Transformer oil Dissolved Gas Analysis in the DGA(Dissolved Gas Analysis) is an important method for transformer fault diagnosis. According to the characteristics of transformer faults

an adaptive immune genetic algorithm is proposed for on-line fault diagnosis. The algorithm combines the traditional immune algorithm with the constraint independent component analysis c ICA

and takes the prior information of the object as a constraint condition

so that the new algorithm only converges to the fault signal of interest

faster and more efficient classification. Promote the number of antibody populations and update probability to make the antibody library more effective. The algorithm is applied to DGA power transformer data analysis

and the transformer fault on-line diagnosis is realized. Experimental data shows that this algorithm can effectively identify the samples and improve the pertinence and effect of fault diagnosis.