powerful digitalization technologies include:

  • Adaptive hybrid digital twins with real-time data-driven models synthesized into high-fidelity physics-based dynamic models.
  • Cyber-physical systems integrating digital twins with physical assets to reduce design time and operational risks.
  • Digital twins for developing and testing cybersecurity systems to detect, localize, and neutralize cyber attacks.

Figure 1 shows a human-machine interface (HMI) display from the power plant digital twin. (Click to enlarge)

Figure 2 shows an engineer using the digital twin in a control room environment with the 3D virtual plant on the large screen in front of him. (Click to enlarge)

Figure 3 shows a screen capture from the 3D immersive virtual-reality interface to the digital twin. (Click to enlarge)


Status of Power System Transformation 2018 IEA, 2018

Impact of Load Following on the Economics of Existing Coal-Fired Power Plant Operations, 2015 NETL 2015

Building a Digital Twin of Your Process Plant with Unified Lifecycle Simulation, By Norbert Jung and Cal Depew, AVEVA, 2017;

Increasing ROI through Simulation and the “Digital Twin” GSE, 2016

Teaching future power plant operators WVU, 2019

Commercial Power Production based on Gasification NETL, 2019

Dynamic IGCC system simulator Zitney et al., 2017

Dynamic Model-Based Digital Twin, Optimization, and Control Technologies for Improving Flexible Power Plant Operations Zitney, 2019

Sensors and controls NETL, 2019

Nonlinear Dynamic Model-Based Multiobjective Sensor Network Design Algorithm for a Plant with an Estimator-Based Control System by Prokash Paul, Debangsu Bhattacharyya, Richard Turton, and Stephen E. Zitney, 2017,

Sensor network design for maximizing process efficiency: An algorithm and its application

Prokash Paul, Debangsu Bhattacharyya, Richard Turton, Stephen E. Zitney 2015;

Plant-wide control system design: Secondary controlled variable selection, Dustin Jones, Debangsu Bhattacharyya, Richard Turton, Stephen E. Zitney 2014a,

Find out more

  • Stephen E. Zitney, U.S. Department of Energy