Studies On Application Of Artificial Neural Network To Superheater's Tube Wall Temperature Calculation And Life Appraisal

With the development of science and technology, the equipment becomes more and more complex, it has become the necessary tendency to increase the reliabilitysafety of superheater operation,predict its remanent lifespan and guide the operation and maintenance of power plant.At present, there is no effective means to measure the temperature of the superheater's tube wall. Aiming at platen superheater of power plant, this paper perfect the general calculation method of heat deviation and wall-temperature based on our scientific researchers' production and "the Soviet standard of thermal calculation method"; then through analyzing its theory, we consider that there is a complicated non-linear function between highest temperature of superheater's tube wall and some parameters such as steam temperature, pressure,steam flux and so on. After introducing the BP N-N theory, we establish the BP N-N model to predict the tube wall temperature by MATLAB and we also discuss something about the model needs to improve.This paper discusses the mechanism of lifetime wastage of high temperature superheater. It indicated that the creep and corrosion and the abrasion are the main reacauses of the life wastage. It considers the operation temperature and the pressure are the main factors of the lapse. This paper introduces several life diagnosis methods. Then it introduces the method which uses neural network model to predict creep-rupture strength based on the actual alloy compositions and creep-rupture strength at different temperatures and times. And it points out the problem awaiting further research in this field. And then, we use Larson-Miller method formula to estimate the platen superheater's life.