论文部分内容阅读
负荷短期预测是电力系统运行和调度每年的重要工作,尤其在市场环境下负荷短期预测更显重要。对于电力系统短期负荷的随机性和突变性的特点,提出了应用遗传算法(GA)和模拟退火(SA)优化的Elman神经网络的短期负荷预测模型。其特点是模型简单、运算效率高,并具有较好的全局最优性能,从而很好地克服了传统BP算法容易陷入局部极小点的缺陷。利用改进Elman网络的良好学习能力,同时利用遗传和模拟退火优化算法对Elman动态递归网络的前馈和反馈值进行优化,实现全局最优的拟合结果。比较了Elman网络和BP网络结构的建模效果,仿真实验证明了利用遗传和模拟退火算法优化的Elman神经网络具有动态特性好、逼近速度快、精度高等特点,说明Elman网络是一种新颖、可靠的负荷预测方法。
Short-term load forecasting is an important annual work of power system operation and dispatching. It is even more important to forecast short-term load especially in the market environment. For short-term load of power system, the short-term load forecasting model based on Elman neural network optimized by genetic algorithm (GA) and simulated annealing (SA) is proposed. It is characterized by simple model, high computational efficiency, and good global optimal performance, so as to overcome the defect that the traditional BP algorithm is apt to fall into a local minimum. The improved learning ability of Elman neural network is improved, and the feedforward and feedback values of Elman dynamic recursive network are optimized by genetic and simulated annealing optimization algorithms to achieve the global optimal fitting result. The simulation results of Elman neural network and BP neural network are compared. The simulation results show that the Elman neural network optimized by genetic algorithm and simulated annealing algorithm has the characteristics of good dynamic characteristics, fast approaching speed and high precision. It shows that Elman network is a new and reliable Load forecasting method.