Doing the Right Work at the Right Time in the Power Plant of Tomorrow

By Steven Seachman
Senior Technical Leader, Electric Power Research Institute

Cisco estimates 21 billion devices will connect to the Internet by 2018 (three times the world population and up from 14 billion in 2013).1 This number will include sensors and other devices that aid in the supply and use of electricity. The proliferation of these sensors, the data they collect, and sophisticated new technologies that enable transformational applications of that data will profoundly change society, including the way we generate, distribute, and use electricity.

There are, however, some challenges. They include the sheer volume of data; proprietary legacy systems; the need for enhanced security; inconsistent life-cycle timescales of utility assets and connectivity technologies; rapid technological change; and effective integration of technologies into the power system, including intelligent devices, sensors, advanced metering, and even customer technologies. Fortunately, these challenges also bring opportunities spanning the energy value chain. A digital world full of new technologies with vast potential for addressing most needs via a pocket-held device offers an as-yet untapped arsenal of tools that could meet these challenges and transform the industry’s generating assets.

The I4GEN concept includes connectivity at every level of information, providing the data needed to the proper systems and personnel. (Courtesy EPRI)

The bottom line is this: The power plant of tomorrow is likely to generate electricity in much the same way it has in the past, but the way in which the plant is controlled and operated—using digital technology—will change dramatically. And these changes will be driven by new, advanced sensors and data analytics.


Today’s typical large-scale power plants have existing systems, communication capabilities, and work processes. Some of these processes are paper based and communication occurs through a variety of disconnected, ad hoc channels such as e-mail, phone, text, radio, paper, and electronic data/entry into various software-based solutions/data management tools. Considerable time is wasted seeking, assembling, and aggregating data, as well as re-entering data in disconnected systems, which limits the amount of time available to analyze data and develop a comprehensive understanding of an issue or solution.

I4GEN includes digital worker technologies that provide all needed information to the technician based on component and job requirements. An EPRI video ( provides an example of a plant worker using a wearable technology to assess equipment condition. (Courtesy EPRI)


The future power plant operating workforce will use a digital platform in which information produced in real time is used to estimate equipment condition, and algorithms are used to forecast a set of operating conditions. The connected and integrated digital networks automatically integrate data and produce the information for various systems, functions, analysis, personnel, and actions. The computation, communication, and linking of systems are embedded with interfaces that are easy to use, but secured at various levels within the network. An open-architecture modular system allows for “plug and play” of new devices, software, and other algorithms regardless of the developer or vendor.

The Electric Power Research Institute (EPRI) and other companies are developing detailed visions of tomorrow’s power plant with seamless integration of data, autonomous communication of information, and corresponding response, action, or control. EPRI’s I4GEN (Insight through the Integration of Information for Intelligent Generation) concept is a holistic approach that creates a digitally connected and dynamically optimized power plant by using a modular and scalable platform of tools, techniques, and technologies that integrate the business, maintenance, and operational aspects of generating power. The goal is to enhance performance, reduce failures, increase availability, improve flexibility, and minimize cost. Dynamically optimizing a plant requires collection and aggregation of data and production of real-time information; embedded distributed and adaptive intelligence that supports decision-making; and identification of actions and responses to account for risk, reward, and uncertainty.

The I4GEN concept is to optimize the use of all available information to make better decisions for the plant, be it autonomous or by the plant operator. Emerging data analytic algorithms are intended to detect degradation at early stages, diagnose the most likely cause of degradation, and estimate the remaining time to take action before an operability limitation will be reached. Successful development and deployment of detection and diagnosis algorithms can result in increased plant reliability, decreased plant O&M costs, reduced forced outages, more efficient use of O&M resources, and minimization of the impact resulting from flexible operations. This is performed through many different avenues. Data can come from operating data from instrumentation in the plant; external sources such as weather conditions, expected demand on the plant and plant flexibility; archived data; and subject-matter expert (SME) specifications and guidelines.

Successful collection and analysis of the most appropriate data enables important advances such as prognostics—doing the right work at the right time—maintaining, repairing, or replacing equipment when needed, instead of when recommended based on hours of service or other, less precise metrics. This enhanced diagnosis of component condition and remaining life can come from optimizing sensor suites. Information is then gathered where needed and vital type and location sensors are identified (what types, how many, where they are located, reduced redundancies); sensor groupings and embedding intelligence into sensors so that they continue to retrieve valuable information even as sensor data drifts or certain sensors fail; and data analytics, statistical methods, model-based methods, SME input (when needed), and machine learning.

The ultimate goal of I4GEN is to make the most useful information available to a person at the time it is needed to make a decision or perform an action. A power plant using I4GEN produces, shares, manages, and manipulates information at the appropriate time, within proper context, and at a level of detail sufficient to support a response. It is enabled through an open-architecture communication framework that is scalable, modular, and secure. Its foundation relies on dense and distributed data collection, aggregation, and computational analytics to generate actionable information. Optimization and effectiveness are the result of having actionable information in context and at the appropriate time to achieve an objective.

The drivers for adopting the new digital technologies for power generation envisioned in I4GEN are many:

  • Grid modernization and integration: As part of an integrated grid, power will be generated from a range of sources; the mix of types, sizes, locations, and intermittent operations adds layers of complexity to the grid control. Dynamic, fully integrated generation assets are required to achieve the full benefits of the integrated grid.
  • Critical cost-competitiveness: Maintaining the reliability and availability of power generation is paramount when introducing new facets of flexible operation. Quantifiable benefits from implementing and adapting the I4GEN architecture/framework would be specific to the generating asset, how it is used within a fleet, and how it is used for a region’s dispatch to the grid.
  • Changes in the generation resource portfolio and the need for operational flexibility: The mix of generating assets is changing as the portfolio becomes more diverse with the advent of advanced power cycles, renewables, energy storage, microgrids, and other distributed generating assets.
  • Changing workforce: Highly skilled and experienced workers are either retiring or preparing to retire from the power industry. The new workforce brings a different set of skills and capabilities, including an in-depth familiarity and expertise with digital technologies. Plants facing a significant turnover with staff and a potential loss in expertise may also view an investment in digital technologies as a means to capture and automate the expertise, facilitate training of new personnel, and reduce risk associated with staff turnover.
  • Inherent value in owning, managing, and controlling data and information: Generating plants produce vast amounts of data: Managing and sharing that data is a critical function for moving from a reactive state to a more proactive state. As organizations recognize and assign value to plant data and information, similar to how they currently treat financial data, the opportunities, benefits, and drivers associated with this valuation emerge.

Workflow is optimized and automated through connectivity of all needed information. (Courtesy EPRI)


The I4GEN concept can be applied to all types of generating assets. Remotely located generating assets, such as hydroelectric and wind, which do not maintain large, onsite staffs may develop highly advanced monitoring and diagnostics to support more effective use of onsite inspection and maintenance. Slightly different drivers and emphasis may be placed on large-scale central power stations in which enhancements in process controls may be needed to support operational flexibility.

Besides producing electricity, all these generating assets have something in common, the two keys to success in their transition to the generating plants of tomorrow: advanced sensors and the data analytics needed to realize the value of the information they provide. After the sensor suite has been developed, the sensors are able to communicate, and data is being fed into the data analytics suite, the plant should be able to detect, diagnose, and act rapidly with much less downtime than current methods allow.

The I4GEN concept requires connectivity and communication between systems, hardware, and software users. This requires an increase in data collection; autonomous data integration; methods for massive data management; ability to reconfigure data integration and analysis; incorporation of advanced query capabilities; and application of intelligence algorithms (e.g., cognitive, analytics, artificial intelligence, etc.). Enabling technologies include component and system modeling, augmented reality, visualization, and networked systems (hardware and software) to provide real-time information, distributed and adaptive intelligence, and action and response.

Novel sensor technology, such as this torsional vibration sensor, can be used for early detection of component faults. (Courtesy EPRI)

Real-time data on component status will identify developing problems and support condition-based maintenance to help prevent failures and avoid outages. Improved situational awareness will allow operators to extend maintenance intervals and maximize asset utilization, helping reduce costs and improve productivity without affecting safety and reliability. The ability to monitor key parameters in areas that could not previously be accessed—or only accessed with significant cost and safety implications—will enable operating and maintenance interventions to address incipient problems and otherwise improve the performance of electricity infrastructure.

EPRI’s Technology Innovation (TI) program is pursuing novel sensor designs for steam turbine and combustion turbine compressor blades and pressure-retaining components, as well as overhead transmission lines, transformers, underground distribution cables, and other applications. This effort also focuses on core technologies for data analysis, decision support, and power harvesting for self-powered sensor technologies.

Three examples of sensor technology developments specific to power generation that are supported by EPRI TI projects include blade vibration sensors, laser-based sensors for coal gasifiers, and fiber Bragg grating (FBG) sensors for nuclear plant applications. Since 2009, EPRI has been leading efforts to create a microelectromechanical (MEM) sensing system for direct online vibration monitoring of large blades in low-pressure steam turbine stages to detect incipient damage and avoid catastrophic failure, which poses major cost and safety risks at thermal power plants. The system is also applicable for combustion turbine compressor blades.

The shaft systems on large grid-connected steam turbine generators can be subjected to dynamic torque (or “twisting”) oscillations caused by negative sequence currents in the generator. To date, prototype EPRI torsional sensors to detect these oscillations have been installed and field tested on three shaft locations on two separate generating units. The accumulated operating time represents a total of 11 unit-months of operation without failure of the shaft sensors, circuit boards, or the stationary antenna assemblies. Compared to earlier technologies, the system is highly sensitive and provides more data, with a higher degree of granularity, making it easier to tell where the torsional peaks are. The data provides an accurate picture of each individual torsional node, enabling assessments of the failure mechanisms. It is anticipated that highly sensitive and low-noise shaft strain monitoring will be the basis for advances in a range of new condition-monitoring applications on a wide array of power generating equipment.

Reliability problems challenge the economics of integrated gasification combined-cycle (IGCC) plants, creating the need for a system to prevent the temperature excursions that damage refractory linings. Based on exploratory research initiated in 2000, EPRI demonstrated the feasibility of applying tunable diode laser (TDL) technology for this application. An initial prototype delivered accurate temperature readings and the first direct, real-time measurements of key chemical species in the high-temperature, high-pressure, particulate-laden gasifier environment. Follow-on scale-up experiments supported development of a TDL sensor system incorporating advanced spectroscopy techniques and multiple diodes tuned to the wavelengths of targeted species. Commercial TDL sensors are expected to provide real-time data for precise monitoring and control of the gasification process to improve reliability, conversion efficiency, and environmental performance at IGCC plants.

Other EPRI TI advanced sensors and data analytics projects include:

  • Transient analysis methods: Much like autos running at highway speed, most bulk power generation assets operate most efficiently in steady states at high capacity. Their components are under the most stress during start-up, load change, and shut-down cycles. EPRI has explored use of transient analysis methods to uncover data anomalies and trends that indicate the onset of aging or failure. In a proof-of-concept study, aging-related performance degradation of a high-speed motor not evident in steady-state data was clearly detected in start-up data. Follow-on work has established a novel method, sharp time distribution mapping, as a generalizable approach for applying transient sensor data to improve anomaly detection for power generation and delivery system components.
  • Decision-support technologies: EPRI is creating a knowledge and capability base to enable accurate and timely decision-making in power plant and grid control centers where personnel are challenged to handle large amounts of complex and diverse data from sensors and other sources with support from growing amounts of automation. A long-term, multidisciplinary research plan has been defined for addressing the industry’s decision-making needs with interactive human–system interface (HSI) technologies, including analysis, visualization, and simulation tools. In addition, based on advanced HSI test cases, design guidelines have been developed to ensure that decision-support technologies meet application-specific needs while avoiding loss of situation awareness and other new types of errors associated with increased automation. The guidelines will aid in the design of user-centered HSIs, responsibilities, and workloads to support decision-making in environments where cognitive processing is essential, such as control rooms, control centers, and monitoring and diagnostics centers.
  • Power harvesting and storage: A number of promising methods have been identified for exploiting ambient energy sources to support the autonomous operation of sensors and associated electronics within power generation and delivery infrastructure. A laboratory test bed for evaluating harvesting technologies has been constructed, and this project continues to focus on applications of self-powered sensors in nuclear, fossil, and renewable generation systems and on the transmission and distribution grid.


Adopting the I4GEN approach in totality will be a large undertaking. In many cases, adoption of selected digital technology platforms and capabilities over time, with short-term returns on investment and measurable benefits, is a more likely scenario. Near-term development and demonstration opportunities that offer tangible benefits and utilize many of these enabling technologies include:

  • Digital worker: Tablets, smart phones, laptops, wearable monitoring devices, headsets, and augmented reality devices can be used by staff to carry out or complete a given job function. Wireless communication capabilities; developing relevant information, procedures, guidance, equipment tagging, operator rounds, and work order entries in a digital format that is easy to access and intuitive; and ergonomics, safety, and impact on situational awareness need to be evaluated under different scenarios.
  • Virtual reality and simulated plant analytics and operation: 3-D interactive images of workspaces and equipment layouts support training and assists in work planning and can be linked to digital devices to support work execution. A plant process simulation running in near time using data from the plant can be used to assist in optimization of the process and provides a safe environment to forecast a number of operating conditions to aid the plant in performance.
  • Low-cost sensing and monitoring: Additional data can help produce information and insights about the process, equipment and component conditions, and other operational aspects, but the cost to purchase, install, and maintain these sensors can be a challenging hurdle. Many new types of sensors are entering the market that are both low-cost and wireless; installation costs can be comparably less if a wireless network is available and powering the sensors can be managed cost-effectively. A number of low-risk opportunities exist to demonstrate new sensor technology and evaluation of additional data may bring in greater insights and proactive operations and maintenance.

The I4GEN concept uses monitoring and diagnostics centers to provide needed information from the equipment level to the fleet-wide operations level. (Courtesy TVA/EPRI)

EPRI plans to develop and implement the I4GEN technologies discussed above that will support power generation companies at various stages of maturity. Whether an organization is at the beginning stages of learning about I4GEN-associated technologies and applications or have already invested in advanced remote monitoring centers, digital worker applications, etc., this development and implementation will provide opportunities for collaboration, technology transfer, and development of industry guidelines to further work in this area.


  1. Cisco®. (2016). Visual Networking Index: Global mobile data traffic forecast update, 2015–2020. White Paper. San Jose, CA: Cisco Systems, Inc.,


The content in Cornerstone does not necessarily reflect the views of the World Coal Association or its members.
Receive e-mail alerts when the new issue comes online!
Click here to opt-in or opt-out.
Receive the new edition in print!
Click here to opt-in or opt-out.