| 摘 要: 为提高短期风力发电功率预测的精度，经过对比选择了内核极限学习机(Kernel Extreme Learning Machine，KELM)预测模型的原始模型，对该模型的内部参数进行研究，选择了多元宇宙优化算法(Multi-Verse Optimizer, MVO)对其参数进行优化。还提出用鲸鱼优化算法(Whale Optimization Algorithm, WOA)来给MOV初始化种群，以使MVO算法更不容易陷入局部最优，从而有着更好的求解能力。通过此预测模型进行发电功率预测，获得一个均方根误差(RMSE)值为0.0032、平均预测误差为0.00033的预测结果，最后进行对比实验验证其具有较好的预测效果。 |
| 关键词: 发电量预测；多元宇宙优化算法；鲸鱼优化算法；内核极限学习机 |
|
中图分类号: TP301
文献标识码: A
|
|
| Short-term Wind Power Prediction based on Optimized Kernel Extreme Learning Machine |
|
XUAN Chang
|
(Sussex School of Artificial Intelligence, Zhejiang Gongshang University, Hangzhou 310018, China)
xuan_chang@qq.com
|
| Abstract: In order to improve the accuracy of short-term wind power prediction, this paper proposes to use the original model of Kernel Extreme Learning Machine (KELM) prediction model and study its internal parameters after comparison. Multi-Verse Optimizer (MVO) is selected to optimize its parameters. This paper also proposes to use Whale Optimization Algorithm (WOA) to initialize MOV population, so that MVO is much less likely to fall into local optimum, and thus it has better solving ability. The proposed prediction model is used to predict a power generation, and obtain a prediction result of RMSE value of 0.0032 and average prediction error of 0.00033. Finally, a comparative experiment verifies its good prediction effect. |
| Keywords: power generation prediction; multi-verse optimizer; whale optimization algorithm; kernel extreme learning machine |
