Computer image processing and neural network technology for boiler thermal energy diagnosis

Abstract

The paper uses the flame image processing technology to diagnose the furnace flame combustion achieve the measurement of boiler heat energy. The paper obtains the combustion image of the flame image processing system, and extracts the flame image characteristics of the boiler thermal energy diagnosis, constructs the neural network model of the boiler thermal energy diagnosis, and trains and tests the extracted flame image feature parameter values as the input of the neural network. A rough diagnosis of the boiler?s thermal energy is obtained while predicting the state of combustion. According to the research results, a boiler thermal energy diagnosis system was designed and tested on the boiler of 200 MW unit. The experimental results confirmed the applicability of the system, which can realize on-line monitoring of boiler heat energy and evaluate the combustion situation.