Artificial Immune System-based Machine Fault Diagnosis Technology

Artificial Immune System, which has powerful information processing ability, simulates Biology Immune System. The main function of Biology Immune System provided with the ability of recognizing "self " and "non-self is to detect, damage and kill the antigen from inside or outside body in the real time. Thus, without the transcendent knowledge of abnormal mode, the algorithm of Negative Selection derived from it can recognize the natural (self) or none natural (non-self) mode of the detected objects efficiently. Therefore this algorithm provides the new idea and method to the field of fault diagnosis.In the Negative Selection Algorithm, the detector shall pass through the processing of very complicated training computation and the period of lengthy time. Based on the principles of Vaccine Inoculation in Artificial Immune System and of Clone Selection, this paper illustrates two kinds of Mutation Searching Method, which simplifies the computing complexity in detector training and, combined with the distribution and diversity of immunity, mends the binary-coded mode.Currently there are many shortages, even infeasibility in the application of revolving machine fault diagnosis when applying the Negative Selection Algorithm combined with the data of time domain oscillating wave. Based on many experiments and deep researches, this paper points out the rationality, feasibility and validity when applying the Non-Dimensional-Parameter to diagnose in the Negative Selection Algorithm. Furthermore, based on the characteristics of distribution and diversity in the Artificial Immune System, this paper, by the difference of frequency segment of fault feature, firstly defines three Non-Dimensional-Parameter classes — displacement class, speed class and acceleration class. Secondly, it redefines five kinds of Non-Dimensional-Parameter Immune Detector (Peak Value Parameter Immune Detector, Margin Value Parameter Immune Detector, Impulse Value Parameter Immune Detector, Kurtosis Value Parameter Immune Detector, Waveform Value Parameter Immune Detector) for each class. The results of experiments prove the practicability and validity of this detector definition.Based on the concepts of Immune Evolvement, Immune Learning, Immune Memory and Immune Acknowledge, this paper, by the difference of training processing, defines the Creating Detector, the Mature Detector and the Excellent Detector, corresponding to the process of exclusive fault space detector set obtained...