Primary superheater spray control valve modeling based on Levenberg-Marquardt learning algorithm

Abstract

Main steam temperature is one of the most important parameters in a coal fired power plant and its characteristics are non-linear and having large inertia with long dead time. Successful control of main steam temperature within ± 2 deg C from its setpoint is the ultimate target for the coal fired power plant operators. Two of the most common main steam temperature circuit are primary superheater spray and secondary superheater spray. This paper present the primary superheater spray control valve modeling based on Levenberg-Marquardt learning algorithm. The neural network algorithm will be trained using actual plant data. The result of the simulation showed that the primary superheater spray control valve modeling based on neural network with Levenberg-Marquardt learning algorithm is able to replicate closely actual plant behavior. Generator output, main steam flow, total spraywater flow and secondary superheater outlet steam temperature are proven to be the main parameters affected the behavior of spray control valve opening in the unit.