| Real-time diagnosis of abnormal variations in processing has always been a challenging task for most manufacturers with continuous/contiguous processing sequences. Despite the rapid advancements in technology in quality monitoring systems, attaining quality as per customer requirements is still a predominant issue for these companies. Hence, there is an urgent need for development of a new and better method for modeling and monitoring of abnormal variations to optimize quality costs in production, and consequently improve profits. This research study is a part in response to this call. A dynamic quality control system was developed, by integrating and enabling three key concepts, namely---concept of bias separation and estimation through structural equations, concept of variance tolerancing, and concept of dynamic control limits. According to this novel system, the changes observed in prior processes update process averages and control limits of the current process through structural relationships that link those two stages. By applying the concept of variance tolerancing, a new set of control limits are obtained, which are functions of dynamic process averages and variances.;C.L = f(muy,sigma y) +/- kf(muy,sigma y).;As a case study, this novel concept is applied to ring spinning and continuous dyeing processes that are serially connected with time lags. Structural equations based on key process parameters---mass variation and amount of dye absorbed for ring spinning, and continuous dyeing processes respectively have been scouted-out from literature. By consolidating and tolerancing these structural equations, a set of dynamic control limits are obtained. Using these accurate control limits, the root causes of the out of control situations can be determined precisely, and unnecessary corrective actions that are detrimental to quality of the output product can be minimized. |
