约束型ME-PP模型在大坝安全评价中的应用

鉴于大坝安全综合评价中的权重选取关系到评价结果的可靠性、合理性,为充分挖掘大坝监测数据所携带的信息,并考虑影响大坝基岩安全的因素特征,从投影寻踪方法的优化函数和约束条件两个方面改进传统算法,即采用最大熵原理与投影寻踪耦合模型改进投影寻踪方法的优化函数,结合实测指标的特征改进投影寻踪法的约束条件,从而得到约束型最大熵投影寻踪耦合模型。将该模型计算得到的各实测指标权重运用于灰色聚类综合评价,并利用聚类系数向量的灰度进行方法优劣的判别。实例应用结果表明,将本文方法求得的权重用于综合评价是合理可行的,相比于熵权法、传统投影寻踪法,更符合大坝基岩实测数据的特征,且得到的评价结果灰度更小,更易区分所属灰类。

Application of Constrained ME-PP Model in Dam Safety Assessment

Selecting weight in comprehensive evaluation of dam safety is related to the reliability and the rationality of the evaluation result. In order to fully tap the information of dam monitoring data, considering the characteristics of the dam bedrock safety factors, tradition projection pursuit (PP) model is improved from two aspects of optimization function and constraints. The optimization function of PP model is improved by using maximum entropy principle coupled with projection pursuit model (ME-PP) and the constraints are improved by combining with the features of measured parameters. Thus, constrained ME-PP model is established. The weight of each index obtained by the model is applied to grey clustering comprehensive evaluation. And the gray of clustering coefficient vector is used to judge the performance of each method. Example results show that the proposed weight model is reasonable and feasible for comprehensive evaluation. Comparing with the entropy method and the traditional projection pursuit, the model is more in line with the characteristics of measured data of dam bedrock, and the gray degree of evaluation result is smaller, which is easier to distinguish the classification of gray.