Comparison of Fuzzy logic and Neural Network in life prediction of boiler tubes

The life prediction of equipments guarantees reliable, safe, efficient and continuous operation of thermal power plants. In this paper wall thickness of reheater tubes of boiler number 3 of Neka power plant in north of Iran are measured at several points during maintenance shut down periods. Thickness dependency vs. time has been investigated. Artificial Neural Network (ANN) as a tool has been acquired to determine this dependency. Extrapolation of thickness function vs. time applying maximum normal stress criteria results in corresponding thickness. The maximum wall reduction rates have been calculated by two schemes namely Fuzzy function (FF) and Neural Network (ANN) applying numerical calculation and Genetic Algorithm. Results have been compared with the existing relevant data from the literature and measured data of the plant in order to determine the accuracy and verify the validity of the methods.