A feasibility study on using neural networks in performance analysis of coal-fired power plants

Coal-fired power plants are highly complex nonlinear systems. Several performance-monitoring techniques based on linearization and empirical estimations have been developed. However, there is a need for nonlinear modeling for the power plant performance analysis in order to meet the growing demands of economic and operational requirements. In the present research neural networks are used to model the thermodynamic process of a coal-fired power plant, based on actual plant data and simulated data obtained from mathematical models that provide information that is currently not directly available. A sensitivity analysis study is performed to determine the effect of various plant variables on an essential performance parameter, namely, coal flow rate. The safe operation of a coal-fired power plant also requires correct operation of plant instrumentation. Failed instruments provide inaccurate information on the state of a process, which can lead to undesirable or inefficient operation of the power plant. Artificial neural networks are used to develop the analytical redundancy to infer the state of important plant parameters. A sensitivity analysis study is performed to determine the critical parameters influencing the estimated plant parameters.