Identification of groundwater pollution sources based on self-adaption Co-Kriging multi-fidelity surrogate model
AN Yong-kai1,2, ZHANG Yan-xiang3, YAN Xue-man4
1. School of Water and Environment, Chang'an University, Xi'an 710054, China; 2. Key Laboratory of Subsurface Hydrology and Ecological Effects in Arid Region of the Ministry of Education, Chang'an University, Xi'an 710054, China; 3. Power China Northwest Engineering Corporation Limited, Xi'an 710065, China; 4. College of Urban and Environmental Sciences, Northwest University, Xi'an 710127, China
Abstract：To identify groundwater pollution sources efficiently and accurately, the Co-Kriging method integrating Differential evolution was used to establish a multi-fidelity surrogate model based on comprehensive application of high fidelity and low fidelity numerical simulation models for solute transport. On this basis, the Markov chain Monte Carlo (MCMC)-DREAM(D) algorithm and the adaptive updating multi fidelity surrogate model strategy were applied to identify groundwater pollution sources. To verify the effectiveness and feasibility of the above methods, this study conducted the numerical case study. The results showed that compared with the Kriging surrogate model based only on the input-output samples of the high fidelity model, the Co-Kriging surrogate model based on the joint use of input-output samples of the high fidelity and low fidelity model has higher approximation accuracy to the simulation model. The joint application of coupled multi fidelity Co-Kriging surrogate model and MCMC-DREAM(D) algorithm can not only obtain accurate identification results, but also significantly reduce the calculation load. At the same time, the adaptive updating multi fidelity surrogate model strategy can further improve the identification accuracy for groundwater pollution sources.
[1] 葛渊博,卢文喜,白玉堃,等.基于SSA-BP与SSA的地下水污染源反演识别[J]. 中国环境科学, 2022,42(11):5179-5187. Ge Y, Lu W, Bai Y, et al. Inversion and identification of groundwater pollution sources based on SSA-BP and SSA[J]. China Environmental Science, 2022,42(11):5179-5187. [2] Yan X, Dong W, An Y, et al. A Bayesian-based integrated approach for identifying groundwater contamination sources[J]. Journal of Hydrology, 2019,579:124160. [3] 闫雪嫚.基于贝叶斯理论的地下水DNAPLs污染源反演识别研究[D]. 长春:吉林大学, 2021. Yan X. Identification of DNAPLs-contaminated groundwater pollution sources based on Bayes Theory[D]. Changchun:Jilin University, 2021. [4] 罗成明,卢文喜,潘紫东,等.基于替代模型和流向算法的地下水污染源反演识别[J]. 中国环境科学, 2023,43(11):5823-5832. Luo C, Lu W, Pan Z, et al. Inversion and identification of groundwater pollution sources based on surrogate model and flow direction algorithm[J]. China Environment Science, 2023,43(11):5823-5832. [5] 王景瑞,胡立堂.地下水污染源识别的数学方法研究进展[J]. 水科学进展, 2017,28(6):943-952. Wang J, Hu L. Advances in mathematical methods of groundwater pollution source identification[J]. Advances in Water Science, 2017, 28(6):943-952. [6] Atmadja J, Bagtzoglou A C. State of the art report on mathematical methods for groundwater pollution source identification[J]. Environmental forensics, 2001,2(3):205-214. [7] 肖传宁,卢文喜,赵莹,等.基于径向基函数模型的优化方法在地下水污染源识别中的应用[J]. 中国环境科学, 2016,36(7):2067-2072. Xiao C, Lu W, Zhao Y, et al. Optimization method of identification of groundwater pollution sources based on radial basis function model[J]. China Environmental Science, 2016,36(7):2067-2072. [8] Guo J, Lu W, Yang Q, et al. The application of 0-1mixed integer nonlinear programming optimization model based on a surrogate model to identify the groundwater pollution source[J]. Journal of Contaminant Hydrology, 2019,220:18-25. [9] Ma X, Zabaras N. An efficient Bayesian inference approach to inverse problems based on an adaptive sparse grid collocation method[J]. Inverse Problems, 2009,25(3):035013. [10] Zhang J, Zeng L, Chen C, et al. Efficient Bayesian experimental design for contaminant source identification[J]. Water Resources Research, 2015,51(1):576-598. [11] 陆乐,吴吉春,陈景雅.基于贝叶斯方法的水文地质参数识别[J]. 水文地质工程地质, 2008,(5):58-63. Lu L, Wu J, Chen J. Identification of hydrogeological parameters based on the Bayesian method[J]. Hydrogeology and Engineering Geology, 2008,(5):58-63. [12] Mo S, Zabaras N, Shi X, et al. Deep autoregressive neural networks for high-dimensional inverse problems in groundwater contaminant source identification[J]. Water Resources Research, 2019,55(5):3856-3881. [13] 江思珉,孔维铭,吴延浩,等.基于稠密连接网络的地下水污染替代模型研究[J]. 同济大学学报(自然科学版), 2023,51(2):229-237. Jiang S, Kong W, Wu Y, et al. A Dense net network based surrogate model for simulating contaminant transport in groundwater systems[J]. Journal of Tongji University (Natural Science), 2023,51(2):229-237. [14] 安永凯,卢文喜,董海彪,等.基于克里格法的地下水流数值模拟模型的替代模型研究[J]. 中国环境科学, 2014,34(4):1073-1079. An Y, Lu W, Dong H, et al. Surrogate model of numerieal simulation model of groundwater based on Kriging[J]. China Environmental Science, 2014,34(4):1073-1079. [15] Razavi S, Tolson B A, Burn D H. Review of surrogate modeling in water resources[J]. Water Resources Research, 2012,48(7):W07401. [16] 张将伟,卢文喜,曲延光,等.基于Monte Carlo方法的地表水地下水耦合模拟模型不确定分析[J]. 水利学报, 2018,49(10):1254-1264. Zhang J, Lu W, Qu Y, et al. Uncertainty analysis of surface water and groundwater coupling simulation model based on Monte Carlo method[J]. SHUILI XUEBAO, 2018,49(10):1254-1264. [17] Thokala P, Martins J R R A. Variable-complexity optimization applied to airfoil design[J]. Engineering optimization, 2007,39(3):271-286. [18] Kennedy M C, O'Hagan A. Predicting the output from a complex computer code when fast approximations are available[J]. Biometrika, 2000,87(1):1-13. [19] Clare M C A, Leijnse T W B, McCall R T, et al. Multilevel multifidelity Monte Carlo methods for assessing uncertainty in coastal flooding[J]. Natural Hazards and Earth System Sciences, 2022,22(8):2491-2515. [20] 刘佳,杨克巍,姜江,等.基于多保真度代理模型的装备系统参数估计方法[J]. 系统工程与电子技术, 2020,43(1):130-137. Liu J, Yang K, Jiang J, et al. Parameter estimation method equipment system based on multi-fidelity surrogate model[J]. System Engineering and Electronics, 2020,43(1):130-137. [21] Sobester A, Forrester A, Keane A. Engineering design via surrogate modelling:a practical guide[M]. John Wiley & Sons, 2008. [22] Zheng Q, Zhang J, Xu W, et al. Adaptive multifidelity data assimilation for nonlinear subsurface flow problems[J]. Water Resources Research, 2019,55(1):203-217. [23] Forrester A I J, Sóbester A, Keane A J. Multi-fidelity optimization via surrogate modeling[J]. Proceedings of the royal society a:mathematical, physical and engineering sciences, 2007,463(2088):3251-3269. [24] Price K V. Differential evolution[M]. Berlin, 2013:187-214. [25] 朱嵩.基于贝叶斯推理的环境水力学反问题研究[D]. 杭州:浙江大学, 2008. Zhu S. Research on the inverse problems of environmental hydraulics by Bayesian inference[D]. Hangzhou:Zhejiang University, 2008. [26] 骆乾坤,吴剑锋,杨运,等.基于DREAM算法的含水层渗透系数空间变异特征识别[J]. 南京大学学报(自然科学), 2016,52(3):448- 455. Luo Q, Wu J, Yang Y, et al. Identification of spatial variability of aquifer hydraulic conductivity[J]. Journal of Nanjing University (Natural Science), 2016,52(3):448-455. [27] Vrugt J A, Ter Braak C J F, Clark M P, et al. Treatment of input uncertainty in hydrologic modeling:Doing hydrology backward with Markov chain Monte Carlo simulation[J]. Water Resources Research, 2008,44(12):WR006720. [28] 张双圣,强静,刘汉湖,等.基于贝叶斯公式的地下水污染源识别[J]. 中国环境科学, 2019,39(4):1568-1578. Zhang S, Qiang J, Liu H, et al. Identification of groundwater pollution sources based on Bayes' theorem[J]. China Environmental Science, 2019,39(4):1568-1578. [29] 顾文龙,卢文喜,张宇,等.基于贝叶斯推理与改进的MCMC方法反演地下水污染源释放历史[J]. 水利学报, 2016,47(6):772-779. Gu W, Lu W, Zhang Y, et al. Reconstructing the release history of groundwater contamination sources based on the Bayesian inference and improved MCMC method[J]. Shuili Xuebao, 2016,47(6):772- 779. [30] Zhang J, Vrugt J A, Shi X, et al. Improving simulation efficiency of MCMC for inverse modeling of hydrologic systems with a kalman-inspired proposal distribution[J]. Water Resources Research, 2020,56(3):WR025474. [31] Vrugt J A, Ter Braak C J F. DREAM(D):an adaptive Markov Chain Monte Carlo simulation algorithm to solve discrete, noncontinuous, and combinatorial posterior parameter estimation problems[J]. Hydrology and Earth System Sciences, 2011,8(12):3701-3713. [32] Gelman A G, Rubin D B. Inference from iterative simulation using multiple sequences[J]. Statistical Science, 1992,7(4):457-472.
