A Concurrent Simulation Framework of Power Plant for Online Fuel Analysis Based on GRNN Neural Network.

A concurrent simulation framework is proposed for biomass-coal blending combustion. General Regression Neural Network (GRNN) is employed as a fast 'black box' model for biomass property modeling. Higher Heat value (HHV) and fractions of C,H,O,N,S are selected as output and input parameters and the training is based on data collected from literatures. Then the model is combined with simulator of power plant to simulate operation process concurrently. Experiments on a 750 MW and a 550MW power plant system show that the framework can provide fast and flexible simulation of a batch of operation process concerning fuel changing, whilst keeping the computing accuracy at an acceptable level.