基于改进PCA算法的火电机组供电煤耗负荷特性建模分析

Modeling and Analysis of Power Supply Coal Consumption-Load Characteristics of Thermal Power Unit Based on Improved PCA Algorithm

针对火电厂机组供电煤耗较高、分析精度较差的问题,研究基于改进主成分分析(PCA)算法的火电机组供电煤耗—负荷特性建模分析方法。基于火电机组阀点效应的煤耗—出力函数,充分考虑火电机组在平稳负荷出力、升负荷出力、降负荷出力时的负荷特性,构建火电机组供电煤耗—负荷特性模型,以PCA为基础加入遗传神经网络(GABP),生成改进PCA算法,通过PCA降低网络输入变量冗余,增强学习效率,再利用GABP实现神经网络权值的优化,训练神经网络得出最优解,实现火电机组供电煤耗—负荷特性模型分析求解。试验结果表明:该方法分析获取的供电煤耗率低,节能效果较好,平均总煤耗成本为10.25万元,可精准地显示火电机组的实际运行情况,为降低火电机组供电煤耗提供理论依据。

Aiming at the problems of high coal consumption and poor analysis accuracy of power supply units in thermal power plants,a method for modeling and analysis of power supply coal consumption-load characteristics of thermal power units based on the improved PCA algorithm is studied.Based on the coal consumption-output function of the valve point effect of the thermal power unit,the load characteristics of the thermal power unit under stable load output,increasing load output and decreasing load output are fully considered.The coal consumption-load characteristic model of thermal power units is constructed,and the Genetic Algorithms Back Propagation(GABP) is added on the basis of Principal Component Analysis(PCA) to generate an improved PCA algorithm.PCA was used to reduce the redundancy of network input variables and enhance the learning efficiency.GABP was used to optimize the weight of the neural network,and the optimal solution was obtained by training the neural network to analyze and solve the coal consumption-load characteristic model of power supply for thermal power units.The experimental results show that the coal consumption rate of power supply analyzed by this method is low,and the energy saving effect is good.The average total coal consumption cost is 102,500 yuan,which can accurately display the actual operation of thermal power units and provide a theoretical basis for reducing power supply coal consumption of thermal power units.