Controle estatístico de processo aplicado à qualidade de águas no abastecimento de Campina Grande - PB

Abstract

Water is essential for sustaining life on the planet and the dynamics of nature it permeates all the cycles of nature. The degradation of this resource affect the health, well-being of people, flora and fauna. In the characterization of water quality are determined various parameters representing the physical, chemical and biological characteristics, indicators of water quality, which are impurities when they reach values ˆaˆahigher than those established for a particular use. For an evaluation of efficient and fast data that indicate the quality of drinking water techniques of statistical process control were used. This work preferenciou the use of techniques of statistical process control in assessing the quality of drinking water in the city of Campina Grande in order to compare the data with the resolution of N0 357/2005 of the National Environmental Council (CONAMA), detect better control chart for the univariate process and optimize the process through multivariate chart. To assess the water quality in Campina Grande a literature search was performed using data provided by the water company and Sewers of Para´ıba in the period 2011-2013 a total of 36 samples. The station treatment in the municipality of Queimadas - PB, responsible for supplying Public water of Campina Grande and adjacent municipalities with the flow treated water averaged 5.400m3/h. The variables chosen for the evaluation of water were: pH, chlorine, turbidity and color as they are considered relevant features on the water distributed for consumption to be potability standards. The analysis of data was through graphs of individual control Shewhart, exponentially moving average weighted (EWMA), cumulative sum (CUSUM), and multivariate capability index univariate and multivariate process. Except for the pH variable, the other variables showed high variability in the data. In Shewhart charts observed that the color and turbidity variables exist sequences and points above the upper limit points, where their processes are shown outside the statistical control. In the graphs of CUSUM and EWMA observed pH and chlorine, the course of the process is an increase in variability, indicating an instability in the process, confirmed by the process capability index, which indicated the process variables for pH and chlorine are not able to meet your specifications, however remained within statistical control. The bivariate correlations between the variables of water quality, chlorine and turbidity showed a moderate correlation (r = 0.673). For the univariate case, the EWMA chart proved to be the best in the monitoring of water quality due its efficiency as it is sensitive to small variations of the order of 1.5 standard deviations. The multivariate control chart is beneficial from the point of view of time in monitoring and analysis of variables simultaneously. The degree of impairment of quality of drinking water and low, because only variable chlorine introduced himself outside the limits stipulated by Resolution CONAMA, which provides for the classification of water bodies.