基于灰色投影随机森林算法的配网故障量预测模型

Distribution Network Fault Prediction Model Based on Random Forest with Grey Relation Projection

科学合理的配网故障抢修管理能及时对配网故障进行抢修,进而减小电力公司损失,提高供电服务质量。本文提出了一种基于灰色投影随机森林算法的配网故障量预测模型。该模型可以根据历史数据来对未来不同区域、不同电压等级的设备故障量进行预测,预测结果可为各单位进行抢修资源合理调配提供依据。该预测模型采用随机森林算法建立预测模型,并利用灰色投影筛选出相似日的样本集合作为训练数据。在真实电力数据上的仿真验证结果表明,该模型的预测效果令人满意。

The scientific and rational distribution repair fault management can achieve timely repair of the rational distribution fault, reduce the loss, and improve service quality. This paper proposes distribution repair fault prediction model based on random forest with grey relation projection. This model can predict the equipment fault rate of future time. The prediction results can be used as the basis of allocation of repair resources. This model is based on random forest algorithm, and it uses grey relation projection to select similar day data set as the training data of random forest. The simulation results on real power data show that, this prediction model can achieve satisfactory prediction results.