Modelagem preditiva da formação de trihalometanos pela cloração de águas com diferentes concentrações de matéria orgânica

• Autor(a) principal: Almeida, Ana Elis Moraes de
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFLA
• Texto Completo: http://repositorio.ufla.br/jspui/handle/1/38802
• topic: Trihalometanos; Cloração; Desinfecção; Regressão linear múltipla; Trihalomethanes; Chlorination; Disinfection; Multiple linear regression; Tratamento de Águas de Abastecimento e Residuárias

Water and sewage treatment is linked to the prevention of waterborne diseases. The disinfection step is important in pathogen inactivation and mandatory to achieve microbiological potability standards. Chlorine is a powerful chemical oxidant, easy to apply, low cost and highly effective in disinfecting and removing color from water, however the oxidation of organic matter by it results in byproducts of carcinogenic potential, such as trihalomethanes (THM). The formation of THM is influenced by pH, dosage and contact time with chlorine, concentration of organic matter, bromide and iodide. Thus, this work aimed to evaluate the influence of organic matter concentration (COD), hypochlorite dosage, temperature, pH, turbidity and color on THM formation in chlorinated waters through the construction of regression models. Effluent dilutions of the UFLA Sewage Treatment Station were prepared by simulating COD removal at 50 (D1), 80 (D2), 90 (D3) and 95% (D4). These samples were chlorinated in triplicate with different dosages of hypochlorite: 7 (Cl1), 11 (Cl2), 15 (Cl3) and 25 mg L-1 (Cl4). After 24 hours, residual chlorine was neutralized with sodium thiosulphate and THM species concentrations were obtained. Temperature, pH, color, turbidity data were collected before and after chlorination. Concentration data of THM species were submitted to Friedman ANOVA test and data collected after chlorination were submitted to Stepwise method, using the Akaike Information Criterion (AIC), using R language, to adjust the best representative model of total trihalomethanes formation (TTHM). The amount of TTHM formed increased by 125, 126, 196 and 410% with increasing hypochlorite dosage from 7 to 25 mg L-1 for D4, D3, D2 and D1, respectively. There was an increase of 17, 40 and 35% in TTHM formation for Cl1, Cl2 and Cl3, respectively, with an increase in COD from 33.7 to 101 mg L-1 of O2, from which there was a 20% decrease. in the amount formed for Cl1 and Cl2 and 0.2% for Cl3 with increasing COD to 235.6 mg L-1 O2. For Cl4, there was a 111% increase in TTHM formation with an increase in COD from 33.7 to 235.6 mg L-1 O2. Friedman's ANOVA test, at a significance level of 5%, resulted in a non-significant difference in the amount of brominated species formed. Three iterations in the Stepwise selection were required, resulting in a linear model to predict TTHM formation, composed by the following predictive variables: Turbidity, pH, hypochlorite concentration ([OCl]) and Color, the last one negatively related to the formation of TTHM. Adjustment of the model with normalized data resulted in the order of relevance of the predictive variables: Turbidity > Color > [OCl] > pH. By the adjusted linear model, 85% of the variability of TTHM formation are contained in the variability of the variables pH, Turbidity, Color and [OCl], with high reliability (p-value = 10-16).