Controle estatístico de processo aplicado a qualidade de águas superficiais

Abstract

Water is a natural resource, biochemical component, a way of life of all species, a source for social and cultural values, a factor of production. The water pollution harms human health and the natural environment. In the characterization of water quality, we use variables that represent physical-chemical and biological indicators of water quality, which represent impurities when they exceed established values. To assess quickly and efficiently these data that indicate the environmental situation, especially the water quality, it can use technical tools of statistical process control. Thus, this study sought to evaluate the tools of statistical process control in assessing the water quality of Mandurim water to compare with CONAMA Resolution 357/2005, and identify which is the best control chart for such a process and calculate the Index of Water Quality (WQI), trophic index, univariate and multivariate process capability index. To assess the water quality of Mandurim river it was performed a literature search and a field research with data from collections made from 2003 to 2009 totaling 30 samples. The 1st point was located upstream, in the geographic coordinates Latitude: 26 º 11 '11 "South and Longitude: 53 º 06' 01" West, at an altitude of 637m. The 2nd point is downstream, near the mouth of Mandurim River, Rio Marrecas, located in the geographic coordinates Latitude: 26 º 08 '03 "South and Longitude: 53 º 06' 08" West, and altitude of 563m. The parameters chosen to evaluate water quality were: dissolved oxygen (DO), temperature of the water, total and fecal coliforms, biochemical oxygen demand (BOD), chemical oxygen demand, phosphorus, hydrogen potential, nitrogen, total solids, turbidity. The analysis of the data was through individuals Shewhart control charts, exponentiated weighted moving average (EWMA), cumulative sum (Cusum), univariate and multivariate process capability index, index of water quality, trophic index. With the exception of variable pH, the variables at all control points showed high variability in the data. In the CUSUM chart, it was observed the variables of pH, nitrogen, BOD, total solids, temperature and downstream DO, phosphorus, turbidity upstream, which processes showed to be out of statistical control by submitting more than two points above or below the limits of the chart. Among the EWMA chart, the upstream OD parameter and the nitrogen and total solids downstream parameters showed with points above or below of the control limits, and this chart was more effective when working with water quality, followed by Shewhart and CUSUM. The multivariate control chart is advantageous from the standpoint that it saves time in tracking and analyzed several parameters simultaneously. The bivariate correlations between variables of water quality showed a high correlation between the COD and BOD, beyond values greater than 0.05 in most parameters, justifying the correlation between variables. The WQI has resulted upstream 91 and downstream 86, considered as excellent quality. The TSI was considered good, since is below 44 shows a very low amount of organic matter. For the process capability index, it was observed except for pH, nitrogen, temperature, solids upstream and downstream, that the processes of the other variables are not able to meet the specifications. The results presented as a multivariate index 0.52 to 0.41 upstream and downstream, following the WQI. In the other words, when the WQI increases the capability index process also increase.