Entropy and normal cloud model-based grade evaluation of soft rock

Soft rock is one of the main engineering geological hazards occurring in the underground engineering construction. The classification of soft rock is an important premise for safe construction. Six quantitative indices including uniaxial compressive strength σ_c, integrity coefficient K_v, cohesion σ_n, softening coefficient K_f and softening index f_s are chosen as the predictor variables of soft rock according to the unascertained factors of classification prediction of soft rock. The entropy method was adopted to determine the weighting coefficient for each evaluation index. Based on entropy method and the unascertained measurement theory, with 4 groups of typical engineering examples, an entropy and the normal cloud model to predict the possibility and classification of soft rock was established. Then the proposed model was validated with four typical soft rock projects and compared with the results of unascertained degree method, fuzzy evaluation method and BQ method. The obtained results show a practicability and reliability in the grading of soft rock. It may provide a new way to predict the type of soft rock.