Research on capability evaluation method for multi-stage measurement system

In the process of measurement system capability evaluation, the multi-stage measurement system that applies for a continuous stage of measurement of the same products with different quality characteristics is of great importance. The multi-stage multivariate measurement system model based on one-factor measurement system analysis (MSA) theory and the state space model for the multistage manufacturing process is built in this paper. The new model identifies the variance of local stage that is affected by upstream stage in the measurement system. After using partial least squares regression (PLSR) algorithm to eliminate correlation between the measurement system stages, one factor variance analysis and P/T %, R&R % indexes are used to measure the capability of system analysis. This paper proves its advantages with a case study in which four stages of the single variable measurement system capability analysis is carried out. Further simulations are also carried out to validate the model under four different conditions.