| dc.description.abstract | The objective of this work was to develop a mathematical model for the prediction of settlements, applied to the sanitary landfill of São Leopoldo, managed by Companhia Riograndense de Valorização de Residuos (CRVR). Firstly, confined compression tests were performed on a sample of MSW collected at the CRVR landfill, obtaining the parameters: primary compression index (Cc = 0.17), secondary compression index (Cα = 0.14), void index (e = 0.224 to 0.241) which were applied to the classical settlement models (Sowers, 1973; and Bjarngard & Edgers, 1990). The models showed low adherence to real data, measured in the field, with errors from 67.6% to 99.9%. After, were analyzed the CRVR leachate and settlement monitoring data, these were applied to the Gomes & Caetano (2010) model. This model showed good adherence to real data, with a mean error of 27.9%. Finally, the mathematical model of settlement prediction was generated, through statistical analysis of multiple linear regression, including factorial analysis by principal component method, via orthogonal rotation of Varimax. The model used the CRVR leachate monitoring parameters and presented an average error of 20% to the actual settlement data. It was found that the use of regression prediction models, through regression analysis, is adequate to estimate the useful life of landfills, considering some adjustments in data monitoring (frequency of monitoring and continuous monitoring), and that each landfill has its particularities, and it is necessary to analyze each case to make sure that a model meets a given landfill. | |
