Study on Danger Theory Condition Monitoring Algorithm Adapted to Mechanical System

摘要

According to biological immune danger theory, a condition monitoring algorithm based on immune danger theory was presented. In the algorithm, the whole diagnosis feature space is divided into danger feature space, normal feature space and abnormal feature space. The algorithm consists of primary testing module, APC testing module and self-adapting testing module. The algorithm can make judgment according to whether existing danger signals. So the algorithm can reduce false rate and adjust databases online. The algorithm was applied to axle driving of Farm machinery condition monitoring. Compared with testing result of advanced negative selection algorithm which based on self-nonself recognition, the testing result has a lower false rate.