Diagnosis for Systematic Errors Using Grey System Theory

It is very important to diagnose systematic errors for improving working performance of manufacture systems. At present, many methods of diagnosis for systematic errors require the certain probability distribution and a great deal of data, primarily based on statistics. Therefore a method using grey relational analysis is proposed to resolve the problem. This method can diagnose systematic errors only with small sample, without special requirements for probability distribution. The concepts, grey confidence level, grey difference, weighting coefficient and weighting function mapping, are defined to test reliability of the diagnosis results. This method is validated by computer simulation and engineering experiments. And the grey confidence level is proved to be 95%.