A novel coupling model for risk analysis of swell and shrinkage of expansive soils

Risk analysis of swell and shrinkage is essential to properly evaluate engineering stability and safety of expansive soils. Herein a novel model for risk analysis of swell and shrinkage of expansive soils was proposed by means of coupling set pair analysis with stochastic simulation of triangular fuzzy numbers. A concept ofmulti-element connection number for a set pair, which consisted of measured evaluation index sets and classification standard sets of shrink and expansion grade, was described to express hierarchy and fuzziness of membership between the evaluation sample and the classification standard. Moreover, triangular fuzzy numbers simulated by the stochastic simulationmethodwere presented to depict the changing process and fuzziness of component coefficients of discrepancy degree. Combined with the weight of the evaluation index, integrated connection number was calculated to specify the swell and shrinkage grade of the evaluation sample. A confidence level of evaluation result was also given corresponding to a confidence interval. Finally, a practical example was used to confirm and to compare with the fuzzymathematic method and extension method. The results show that this presented method is more feasible and easy to operate, and the result is good. It can provide a reliable degree of evaluation results for risk design and risk management of projects on expansive soils. © 2011 Elsevier Ltd. All rights reserved.