Prompt:
Present literature review on statistical process control highlighting the development of the concept from the time it was developed to-date. You have been provided with historic data relating to temperature of combined effluence discharged by company ABCD. The data provides temperatures recorded four times a day during the month of September 2022. The company’s effluent discharges are usually controlled within the 250C to 350C range. The company (a brewery) usually performs weekly maintenance on the balancing system, whose effect is to neutralise the pH of the effluent and in the process heats up the discharge. The maximum temperature allowed for the discharge is 400C. Using the data, visualise the performance of the company’s effluent control process. Ensure that you describe the analytical approach you have applied and include any graphs you have produced. Based on your analysis/visualisation, how well has the effluent control process performed? What priorities should the company adopt for quality improvements? – This is a statistical process control case study.
Data Sample: 27.508, 33.19, 30.06, 30.01 … (120 data points in total).
Statistical Process Control Case Study
Introduction
Statistical process control (SPC) is the application of statistical methods to monitor, control, maintain, and improve the performance of a process (Jiju & Mehmet, 2003; Jamadar (2020). SPC is also a tool that applies time series plots, for checking whether products or processes confirm to their design requirements (Qiu, 2014). SPC, a brainchild of Dr. Walter Shewhart, was first developed in the 1920s as a tool for monitoring and controlling manufacturing processes (Best & Neuhauser, 2006). The tool’s versatility was recognized by Shewhart and Dr. Edwards Deming, who acknowledged that repeated measurements would exhibit some variation. Shewhart later realized that the tool had the potential for use in other types of processes in addition to manufacturing (Best & Neuhauser, 2006; Niavand & Tajeri, 2014). True to this finding, in contemporary times, SPC extends its application to diverse contexts such as management, health quality assurance and improvement, survey, among countless other settings (Jin et al., 2019; Qiu, 2019).
Common Cause Variation and Special Cause Variation
In statistical process control, common cause variation and special cause variation are crucial concepts (Montgomery, 2009; Qiu, 2014). Common cause variation refers to the observed variation that come into being as a result of random fluctuations, representing the natural variability or “background noise” inherent in a process (Montgomery, 2009). This variation is anticipated based on the underlying distribution, assuming variables remain constant with the passage of time. When common cause variation is present, the process is considered to be naturally stable and predictable, and the process is considered to be in “in statistical control” or simply, “in control” (Qiu, 2014). A stable process exhibits predictable variation described by a statistical distribution, such as normal, Poisson, geometric, or binomial distributions. In a process that follows the normal distribution, normally about 95% of future measurements are expected to fall within +/- 2 standard deviations of the mean. Furthermore, regardless of the statistical distribution, almost all measurements are expected to fall within +/- 3 standard deviations about the mean, when the process is in control (Benneyan et al., 2003; Goedhart & Woodall, 2022).
On the other hand, special cause variation represents observed variation beyond what can be attributed to chance alone. This kind of variation results from external factors or special causes, constituting unnatural variation due to circumstances, changes, or events that are not naturally part of the regular process (Carroll & Johnson, 2020). In contrast to traditional hypothesis testing, where special cause variation is analogous to statistically significant differences, SPC identifies changes graphically over time and often involves the collection of a (relatively) few samples (Benneyan et al., 2003; Montgomery, 2009). Continue reading …