电能质量的群体综合评估方法与可靠性优化

Power quality comprehensive evaluation based on group decision-making and reliability optimization

电能质量综合评估对改善电力系统供用电质量具有重要意义。在传统的电能质量评价指标基础上,针对现代电力负荷对电能质量要求的新变化,引入IEC 61000-2-8标准中规定的系统平均有效值波动频率指标来评估电压暂降与短时间中断这类暂态电压扰动现象。在电能质量综合评估方法方面,鉴于现有研究主要集中在个体评估层面,提出了基于相对熵集结模型的电能质量群体综合评估方法。首先由不同电力专家分别给出各待评估监测点的电能质量综合评价值向量,然后依据相对熵的概念及其优化原理建立了电能质量群体综合评估的相对熵集结非线性规划模型,通过求解该模型得到与所有个体评估结果具有最高一致性的群体综合评估结果。之后,为了使得到的电能质量群体综合评估结果具有较高的可靠性,基于熵的思想对该结论进行可靠性分析,并构造了可靠性量化指标。对于不满足可靠性水平的群体综合评估结果,需要不断调整各电力专家的个体评估结果,最终使群体综合评估结果的可靠性达到满意程度。最后,采用某地区10 kV配电系统中的电能质量监测数据,对所提出的方法进行了验证。

Power quality comprehensive evaluation is very useful for improving power quality in power systems. Based on existing power quality evaluation indices, an index named system average root mean square variation frequency index （SARFI） is employed to evaluate voltage sag and shortterm interruption in accordance with the IEC 6100028 stand ard, which aims at satisfying the demands of modern loads on power quality. The existing comprehensive power quality evaluation methods mainly focus on individual decisionmaking. Given this background, a group decisionmaking method based on the relative entropy aggregation model is proposed to comprehensively evaluate power quality in this paper. The comprehensive evaluation values of power quality at all specified monitoring points are respectively determined by all indi vidual experts. The relative entropy aggregation model is established for comprehensively evaluating power quality, and is formulated as a nonlinear programming problem. The final group decisionmaking result, which shows the highest consis tency with all individual decisionmaking results, is obtained by solving this nonlinear programming model. Then, relia bility analysis on the group decisionmaking results is performed based on the decision entropy, and a quantitative relia bility index is proposed. If the reliability requirement of the group decisionmaking results is not matching, the individual comprehensive evaluation results should be adjusted for assuring that the specified reliability requirement is matching. Fi nally, a distribution system with a 10 kV voltage level is served for demonstrating the developed method.