Neural network-based modeling for a large-scale power plant

A large-scale power plant, specifically, a 500MW, once-through type, super-critical boiler plant, requires investigation for the development of a control system. Using data from the power plant, a model can be realized using intelligent techniques. In this paper, a Neural Network-based Model (NNM) is presented as an alternative methodology to expand the modeling techniques for developing a new power plant. The developed Neural Network-based Combined Model (NNCM) consists of many processes which include air/flue gas, pulverizer, water/steam, and turbine/generator systems. The major inputs/outputs of the processes will be mass flow rate, temperature, pressure, and enthalpy of fluid. Moreover, control variables are utilized for driving the plant to desired states. For validation of the proposed model, a comparison of Rankine cycles between actual data and the output of the NNCM will be shown. The results of the NNCM will also be compared to actual plant data for major outputs.