Research Of Dam Seepage Safety Comprehensive Evaluation Based On Improved Finite Interval Cloud Model

The safety of dam seepage is directly related to the safety and stability of the dam,and its comprehensive safety evaluation is an important research content in the field of dam safety.Affected by factors such as monitoring technology and human operation,The comprehensive evaluation index of dam seepage safety obtained based on site monitoring is random.At the same time,due to the limitations of expert experience,the boundaries of the evaluation index levels are inevitably fuzzy.Therefore,the comprehensive evaluation of dam seepage safety is a complex problem with randomness and fuzziness.The cloud model has the advantages of considering both randomness and fuzziness in the evaluation,so it is used in the comprehensive evaluation of dam seepage safety.However,there are some problems in the existing comprehensive evaluation research of dam seepage safety.The grade division of the evaluation index is too subjective;the combination weighting method of the evaluation index presupposes that the reliability of each single weighting method is the same,which is difficult to fully express the subjective and objective weight information.At the same time,most traditional cloud models assume that the evaluation indexes completely obey the normal distribution,ignoring the actual distribution characteristics of the evaluation indexes,and use the mean value of a large number of cloud drops randomly generated based on the cloud generator algorithm as the grade membership degree of the evaluation index,which fails to fully consider the randomness of cloud drop generation.To solve these problems,based on backward cloud generator algorithm,improved game theory combination weighting method,improved finite interval cloud model and the principle of maximum entropy,this paper carried out the research of dam seepage safety comprehensive evaluation based on improved finite interval cloud model.The main research results are as follows:（1）The method of index grade division of the comprehensive evaluation for dam seepage safety based on backward cloud is proposed.The existing grade division methods of comprehensive evaluation indexes for dam seepage safety are mostly based on the subjective determination of expert experience,which is too subjective,ignoring the distribution law of the monitoring data itself.Based on the experience of experts and combined with large amount of monitoring data,this paper uses the backward cloud generator algorithm to transform the quantitative dam seepage safety monitoring data into the qualitative performance of seepage safety expressed by the digital characteristics of the cloud（Ex,En,He）,and then determines the grade division based on the"3En"principle to improve the rationality of the grade division of evaluation indexes.（2）The subjective weighting method of group decision making based on hesitant cloud linguistic term sets,the objective weighting method of CRITIC method improved with the entropy method,and the combination weighting method based on improved game theory is proposed.In terms of subjective weight calculation,the group decision making based on hesitant cloud linguistic term sets（HCLTSs）is adopted to obtain subjective weight,which introduces HCLTSs that considers the hesitation of experts and invites several experts to participate in decision-making using the wisdom of the group to improve the rationality of subjective weight.In terms of objective weight calculation,the CRITIC method improved with the entropy method is proposed.The coefficient of variation is used to measure the difference within the index.At the same time,the CRITIC method is improved by the entropy method that can effectively measure the discreteness of data,which realizes the unification of the discreteness,relevance and difference.As for combination weight calculation,the combination weighting method based on improved game theory is proposed.The deviation minimization objective function is improved to ensure that the obtained weight combination coefficients are all positive.The consistency test is used to judge and adjust the possible conflict weights effectively,and the weight fusion can be carried out after passing the consistency test,so that the information contained in the subjective and objective weights can be fully expressed.（3）The comprehensive evaluation method for dam seepage safety based on improved finite interval cloud model is proposed.In this paper,the finite interval cloud model that can reflect the distribution characteristics of the evaluation indexes is used for the comprehensive evaluation of dam seepage safety.The limited interval cloud model is a modification and improvement of the traditional cloud model.When the evaluation index is outside the mean value of the two ends of the grade cloud,the cloud drop no longer belongs to the normal distribution,but uniform distribution with the membership degree of 1,which can reflect the characteristic of finite interval distribution of evaluation indexes.At the same time,the probability distribution of cloud drops is calculated based on the principle of maximum entropy.Then the comprehensive membership degree of the evaluation index considering the randomness of cloud drop generation is obtained,which improves the reliability of the evaluation result.（4）Taking the actual project as an example,the comprehensive evaluation of dam seepage safety based on the improved finite interval cloud model is carried out.The comprehensive evaluation method proposed in this paper is used to carry out dam seepage safety comprehensive evaluation of an earth-rock dam in southwest China.The dam seepage safety comprehensive evaluation level is divided into five levels{safe（L₁）,basically safe（L₂）,mild dangerous（L₃）,dangerous（L₄）,relatively dangerous（L₅）},and the grade division of each evaluation index is determined by the backward cloud.The evaluation results show that the seepage safety behavior of the dam is between safe（L₁）and basically safe（L₂）,which is consistent with the actual situation.Comparing the evaluation index weights got from the combined weighting method based on the improved game theory with the results of the single weighting method and the traditional game theory combination weighting method,the rationality of the combination weighting method proposed in this paper is verified.Further,the index grade division method based on the backward cloud is compared with the method based on expert experience to verify the rationality of the evaluation index grade division method proposed in this paper.The comprehensive evaluation method based on improved finite interval cloud model in this paper is compared with the traditional cloud model,finite interval cloud model,traditional cloud model improved by the principle of maximum entropy,fuzzy comprehensive evaluation and set pair analysis to verify its consistency,validity and reliability.