基于径向基神经网络的风电场无功补偿优化算法

 针对风电场无功补偿容量计算工作量大、计算过程复杂的问题，提出了应用径向基神经网络优化风电场无功补偿容量计算的方法。首先建立了含风电场的电力系统潮流计算模型，以某风电场实际有功功率作为模型的输入，计算该风电场所需的无功补偿容量；以有功功率作为输入数据，以计算所得的无功补偿容量作为目标输出，建立径向基神经网络，并对该神经网络进行训练。用训练后的径向基神经网络代替潮流计算模型，对该风电场所需无功功率进行计算，结果表明，该方法计算复杂度比潮流计算模型低，计算量少。研究表明可用训练后的径向基神经网络模型代替潮流计算模型，实时计算风电场无功补偿容量。

Reactive Power Compensation Based on Radial Basis Function Neural Network for Wind Farm Connected to Power System

This paper proposes an optimization algorithm based on radial basis function (RBF) neural network to deal with heavy workload and complex calculation process of wind farm reactive power capacity calculation. First, a model for power flow computation of power systems containing wind farm is established, and the actual active power of a wind farm is taken as the input of the model, to calculate the reactive compensation capacity required. Second, the actual active power of the wind farm is used as input data, and the resulting reactive power compensation capacity as the target output, to establish a RBF neural network and train it. Finally, with the trained RBF neural network replacing the power flow calculation model, the reactive power compensation capacity for the wind farm is calculated. Calculation results show that the computational complexity of RBF neural network model is lower than that of the power flow calculation model, and the workload is reduced. Thus, the RBF neural network model can be trained to replace the power flow calculation model to calculate the reactive power compensation capacity of wind farm in real time.