Methods And Application On Attribute Measurement System Analysis

Measurement System Analysis is of vital importance in a Six Sigma project. Attribute measurement system takes the advantage of being practical and efficient. Unfortunately, the study on Attribute MSA is far less than what we expected compared with the Variable MSA and some problems in application need to be solved.Based on the above analysis, such method as proportion of agreement and Kappa statistic is elucidated in detail by being applied in the sealing process of thermostat. It is demonstrated when to use the above method and how to use it correctly. And then some attempts have been made to evaluate the attribute measurement system with the help of some widely used method in other fields. What's more, the proper sampling method has been made for the attribute measurement with programming in the environment of MATLAB. Namely, at least 20 parts should be picked for the Attribute MSA, especially for the measurement system whose proportion of agreement is moderate (e.g. p=50%), more parts than 40 are required. Taking the improper use of the present standard into consideration, an easier and feasible method to evaluate the binary measurement system has been proposed with the aim of lowering the producer's risk as well as the customer's risk. This method turns out to be efficient in the case study, in which the producer and the customer totally reach an agreement on how to evaluate binary measurement system and improve it continuously.Futhermore,a software, streamlining the calculating process of attribute agreement analysis, has been developed, which can easily calculate proportion of agreement, Cohen's and Fleiss's Kappa statistics.