基于神经网络的并行差分方程综合负荷模型

Parallel Difference Equations Load Model Based on Neural Network

为准确地建立电子系统负荷模型,结合Elman反馈型神经网络并行结构的特点,构造了一个并行差分方程动态负荷模型。该模型克服了传统差分方程负荷模型输出无耦合的缺点,更加符合电力系统综合负荷的实际情况。把此模型应用于实际变电站的综合负荷建模,分别用两组数据进行了模型辨识及验证,结果表明所采用的模型是有效的。

In order to establish load model of power system accurately, a new parallel difference equation dynamic load model is constructed combined with the structure characteristics of Elman artificial neural network（ANN）. The new model overcomes the limitations of tranditional difference equation load model output and realizes the ingenious couple of active power and reactive power. The modeling instance to field measured date of power transformer substation shows that the proposed model can effectively describe the composite power load characteristics and is a composite load model with excellent performance.