水电机组效率的神经网络算法

水电站优化运行要求具有连续水位和导叶开度下的效率特性曲线。对此提出用BP神经网络方法计算机组效率，并建立了BP神经网络模型，以现场效率试验数据作为样本进行训练，并用训练好的网络计算该机组的效率。网络的训练速度及计算结果表明，该算法收敛速度较快，精度高，为计算水电站任意水头及导叶开度下的机组效率提供了新思路和新方法，可用于指导水电机组的优化运行。

Artificial Neural Network Algorithm for Calculating Efficiency of Water Turbine Generator Unit

The efficiency curve of a water turbine under continual water head and guide vane angle is needed to optimize the dispatching of a water power station. A BP neural network model and leaming algorithm are proposed. First, the BP neural network is trained with typical onsite efficiency test data of a turbine generator unit, then the trained BP neural network is applied to calculate the efficiency of the unit. The training speed and the simulation results show that the BP neural network method has fast convergence speed and high precision, it can calculate the characteristic curve at any head and guide vane angle and be used to guide the optimal dispatching of units.