组合式粒子群神经网络的96点负荷建模的应用

Ninety-Six-Point Load Modeling Application Based on Particle Swarm and Neural Network

粒子群算法是一种新型寻优策略,具有收敛速度快、收敛精度高的优点。提出一种改进粒子群神经网络的负荷预测模型,通过改进粒子群算法优化BP神经网络的权值和阈值,改善神经网络的缺陷,将优化好的BP神经网络对某电力系统进行短期负荷预测。仿真结果表明,该算法收敛速度快,网络性能良好,并具有较强的自适应能力。

Particle swarm optimization is a new type of optimization strategy with quick convergence and high accuracy advantages.An improved load forecasting model based on particle swarm and neural network was raised to overcome defects of neural network via optimizing weights value and threshold value of particle swarm optimization algorithm and back propagation（BP） neural network,so as to carry out short-term load forecasting for a certain power system with the optimized BP neural network.Simulation result shows that the algorithm is fast in convergence,fine in network performance with stronger self adaptability.