| The traditional process capability analysis theory is based upon the assumption that the observed data which came from the process is independent mutually. However, in many conditions, the process data is collected automatically with the improvement of equipment and instrument automatization and intellectualization, some process data exists obvious auto-correlative due to the reduction of sampling period. These auto-correlated data can not comply with the assumption that every data is independent which based on the traditional statistic theory.Under these circumstances, the traditional process capability analysis theory becomes invalid.In order to find the best method of process capability analysis for stable auto-correlation process, the basic theory and knowledge of the process capability analysis and the characteristics and identification methods of auto-correlated process are first introduced in this paper. Based on AR(1) model in time series analysis theory, the effect of auto-correlation on process capability analysis is studied by simulation. The results show that the sampling method, sample numbers and derivation estimating methods of standard statistical process control method have great effect on process in the auto-correlation processes, and with the increment of degree of auto-correlated, the effect becomes more and more serious. Based on the correct sampling and derivation estimating method, the relationship between sample numbers and the calculation of process capability index, and the relationship between sample numbers and identification of auto-correlation process is studied by simulation. Compared with the yield result of auto-correlation process, the minimum sample numbers for calculating process capability index is proposed. At last, the problem for calculating process capability index when implementing incontinuous sampling is discussed, and requirement for calculating process capability index using incontinuous sample data is proposed. |
PDF Full Text Request