The application of artificial immune network in load classification

Characteristic clustering of dynamic loads is necessary for load modeling practicization. This paper presents a novel method of characteristic clustering of dynamic loads based on fuzzy artificial immune network (FaiNet). Firstly, artificial immune network learning algorithm reflects the samples in a compact network. By analyzing the nodes of obtained network with minimal spanning tree, the cluster number and related cluster centers are easily gotten. At last, load samples are sorted on the basis of fuzzy rules. Test results show that this method not only finds the clustering structure effectively, but also is independent of the initial prototypes selection and predefined class number. To compare with fuzzy C-means method, the FaiNet method fit more for characteristic clustering of dynamic loads.