摘要
针对电动汽车充电站充电功率随机性大的难题,本文建立了ELMAN反馈式神经网络预测模型,然后采用粒子群算法对其进行优化,接着将优化之后的模型与模糊控制相结合,最终建立3种模型相结合的组合预测模型,并以青岛地区某一充电站的实际负荷数据为算例,验证了组合预测模型的有效性,提高了电动汽车充电站短期负荷预测的精度。
The electrical vehicle charging station has a problem of big randomness. In order to solve this problem, this article sets up the ELMAN neural network prediction model. Then this article optimizes the model by Particle Swarm optimization. Then this article combines the optimized model with fuzzy control, and sets up a combined prediction model based on three models. This article collects the real load data of a electrical vehicle charging station in Qingdao. Lastly, results show that the above combined prediction model is effective, and this combined prediction model can improve the predict accuracy.
出处
《电气技术》
2017年第2期59-64,共6页
Electrical Engineering
关键词
ELMAN神经网络
粒子群算法
模糊控制
组合预测模型
短期负荷预测
ELMAN neural network
particle swarm optimization
fuzzy control
combined prediction model
short term load forecasting