Negative selection principle based on the natural immune system makes it possible to develop a new fault diagnosis technique. This paper investigates a fault diagnosis method for the gas valve of the reciprocating compressor based on the negative selection algorithm. The negative selection algorithm combining the clonal selection principle with the mutation theory is used to produce the detectors representing one fault by the sample indicating normal. The detectors representing abnormal are used to perform the fault detection and recognition. The preprocessing using discrete fourier transform to extract the characteristic frequency interval makes the method more efficient and reliable. Experimental data of three kinds of fault from the vibration signal of the gas valve of the reciprocating compressor is used to test the method. The results prove that this method is very efficient.

This content is only available via PDF.
You do not currently have access to this content.