电力计量数据驱动的电力故障预测方法研究

Research on Power Failure Prediction Method Driven by Power Metering Data

传统的电力计量故障诊断方法过度依赖于人工判断,工作效率偏低。针对此问题,提出一种基于LR优化的Tradaboost算法,并将其应用在电力的故障诊断领域,提升故障预测模型的样本容量及预测准确率。由实验结果可知,LR-Tradaboost算法预测误差相对SVR、Tradaboost-R2均降低了误差率。LR-Tradaboost在迭代次数为50时已经收敛,而Tradaboost-R2在迭代次数100次后才收敛,所以LR-Tradaboost算法的预测性能更好。

The traditional power metering fault diagnosis method relies too much on manual judgment,resulting in low work efficiency.To solve this problem,a Tradaboost algorithm based on LR optimization is proposed and applied to the field of power fault diagnosis,it improves the sample size and prediction accuracy of the fault prediction model.The experimental results show that the prediction error of the LR-Tradaboost algorithm is reduced compared to SVR and Tradaboost-R2.At the same time,it is obtained that LR-Tradaboost has converged when the number of iterations is 50,while Tradaboost-R2 only converges after 100 iterations.Therefore,the prediction performance of the LR-Tradaboost algorithm is better.