摘要
针对电力负荷具有的非平稳、随机性、不确定性的特点,提出用EMD-BP神经网络方法对电力负荷进行预测,通过EMD方法将非平稳、随机的电力负荷数据转换成平稳、确定性数据,之后利用BP神经网络进行电力负荷预测。通过仿真试验可以看出,相比于直接使用BP神经网络进行预测,EMD-BP神经网络的预测精度更高、相对误差较小。
Considering power load's unstability and randomness and uncertainty,adopting EMD-BP neural network method to forecast power load was proposed,in which,the EMD can transform non-stationary random power load data into a smooth deterministic data and then makes use of the BP neural network to predict power load. Simulation test shows that as compared to applying BP neural network in the prediction,the EMD-BP neural network has higher prediction accuracy and smaller relative error.
出处
《化工自动化及仪表》
CAS
2016年第3期305-307,332,共4页
Control and Instruments in Chemical Industry