Research On Multi-Objective Optimization Of RCC Dam Compaction Quality,Inter-layer Bonding Quality,and Schedule Based On Intelligent Construction Quality Analysis

Compaction quality,inter-layer bonding quality and schedule are the three core objectives of roller compacted concrete(RCC)dam construction.However,there is no research on multi-objective optimization between these three objectives.And it’s difficult to quantify compaction quality and inter-layer bonding quality of RCC dam.Also,there is no schedule analysis considering the influence of quality.Focusing on the above problems,research on RCC dam compaction quality,inter-layer bonding quality,and schedule multi-objective optimization is carried out,and the main contributions are as follows:(1)Intelligent compaction quality analysis for RCC damA perception method for layered compaction stiffness is proposed,which overcomes the difficulty of perception of dynamic response characteristics of roller compactor and concrete among multi layers.And the comprehensive perception of layered compaction stiffness,rolling parameters,and concrete performance parameters is achieved.An intelligent compaction quality analysis model based on kernel extreme learning machine is established,whose parameters are optimized by a chaos theory based cuckoo search algorithm.An online update method for intelligent compaction quality analysis model is proposed based on fast leave-one-out cross-validation.Considering actual project,the prediction precision of established model is increased6.3%,4.8%,and 13.8%,respectively,compared with multiple nonlinear regression,BP neural networks,and support vector machine.And under online update,the model’s absolute residual is about 3%,which proves that model’s generalization ability is enhanced.(2)Intelligent inter-layer bonding quality analysis for RCC damA perception method for inter-layer bonding time is proposed based on spatial-temporal matching of concrete production time and unloading position.For unbalanced sample data,the intelligent inter-layer bonding quality classification model is established based on oversampling and cost-sensitive semi-supervised support vector machine,which realizes the intelligent judgment of unqualified inter-layer bonding quality.Under the framework of Ada Boost.RT integrated learning algorithm,an intelligent inter-layer bonding quality analysis model based on relevance vector regression is established.Considering actual project,the G-mean value of the established classification model is 0.908,which indicates that the model has good classification ability for unbalanced data.And the coefficient of determination(R~2)of the established intelligent analysis model is 0.8881,which indicates that the model has high prediction precision for small-sample data.(3)Adaptive construction simulation for RCC dam storehouse surface based on intelligent construction quality analysisAn adaptive simulation framework for RCC dam storehouse construction based on intelligent construction quality analysis is established.An adaptive simulation parameter update method is proposed based on the Dirichlet process mixture(DPM)models and ranking entropy improved sequential update and greedy search algorithm.An adaptive simulation logic chain adjustment method is proposed based on intelligent construction quality analysis.Considering actual project,the deviation rates between the simulated durations and the actual ones are about 3%～4%,where the high simulation precision demonstrates the effectiveness of the proposed simulation method.(4)Multi-objective optimization analysis of RCC dam’s compaction quality,inter-layer bonding quality and scheduleA mathematical model for RCC dam compaction quality,inter-layer bonding quality and schedule multi-objective optimization is established.An improved the third generation non-dominated sequencing genetic algorithm(NSGA-Ⅲ)based on adaptive reference point method is proposed to solve the Pareto optimal solution set.Relative random dominance degree and Technique for Order Preference by Similarity to Ideal Solution method(TOPSIS)are applied to make multi-attribute decisions on Pareto solution sets,which obtains optimal construction plan.Considering actual project,the optimal plan’s duration decreases by about 14%under horizontal layer construction,the mechanical utilization increases by about 9%,the compaction degree increases by about 0.5%,and tensile strength ratio increases by about 2%,which demonstrate the effectiveness of the proposed method.