Modelos matemáticos preditivos para estimativa das propriedades de solos compactados

Abstract

The purpose of this work is to determine predictive mathematical models to obtain compacted soil parameters: maximum dry unit weight and optimum moisture content. The methodology used admitted that, initially, the collect and the data registration of laboratory tests of soils applied in 17 water dams in the state of Ceará, having geotechnical properties that are important for the development of the study. Then, the statistics of the sampled data were performed, being the nonlinear regression model defined as the most adequate to represent the parameters behavior of the compaction parameters against the characterization properties of the studied soils. Determined mathematical models have shown that some indices have a greater influence on the compaction parameters, such as consistency limits, percentage of fines and void ratio. To validate the proposed theory, a sample containing 14 laboratory test results from the Umari Dam (Madalena/CE) was used, which was not part of the process of mathematical modeling.process. The results obtained presented average percentage errors of the order of 2.3% for the maximum dry unit weight and 9.4% for the optimal moisture content.. In addition, the samples were applied to the mathematical models determined by Omar et al. (2003) and Gurtug & Sridharan (2004), showing that such models present values that are not equal to the real. The results were promising, indicating that the empirical models satisfactorily represent the compaction parameters, justifying their use in the initial study phase and preliminary design, for structures that use compacted soils, so that costs and time are reduced.