Comprehensive mitigation strategy of voltage sag based on sensitive load clustering

Summary To reduce the investment cost of voltage sag mitigation equipment and the probability of voltage sag at the installation nodes of sensitive load, a comprehensive voltage sag mitigation strategy based on sensitive load clustering is proposed. The Monte Carlo method is used to get the stochastic estimation model of voltage sag to determine the lines with high fault probability, so as to get the weak links in the power grid, that is, the voltage sag vulnerable area. This paper proposes four sensitive indicators called tolerance magnitude of voltage sag, endurance duration of voltage sag, equipment capacity, and power dependence. The weight of each indicator is determined by the combination weighting method, and the affinity propagation (AP) clustering algorithm is used to classify different sensitive loads. The sensitive loads are mitigated by installing the energy storage equipment and dynamic voltage restorer (DVR), and the energy storage capacity is configured with the goal of minimum life cycle cost under the premise of voltage compensation. The result using MATLAB/Simulink and the laboratory equipment shows that, compared with existing mitigation strategies, the proposed mitigation strategy can not only effectively reduce the investment cost and the probability of voltage sag but also make the voltage distribution more reasonable. To reduce the investment cost of voltage sag mitigation equipment and the probability of voltage sag, we proposed a comprehensive voltage sag mitigation strategy based on sensitive load clustering. Firstly, the line of voltage sag and the duration of voltage sag are selected as random variables, and the estimation model of voltage sag is established by Monte Carlo method. Four sensitive indicators, namely tolerance magnitude of voltage sag, endurance duration of voltage sag, equipment capacity and electrical dependence are proposed. The weights of each indicator are determined by the combination weighting method, and the Euclidean Distance is calculated according to the weight of each indicator. The AP clustering algorithm is used to classify different sensitive loads. According to the clustering results of sensitive loads, one kind of load is configured with energy storage to mitigate, and the other is mitigated by DVR. When determining the capacity of energy storage configuration, the objective function is to minimize the life cycle cost, and the GA is used to optimize the allocation. The results of numerical examples