摘要

Distributed generation (DG) output is intermittent, random, and volatile. Consequently, evaluation of the load supply capacity(LSC) of the distribution network becomes more complicated. Owing to the fact that traditional uncertainty methods can not fully reflect the aleatory uncertainty and epistemic uncertainty of LSC, the affine algorithm based step-varied repeated power flow method (AA-SVRPF) was proposed. The parameterized probability box models of wind speed and light intensity were established by using the distribution parameter interval of probability function. The affine algorithm was used to directly obtain the focal elements of LSC, and the Dempster-Shafer structure of maximum load growing percentage was generated based on the evidence theory. Due to multi-layered cycle nesting and the low efficiency, a fast solution strategy based on matrix operation was proposed to improve the calculation efficiency of AA-SVRPF. In order to reflect the uncertainty of LSC under different load levels, the static load supply risk probability was used as the measurement index of LSC. Comparative analysis of the calculation examples indicates that the proposed method is feasible and effective. ?2022 Chin.Soc.for Elec.Eng. 8153.

全文