Different black-start initial plans may be required for specific power systems, so it is necessary to make decisions and select reasonable black-start plans to ensure that the power system can be quickly and safely restored in the event of a power outage. To address this issue, this study aims to propose a novel black-start schemes evaluation method called genetic algorithm(GA)–self-organizing feature mapping neural network (SOM), which combines the GA and SOM comparison strategy.
First, the index weights are considered as variables, and a fitness function with minimum scheme similarity is established. The GA is used to obtain the optimal index weights when it terminates. Then, a comparison strategy for black-start evaluation methods based on SOM clustering is designed. The SOM clustering algorithm is used to cluster the black-start schemes, and the consistency between the ranking vector and the clustering results is analyzed to measure the accuracy of black-start evaluation methods.
Experimental results indicate that the GA weight determination method enhances the accuracy of the black-start evaluation, and the SOM comparison strategy effectively realizes the quantitative comparison among different evaluation methods.
The proposed method is novel in using GA to determine the index weights and for verifying the accuracy of black-start evaluation methods based on SOM comparison strategy.
