As experienced power plant operators retire and the industry become more decentralized, could AI help fill in the knowledge gaps?
By Stephen Kwan
The utility industry is facing increasingly complex planning and implementation methods, in addition to an industry-wide shift, moving from massive, traditional central power stations to smaller distributed systems. Not to mention, experienced power generation operators across the globe represent an ever-shrinking workforce whose deep domain knowledge and expertise are central to efficient, low risk, and reliable operations.
As the industry works on accelerating its pace of innovation to meet increasing power supply needs, reduce carbon emissions, and solve the challenge of a shrinking workforce, the role of artificial intelligence (AI) has become ever more central for managing the health and performance of key large-scale assets, implementing complex networks of decentralized systems, along with digitizing and democratizing domain expert knowledge.
Revolutionizing Asset Health & Performance Management of Key Large-scale Plants
Advanced AI approaches have great potential to revolutionize the management of critical assets across the power generation domain, combining historical and real-time operational datasets with deep subject matter expertise. Solutions with such capabilities are able to provide recommendations on asset health management while maximizing the overall efficiency and reliability of operations to meet power generation demands and financial objectives.
Production goals are essential to power generation facilities that participate in the grid or provide localized power but optimizing operations to meet these goals while balancing the health and performance of assets can be challenging. One critical value of AI in the power generation sector is the ability to model a layered approach inclusive of multiple assets, plants, and enterprises in order to perform ‘what-if’ scenarios and forecasts. Incorporating these AI approaches can help operators compare asset performance, predict how an action could impact asset health, and determine how various decisions will affect the bottom-line.
DISTRIBUTECH+ will be offering a session called: “How Utilities Are Using Artificial Intelligence for Greater Insight and Improved Decision-making” on March 25. Learn about the session and register to attend at this link.
Power generation companies also have to consider and manage the numerous variables that can impact their production goals and asset health, including environmental changes, fuel cost, maintenance scheduling, and historical operations. Cognitive AI can provide operators with an accurate representation and assessment of the health and capabilities of their assets and operations of their plant, helping them make more informed decisions when considering trade-offs, goals, and constraints at the same time. They do this by combining predictive models (both supervised — learning to recognize and predict specific events and patterns — and unsupervised — learning to detect anomalies and potential events of interest) that are trained from historical data on critical assets and their operations. Such systems are educated and informed by relevant domain expertise captured and digitized from skilled operators’, engineers’, and decision-makers’ knowledge.
AI’s Role in Transitioning to Decentralized Power Generation Systems
The global energy market is transforming from highly centralized power distribution systems to more complex networks of distributed energy resources, including fossil fired power plants, renewables such as solar and wind farms, as well as roof-top solar installations that are putting power back into the grid. Centralized power systems located in remote areas incur large electricity losses, partly due to resistance in electric lines during transmission. Networks of decentralized systems, specifically ones that generate power from renewable resources, have less impact on the environment and produce less harmful emissions, allowing for close proximity from a power source to end-use sites in residential and business areas. This low-carbon shift can be seen across the energy industry from power and utilities to legacy sectors such as oil and gas.
During this transitional period, integrating distributed energy resources has been challenging. However, artificial intelligence solutions, alongside intelligent asset management strategies, can help. AI approaches can model the impact of variables such as an increase to roof-top solar panels or decreases in wind speed, then provide recommendations to operators on how to best remediate the impact. This approach increases grid reliability and provides better maintenance for aging transmission and distribution systems.
Decentralized systems often operate in silos, affecting their ability to detect and account for perturbations that can be introduced by others. Advanced AI can provide the basis to better understand the effects of any one individual’s actions on the stability and reliability of the grid and overall generation capabilities. AI approaches can play a significant role in augmenting, monitoring, and managing capabilities across decentralized resources, leveraging a wealth of current and historical data streams while incorporating the deep industry knowledge required to ensure the adoption and deployment of such AI-based systems.
Digitizing & Distributing Domain Expert Knowledge Across Operations with AI
A plant only operates as well as its leading operator. Across the industry, experienced power generation operators are retiring and taking their knowledge with them. This results in an unfortunate (yet inevitable) knowledge gap. The deep domain knowledge, experience, and expertise of these operators are central to efficient, low risk, and reliable operations. In order to ensure the long-term continuity, efficiency, and reliability of operations across the power generation domain, this subject matter domain expertise needs to be captured, digitized, immortalized, and made accessible across workforces.
Novel hybrid AI approaches such as Cognitive AI represent a necessary avenue for combining the value of data and domain knowledge to tackle numerous challenges currently facing the power generation industry, including the issue of a shrinking global workforce. Cognitive AI technology combines machine learning techniques with encoded human expertise and business logic to solve complex problems using human-like reasoning. The resulting democratization of domain expert knowledge and data provides operators with more reliable recommendations and helps them achieve best practices.
Indeed, AI has the potential to revolutionize the power generation domain, combining historical and real-time operational datasets with embedded deep subject matter expertise to result in explainable, trusted recommendations. Such accessibility is designed to maximize the efficiency and reliability of operations while simultaneously minimizing risk, meeting power generation demands, and achieving financial objectives.
Powering the Future with AI
Artificial intelligence can advance entire operations in the power industry by improving infrastructure and asset monitoring, outage supervision, prediction, and planning. AI systems can also help power generation entities better support dynamic changes to their generation assets and optimize their operations to align with grid forecasts and demands. Such unparalleled insights could yield greater potential for power trading and increased risk mitigation.
Integration and adoption of machine learning and other artificial intelligence solutions are already increasing exponentially. At this pace, the outlook for advanced technologies points towards approaches – like Cognitive AI – becoming the gold standard for the future growth of the power generation and utilities industries.
About Stephen Kwan
Stephen Kwan is Director of Product Management for Power Generation/Grid Management at Beyond Limits, an industrial-grade AI company that builds advanced software solutions for the most demanding sectors, including energy, utilities and healthcare. Stephen has over 25 years of experience in product management and engineering, holding previous positions in the product organizations at OSIsoft and GE Energy. Stephen holds a BS in Engineering Chemistry from SUNY Stony Brook and a Ph.D. in Materials from Pennsylvania State University.