基于改进的BP网络的水机温度模型建立

传统的水电仿真系统中的温度模型的构建方法存在模型可移植性比较差、推导过程必须有水电专家的参与、推导过程复杂度大、模型精确度不高等不足,这些都对水电仿真的进一步发展和准确性产生了一定影响。因此需要一种更新的方法来适应水电仿真的发展。为此提出了一种改进的BP网络神经元网络学习算法,通过改进训练算法以提高神经网络的训练效率以及准确度。将这种算法应用于吉林丰满水电厂水电仿真系统的水机温度模型的建立实验中,并与原有的神经网络方法进行比较,比较结果表明,该方法能提高分类准确率和训练速度。

Hydraulic turbine temperature model establishment based on improved BP neutral network

Temperature model design method in traditional water and electricity simulation system is of many disadvantages,such as bad transplantable ability,complex deduction,lower precision and so on.Those disadvantages have a bad effect on veracity and further development of water and electricity simulation system.So that it is necessary to present an improved method to develop a simulation system of water and electricity power.An improved BP neural network algorithm is presented in this paper,which can improve the efficiency and accuracy of the training,then apply this algorithm to establish a hydraulic turbine temperature simulation model of Jilin Fengman's water and electricity power plant, compare to the old neural network algorithm,the results certified that this algorithm can promote classify accuracy and training speed.