A Model For Detection and Diagnosis of Fault Based on Artificial Immune Theory

An immune learning algorithm using feature vector code is developed to solve problems about anomaly detection. The antigens input are classified as self pattern code (the first kind of antigens) and non-self pattern code (the second kind of antigens). The first kind of antigens is used to generate randomly initial antibodies according to negative selection principle. The second kind of antigens is regarded as learning stylebook of the immune system. Regarding taking the set of each era antibodies mutated in the system learning as a random series, the condition of convergence of the series and a proof are presented. The astringency of the algorithm is proved. The experimental result indicates that the algorithm can realize optimization to distribution situation of the antibodies and clustering of data modes. High veracity of anomaly detection is obtained