基于ANFIS-GM的心墙堆石坝变形预测

本文提出采用自适应网络模糊推理系统(adaptive neuro-fuzzy inference system,ANFIS)优化灰色理论模型(Grey Model,GM)的建模方法来研究预测大坝变形。ANFIS-GM模型综合考虑了由于资料不完备、考虑因素不全面而产生的灰色特性和各影响因素与大坝变形之间存在的模糊特性。该模型相比于GM模型不仅考虑了大坝变形的灰色特性,而且还考虑了水位变化速率、填筑速率与大坝变形的模糊关系。通过心墙堆石坝沉降变形的实例分析,表明该模型比GM模型误差更小。同时,该模型具有处理小样本,自组织、自学习、自适应,模糊推理的综合能力。

ANFIS-GM based-prediction on deformation of rockfill dam with core wall

It is proposed herein to adopt the modeling method of the adaptive neuro-fuzzy inference system optimized grey model（ANFIS-GM）to study and predict dam deformation. In ANFIS-GM model，the fuzzy characteristics among the grey characteristics due to the incomplete data and incomprehensive consideration of the relevant factors，all the impacting factors and dam deformation are comprehensively considered. Compared with grey model（GM），this model considers not only the grey characteristics of dam deformation，but also the fuzzy relationship in-between the water level changing rate，filling rate and dam deformation，and then has the performance to process fuzzy system. Through the case analysis on the settlement deformation of rockfill dam with core wall，it is indicated that the error from this model is smaller if compared with GM model. Meanwhile，The model has the comprehensive capacity to deal with small samples with the abilities of self-organizing，self-learning，adaptive and fuzzy inference.