GRNN算法在电力系统负荷建模中的应用

Application of GRNN-algorithm on Load Modeling of Power System

针对传统负荷模型不能有效地克服负荷建模中的非线性和不连续性等问题,提出了一种基于广义回归神经网络的负荷建模方法。利用广义回归神经网络具有的全局逼近和最佳逼近能力及网络稳健、快速收敛的优点,建立了新的电力系统综合负荷模型。并与两种改进的反向传播网路模型进行了比较,仿真实例证明了该模型对电力系统负荷模型辨识的有效性和准确性。

Since the traditional load model can not overcome the problems such as nonlinear and discontinuous in load modeling effectively, a load modeling method based on general regression neural networks（GRNN） is proposed. The capacity of universal approximation and the best approximation as well as the merits of steady networks, fast convergence is used in GRNN to establish a new comprehensive load model in power system. Then it is compared with two improved back propagation neural networks. The simulation results prove its effectiveness and accuracy when the proposed model is adopted in power system load identification.