Detection Method of Combustion Instability in Combustion Chamber of Gas Turbine

Abstract

This article mainly studies combustion instability. It analyzed the basic composition of the combustion chamber and gas turbine and the principle of combustion instability. In order to detect the abnormality of the combustion chamber pressure, the paper proposed a model based on the combination of Principal Component Analysis (PCA), Mind Evolutionary Algorithm (MEA) and Back Propagation Neural Network (BPNN). The model uses Pearson's rule and PCA to preprocess the data. Then it optimizes the back propagation neural network through the global optimization capability of MEA to obtain the optimal prediction curve. At last, the paper introduced similarity and set an early warning threshold to detect the limit alarm. After verification on the simulation platform of combined-cycle gas and steam turbine power plants and import real data for simulation, the results showed that the PCA-MEA-BPNN algorithm can handle non-linear problems well and detect abnormal combustion and issue an alarm.