Power Transformer Fault Diagnosis Based On Artificial Immune

Power transformers belong to the most important components of power generation and transmission system. Artificial immune system is a new intelligent system which is inspired by biological immune, which is a highly parallel adaptive system with capability in learning, memory acquisition, and pattern recognition and so on. This paper analyses the immune algorithm, combining with fault diagnosis of power transformer, proposed some fault diagnosis approach based on the theory. The detail work as follows:(1) An artificial immune classification approach was proposed for fault diagnosis. The interaction mechanism between antigens and B cells was simulated, where fault sample(feature vector consists of ingredients dissolved gas in-oil in transformers)was regarded as antigen. The principle was that, the highest affinity B cells cloned and mutated for diversification and shape-space exploration, then developed into ARB(artificial recognition ball), at the same time, class information was added to ARB so that it was trained to learn the feature of fault samples better. Here, affinity was computed with weighted Euclidean distance in order to take some single abnormal ingredient into account. Subsequently, antigens were presented to ARBs, and ARBs competed for representation more B cell until stimulation value of each ARB reached the threshold. In this way, the ARB pool was generated and served as a classification tool. After training being completed, the classification was performed in a k-nearest neighbor method.(2) Fuzzy immune recognition approach was presented for power transformer fault diagnosis. In ARIS, resources allocation is done linearly with stimulation; this requires excess resource usage in the system, which probability results converging rather prematurely to a state where a few good ARBs overtake too many resources. To get rid of this problem and balance the level of stimulation and suppression, the fuzzy inference system with nonlinear mapping ability is used to improve resources allocation, which shows well-respect classification capability.(3) Compare the performance of artificial immune classification approach and fuzzy immune recognition approach in power transformer fault diagnosis. And Compare the performance of artificial immune and BP neural networks in power transformer fault diagnosis.(4) This paper proved that the proposed transformer fault diagnosis approach based on artificial immune has prospect application with good diagnostic precision.