Fast computation of voltage/VAR feasible boundaries of wind farms: An adaptive parameter aggregation dimensionality reduction equivalence method

Efficient voltage/VAR feasible boundary (VVFB) assessments for large-scale wind farms can help reduce cascading trip risks. An accurate VVFB assessment result for such large-scale wind farms, which is essentially a transient security constrained optimal power flow (TSCOPF) problem, may involve dynamic characteristics of all wind turbines in a whole wind farm, and the TSCOPF scale of VVFB assessment is quite huge and difficult to use directly for online computations. Therefore, this paper proposes an adaptive parameter aggregation dimensionality reduction equivalence (APA-DRE) method to efficiently and accurately solve the VVFB problem. First, the core parameters are combined into an adjacency matrix as an input aggregation sample based on automatic weighting factors. The proposed between-within proportion index is used to evaluate the aggregation result and determine the optimal aggregation number of input samples. Then, the original VVFB model is accurately transformed into a small-scale problem represented by equivalent wind turbines with optimal numbers based on the principle of equal voltage differences. Finally, actual wind farm results validate that the proposed approach reduces the computation scale of the VVFB and improves the computation efficiency of the original model by approximately three times while ensuring accuracy.

Automated summary

Efficient voltage/VAR feasible boundary (VVFB) assessments for large-scale wind farms can help reduce cascading trip risks. This paper proposes an adaptive parameter aggregation dimensionality reduction equivalence (APA-DRE) method to efficiently and accurately solve the VVFB problem. The proposed approach reduces the computation scale of the VVFB and improves the computation efficiency of the original model by approximately three times while ensuring accuracy.