A New Assorted Approach to Monitor Industrial Processes

Statistical Process Control (SPC) is a multi-discipline technique used to monitor and control the special causes of variation in manufacturing or service processes, since the natural variations are inevitably part of the processes. This bunch of techniques includes check sheet, Pareto chart, cause and effect diagram, control chart etc. Control chart is the most efficient and widely used technique. Series classification of control charts are based on the size of the shift (large or small), the parameter to be measured (location or dispersion) and number of variable of interest (univariate or multivariate). Control charts based on the size of the shift, are classified as memory-less chart and memory control charts. Memory-less control charts (particularly Shewhart-type) are sensitive and more effective in monitoring large shift. They are named memory less because they make use of the present information only, while they operate. On the other hand, memory charts use both past and present information about the process, which makes them more sensitive and effective to small and moderate shifts in the process. Examples of such structures include the Cumulative Sum (CUSUM) and Exponential Weighted Moving Average (EWMA) charts. In this project, we intend to establish a new assorted approach to monitor large, medium and small shift in a single control chart. We will examine the performance of the stated proposal in terms of average run length, standard deviation of run lengths, extra quadratic loss and finally the comparison of its detection ability with its counterparts will be made. Our study will also cover real life application from adhesive manufacturing industry (monitoring the thickness of tape), Electrical processes (Monitoring the charge of Z-source inverter) and semiconductor manufacturing (monitoring the mass flow controller) to explain the significance of the proposal in different disciplines.