Measurement System Error Analysis

Quality starts from data. In Quality management, no matter using SPC (Statistic ProcessControl), or applying DOE(Design of Experiment), or any other optimizing process tocontrol quality, it all depends on the data from processes to analyze. Hence, quality ofmeasurement data is of critical importance to the management practice. In order to get dataof high quality, it is necessary to separating and analyze the varied error source of themeasurement system.This thesis is based on the research into the error of measurement system. It comparesMSA(Measurement System Analysis), theory of measurement uncertainty and the erroranalysis theory of measurement quality engineering (Taguchi), thus seeks their internalrelation and tries to improve the method system of measurement system error analysis.This thesis firstly studies MSA, theory of measurement uncertainty and measurementquality engineering of Taguchi characteristic. On the foundation of the above study, wecompare and further study the three, getting one's advantage to make up for another'sdisadvantage, thus able to apply them comprehensively. We find the internal relationship ofthe three theories, establish means of evaluating measurement uncertainty based on MSA,and able to get measurement uncertainty in dynamic measurement activities fromSN(Signal to Noise) ratio, solving the problem of getting measurement uncertainty at anypoint during the measurement process.This thesis also studies the methods to compare different measurement systems, both ontheory and practice. Based on MSA and measurement quality engineering, it gives themethod of GR&R and through evaluating SN ratio. Further more, this thesis introducesANOVA method into the comparing, providing new means to compare capabilities ofdifferent measurement systems. When the target to be measured is vari-index, this thesisalso gives the method of MTS to compare different measurement systems.