基于BP神经网络的超临界火电单元机组负荷系统建模研究

Supercritical Thermal Power Unit Load System Modeling Research Based on BP Neural Network

针对火电单元机组大范围变负荷、变工况发生时,超临界火电单元机组的控制品质变差,机组负荷、主蒸汽压力等参数难以满足工程需要的问题,基于较为先进的BP神经网络建模方法,对600 MW超临界火电单元机组进行了数学模型的建立。仿真结果表明,网络的输出值与实际模型的输出值间的误差在允许范围内,BP神经网络可有效逼近超临界火电单元机组模型。

When the thermal power unit occurs large-scale variable load and variable operating conditions,the control quality of supercritical thermal power unit will become worse, and the parameters as unit load and main steam pressure can not meet the needs of engineering. This paper establishes the mathematical model of 600 MW supercri- tical thermal power unit based on the advanced BP neural network modeling method. The simulation results show that the error between the output value of the network and the output value of the actual model is within the allowable range, and the BP neural network can effectively approximate model of the supercritical thermal power unit.