Evaluation of the performance of rural wastewater treatment plants using chemical measurements in combination with statistical techniques: a case study

The performance of a wastewater treatment plant was assessed statistically using multivariate cluster and principle component analysis. This was after measuring some physico-chemical properties in the influent, effluent, downstream, and upstream waters over a 4-month period. The cluster analysis grouped the sampling sites into three clusters: relatively non-polluted (upstream), medium polluted (downstream), and polluted (influent and effluent). The polluted water was further subdivided into very highly (influent) and highly (effluent) polluted. The grouping of influent and effluent into one cluster was due to some water quality parameters such as amount of copper, lead, and phosphates that are not efficiently removed by the plant. Using principal component analysis, samples from the same site taken over a period of 4 months were scattered, indicating inconsistencies in the performance of the plant. This was more pronounced during the rainy season, suggesting that increased water volumes from open sewers make the already poorly performing plant worse. The major loading factors found by principle component analysis were phosphate, lead, iron, zinc, copper, pH, and conductivity. Generally, the wastewater treatment system was found to be efficient in removing heavy metals and these were found in the sludge, but not anions. The mean percentage metal removal could be arranged in the following decreasing order: iron (85%)?>?zinc (57%)?>?copper (40%) and lead (38%) following the concentrations (mg?kg^?1) found in the sludge: iron (11,300)?>?zinc (820)?>?copper (180)?>?lead (20)?>?cadmium (3). Phosphate and iron concentrations in the effluent were found to be above the South African Bureau of Standards (SABS) recommendations. The major cause of poor performance is the high volume of the wastewater, exceeding the capacity of the plant 10 times.