基于k-means聚类和变分位鲁棒极限学习机的短期负荷预测方法

Short-term Load Forecasting Method Based on k-means Clustering and Varied Quantile Outlier Robust Extreme Learning Machine

随着售电侧市场的逐步开放,集中式的供售电模式被打破,为获取更精确的区域短期负荷预测值,提出一种基于k-means聚类和变分位鲁棒极限学习机的短期负荷预测方法。首先利用传统的k-means聚类算法对历史电力负荷数据进行负荷模式的提取,获取相同用电行为的用户负荷曲线。然后采用变分位鲁棒极限学习机对不同类负荷曲线分别建立预测模型,最后叠加单个的预测值形成最终的预测结果。通过设定不同的分位值来模拟不同的预测场景,以此得到所有可能性的预测值,即实现变分位-多场景的VQR-ORELM灵活预测。为验证所提方法的有效性,采用2个实际案例进行仿真分析。结果表明,相对于支持向量机、BP神经网络、极限学习机模型、鲁棒极限学习机模型,所提模型在聚类前后预测精度始终最高,进一步验证了所提方法的优越性和灵活性。通过k-means聚类后,所有模型预测性能都有较大提高。

With the development of power sales market,the centralized electricity trading pattern is broken up.In order to obtain more accurate results of short-term regional load forecasting,this paper proposes a short-term load forecasting method by combining kmeans clustering method with varied quantile outlier robust extreme learning machine(VQR-ORELM).Firstly,the k-means clustering algorithm is used to extract different load patterns from historical load data.Then VQR-ORELM is used to establish the forecasting mod.el for all load patterns respectively,obtaining final load prediction result by means of the superposition of individual forecasting results.The different prediction scenarios are simulated by setting different quantile to obtain the predicted values of all possibilities,thus realiz.ing the flexible forecasting of varied quantile and varied scenes.To verify the effectiveness of the proposed method,the simulation analy.sis on two cases is carried out.The results show that the proposed model outperforms all other models in predicting the short-term load with and without clustering all the time,such as support vector machine,BP neural network,extreme learning machine and outlier ro.bust extreme learning machine,further verifying the superiority and flexibility of the proposed method.In the meanwhile,there is a no.ticeable improvement in the prediction accuracy of all these models after k-means clustering.