基于BP神经网络的水轮机调节系统建模与仿真  

BP neural network based-modeling and simulation of hydraulic turbine governing system

作　　者：谢进[1] 陈启卷[1] 李俊益[1] 

机构地区：[1]武汉大学动力与机械学院,湖北武汉430072

出　　处：《水利水电技术》2015年第3期119-122,共4页Water Resources and Hydropower Engineering

基　　金：国家自然科学基金资助项目(51179135)

摘　　要：采用BP神经网络对水轮机综合特性曲线进行处理,将水轮机非线性特性转换为可用于实时仿真的力矩和流量神经网络,可避免建立水轮机特性变量间复杂的函数关系式。将其应用于水轮机调节系统的大波动过渡过程计算,可使计算更为简洁,且由于曲面是光滑的且导数连续,更容易保证计算过程的收敛。建立的机组非线性模型更贴近实际情况,仿真试验表明应用效果良好。Based on BP neural network,the synthetic characteristic curve of hydraulic turbine is processed to transform the nonlinear characteristics of it into the torque and flow neural network that can be applied to the real-time simulation,thus the establishment of the complicated functional relationship among the characteristic variables of the turbine can be avoided. If it is applied to the calculation of the large fluctuation transition process of the hydraulic turbine governing system,the calculation can be made much simplerandeasier,and then the convergence of the calculation can be more easily guaranteedsince the surface is smooth and derivative is continuous. The nonlinear model for hydraulic turbine unit established herein is more close to the actual operation condition of hydraulic turbine and the simulation shows a better application effect.

关 键 词：水轮机调节系统 BP神经网络 水轮机综合特性曲线 

分 类 号：TV734[水利工程—水利水电工程]