Optimization of a Process System in Nuclear Power Plant- A Data Mining Approach

Nuclear Power generation is a capital intensive process requiring continual efforts to reduce cost burdens to stay competitive in the market with many stipulations from statuary, legal and regulatory agencies. This necessitates more emphasis towards safety, efficiency, availability and reliability during operation. Operating the power plant processes inefficiently can result in a low capacity factor of the plant with high significant costs. It will also accelerate the ageing process of equipment and systems. Hence, monitoring the condition of the process, estimating critical parameters based on the variations seen in other parameters and their optimal control is a vital part of today’s plant operation and control schemes. Data science has emerged as a major technology player in implementing concepts like Industry 4.0 and Internet of things. Basic principles of mechanics and process systems remained the same but the efficiency, productivity and reliability were improved immensely with this. Data mining concepts were widely accepted in all the areas since addition of an intelligence factor can optimize any system independent of industry, Data Mining has found promising scope in process optimization and performance improvements in nuclear power plants. This paper deals with application of data mining techniques for a representative process in a nuclear power plant to identify various steady state operating regimes, detect anomalous operating conditions, identify key performance indicators (KPIs) so that steady state estimation of KPIs can be carried out subsequently. This will prompt the operation personnel to adhere to the optimum region of operation and estimate the process conditions for enhanced performance. In addition, this study will also enhance the scope for implementation of predictive maintenance concepts and hence an improved reliability.