电能质量综合评估方法比较

Comparison of power quality comprehensive evaluation methods

对目前几种典型电能质量综合评估方法:概率统计与矢量代数方法、模糊数学方法、物元分析法、层次分析法、人工神经网络法和遗传投影寻踪方法进行了分类,并对其评估原理及流程进行了系统的总结。利用同一组电能质量测试数据对各种评估方法进行了实例比较分析,分析说明概率统计与矢量代数、模糊数学方法、物元分析法均不同程度地存在人为主观因素的影响,评估结果具有较大的不确定性。而人工神经网络法和遗传投影寻踪法实现了评估过程的客观性,具有较高的可信度。最后,对电能质量综合评估工作的研究方向进行了探讨。

The methods of PQCE（Power Quality Comprehensive Evaluation） are classified, including probability and vector algebra, fuzzy mathematics, matter- element, analytic hierarchy process, artificial neutral network and genetic projection pursuit. Their theories and flowcharts are summarized systematically. The comparative analysis is carried out based on same measured data,which shows that,there is uncertainty in the evaluation results of the probability and vector algebra method, fuzzy mathematics method,matter- element method,while the methods of artificial neutral network and genetic projection pursuit realize the objectivity of evaluation process with higher reliability. The direction of PQCE research is also discussed.