基于自适应模糊神经网络的水轮机特性辨识研究

Research on identification of hydraulic turbine model based on adapting fuzzy neural networks

利用自适应模糊神经网络(ANFIS)较强的非线性逼近能力,建立了辨识模型,对水轮机非线性特性进行了辨识.训练算法采用最小二乘和梯度下降结合的算法来训练参数,模型能很好地辨识水轮机特性,并有一定的透明性,为研究智能水轮发电机控制策略提供了有效的建模方法.

The identifying model of hydraulic turbine based on ANFIS neural networks are established by using the strong approaching ability of ANFIS network. In the design, parameters are trained according to minimization principle and steepest descent method. The designed model can well distinguish the characteristics of hydraulic turbine and is transparent to express the relation between input and output. Thus, the identifying model can lay the good foundation for the study on the intelligent control strategies for hydraulic turbine governor.