基于ACGWO-SVR的高寒地区心墙堆石坝压实质量评价模型

ACGWO-SVR-based evaluation model for compaction quality of rockfill dam with core wall in alpine region

为了提高高寒地区心墙堆石坝压实质量评价的精确度、泛化能力和鲁棒性,保证大坝施工质量,在考虑寒区气象参数的基础上,提出基于自适应灰狼优化支持向量机回归算法的压实质量评价模型。首先,在压实质量评价模型建立过程中考虑温度、湿度等气象参数的影响。其次,采用ACGWO对SVR的惩罚因子和核参数进行寻优,以提高支持向量机参数选择效率及泛化能力,进而建立基于ACGWO-SVR的压实质量评价模型;其中,ACGWO采用自适应缩放因子和混沌理论优化灰狼算法,克服了传统灰狼算法收敛速度较慢、易陷入局部最优等问题。最后,以某高寒地区心墙堆石坝工程为例,并与线性回归、BPNN以及SVR方法进行对比分析。结果表明:基于ACGWO-SVR的压实质量评价模型具有更高的精度、更好的泛化能力和鲁棒性,适用于寒区工程质量评价。

In order to improve the accuracy,generalization ability and robustness of the compaction quality evaluation of rockfill dam with core wall in alpine region and ensure the quality of the dam construction,an ACGWO-SVR(adaptive Chaos Grey Wolf Optimization-based Support Vector Regression)-based evaluation model for the compaction quality is proposed herein under the consideration of the meteorological parameters of alpine region.Firstly,the influences of the meteorological parameters of temperature,humidity,etc.are considered while establishing the compaction quality evaluation model.Secondly,the penalty factor and kernel parameters of SVR are optimized with ACGWO for enhancing the selection efficiency and the generalization ability of SVR and then establishing the ACGWO-SVR-based evaluation model;the adaptive scaling factors and chaos theory are used for ACGWO to optimize GWO,thus the problems of slow convergence speed,tendency to fall into local optima,etc.of the conventional GWO are overcome.Finally,the relevant study is carried out through taking the project of a rockfill dam with core wall in an alpine region as the study case,from which the results are compared with those obtained from the methods of linear regression,BPNN and SVR.The results show that the ACGWO-SVR-based evaluation model for the compaction quality has higher accuracy,better generalization ability and robustness and then is applicable to the evaluation of the construction quality of the project within alpine region.