An effective statistical process control scheme for industrial environmental monitoring

作者：Shamsuzzaman, Mohammad^*; Haridy, Salah; Maged, Ahmed; Bashir, Hamdi; Shamsuzzoha, Ahm; Ali, Ahad

来源：Chemometrics and Intelligent Laboratory Systems, 2022, 229: 104651.

DOI：10.1016/j.chemolab.2022.104651

• Industrial effluent
• Environmental quality monitoring
• Non-normal environmental data
• Statistical process control
• Quality cost
• Manpower cost

One major source of environmental pollution is industrial effluents, especially the heavy metal pollutants, including zinc, in the wastewater discharged from manufacturing industries (hereinafter known as an environ-mental process). The safe disposal of such effluents into aquatic environments is very crucial and of great health and environmental concerns worldwide. To safeguard the quality of the receiving waters, continuous and effective monitoring of zinc in industrial effluents before discharging it into the environment is essential. However, the real-life environmental data are often skewed and do not follow the normal distribution. Therefore, classical statistical process monitoring techniques based on normality assumptions cannot be used directly to monitor the environmental processes. To overcome this limitation, this study designs an optimal X scheme for monitoring non-normal zinc data in industrial effluents. This scheme minimizes the total cost, including quality and manpower costs, based on optimal deployment of manpower. The optimization process ensures that the false alarm rate of the monitoring scheme meets the allowable value. A fractional factorial experiment considering different design specifications shows that the proposed optimal X scheme reduces the associated costs by about 69% and 36% compared to the basic X and improved X schemes, respectively. The robustness of the performance of the optimal X scheme is further investigated under different non-normal, and process shift distributions. A sensitivity analysis is also conducted to explore the impact of the input parameters on the associated cost of the optimal X scheme. Lastly, the design and application of the proposed optimal X scheme are presented using real data on zinc content in the wastewater emitted from a steel production plant.