Use of Data Analytics for Power Plant Operation Optimization

Abstract

Optimized operation and maintenance of a power plant to improve the KPIs is the key to success. Forced outages not only make the unit unavailable but also increase the number of start-ups and hence high cost and deteriorated plant life. Fault diagnostics and prediction of future behavior of equipment by data analytic techniques will be very handy in improving reliability. On the other hand, the cost of generation depends upon how efficiently the plants are operated. Any deviation from the optimized design point, results in losses and hence as far as possible the deviations should be minimized. However, the deviations are unavoidable due to various reasons. Along with efficient operation, Environmental compliance, and flexibility is also to be ensured to meet the statutory requirements. With the advancement of computing techniques, data analytics, and Artificial Intelligence are being used for the optimization of many industrial processes. Given the background, the objective of this paper is to study and suggest how Data Analytics can be utilized for the optimization of power plant operations to improve reliability, efficiency, and flexibility, and minimize the impact on the environment. Many statistical methods are available to identify the hidden pattern in the data which can be used for optimisation of processes. AI is used for process control, diagnosing faults, and predicting of future behaviors so that advanced action can be taken to avoid surprises. The domain expertise along with data analytic methods can be utilized to find solutions to a variety of problems. In this paper, a comprehensive data analytic tool with four modules i.e., “Efficiency optimization”, “Plant Health monitoring and reporting”, “Optimization of life consumption” and “Environment protection” has been conceived for application. These Modules work in an integrated manner and shall monitor, optimize, control, and report/advise. The tool shall have a digital replica of the individual equipment for simulation individually as well as in combination with other related equipment for whole plant performance prediction and diagnosis. The replica shall use its database for machine learning and for running of diagnostic process.