Monitoring Sewage Treatment Plants: Some Quality Control Aspects

Abstract

Common features of environmental quality data are serial correlation, seasonality, missing values, nonconstant variance, and nonnormal distributions. These features are found in air and water quality data, in biological and chemical data, and in data from treatment processes as well as natural systems. In this paper, effluent data from two sewage treatment plants are studied, and it is illustrated how complications such as these—in particular serial correlation—affect the use of Shewhart and cusum charts. It is shown how standard charting procedures can be suitably modified so that the usual assumptions are approximately satisfied and a rigorous, quantitative analysis is possible. The modification proposed is to plot residuals from an adequate stochastic model for the process, rather than the data themselves. But, even in unmodified form, the standard charts have their uses in monitoring work. It is demonstrated how the plotting of raw or transformed data on standard quality control charts can provide valuable qualitative information, even though the assumptions that justify the use of these techniques are typically violated by environmental quality data. Such charts may aid in interpreting past system performance or identifying an appropriate stochastic model for the process, and thereby help in improving plant operation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 138 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.