2024年6月20日 星期四
基于多目标机器学习模型的城市内涝快速模拟研究
Fast Simulation of Urban Waterlogging Based on Multi-Objective Machine Learning Model
摘要

近年来频发的城市暴雨内涝灾害已给国家造成了严重的经济损失和人员伤亡。内涝数值模拟是灾害预警预报及防控的一种重要手段, 然而传统的数值物理模型存在计算效率低下的缺点, 难以满足暴雨内涝实时模拟和预警预报需求。为此, 本研究结合具有物理机制的耦合雨洪模型和机器学习算法各自的优势, 提出基于多目标机器学习算法的城市暴雨内涝淹没水深快速模拟的方法, 分别探讨了K近邻 (KNN) 、多目标随机森林 (MORF) 、极端梯度提升 (XGBoost) 及其集成模型的预测性能, 结果表明: (1) 基于SWMM和LISFLOOD-FP构建的耦合雨洪模型在研究区暴雨内涝模拟中具有良好的适用性, 在此基础上共生成了70种包含不同特征的“暴雨内涝”数据库; (2) KNN、MORF、XGBoost及其集成模型的水深预测效果均较好, 皮尔逊相关系数 (PCC) 值均达0. 812以上, 平均绝对误差 (MAE) 均在6. 9cm以下, 均方根误差 (RMSE) 不超过0. 116; KNNMORF-XGBoost集成模型的总体效果最好, 其MAE、PCC和RMSE的平均值分别为2. 4cm、0. 965和0. 043; (3) 所构建的多目标机器学习预测模型除了预测精度高外, 其预测速度极快, 水深模拟效率比耦合雨洪模型提升20倍以上。本研究可为机器学习在城市暴雨内涝快速模拟方面提供一种新思路, 对内涝灾害的预警预报具有重要价值。

Abstract

The frequent occurrence of urban waterlogging disasters induced by rainstorm has recently caused serious economic losses and casualties in China. Numerical simulation of waterlogging is an important tool for disaster prewarning and forecasting as well as disaster prevention and control; however, the traditional numerical physical models have the disadvantage of low computational efficiency, which makes it difficult to meet the demand for real-time simulation and real-time early warning and forecast. To this end, this study combines the respective advantages of coupled rainstorm-flood models with physical mechanisms and machine learning algorithms, and proposes a rapid prediction and simulation method for inundated depth of urban waterlogging based on multi-objective machine learning algorithms. The forecasting performances of K-Nearest Neighbors (KNN) , Multi-Objective Random Forest (MORF) , Extreme Gradient Boosting (XGBoost) and their integrated models are discussed, respectively. The results show that: (1) The coupled rainstorm- flood model based on SWMM and LISFLOOD-FP has good applicability in the simulation of urban waterlogging induced by rainstorm in the study area. On this basis, the database with a total of 70 scenarios of rainstorm-inundation with different characteristics were generated. (2) The KNN, MORF, XGBoost and their integrated models all have good results in predicting water depth, with Pearson correlation coefficient (PCC) values all above 0. 812, mean absolute error (MAE) below 6. 9 cm, and root-mean-square error (RMSE) less than 0. 116. The KNN-MORF-XGBoost integrated model has the best overall results, with the average values of MAE, PCC and RMSE reaching 2. 4cm, 0. 965 and 0. 043, respectively. (3) In addition to the high prediction accuracy, the prediction speed of the constructed multi-objective machine learning prediction model is extremely fast, and the water depth simulation efficiency is more than 20 times higher than that of the coupled rainstorm-flood model. This study can provide a reference for the application of machine learning in the rapid simulation of urban waterlogging induced by rainstorm, which is of great value for the early warning and forecast of urban waterlogging disaster.  

DOI10.48014/fcws.20220827001
文章类型研究性论文
收稿日期2022-08-29
接收日期2022-11-06
出版日期2023-03-28
关键词城市内涝, 耦合雨洪模型, 机器学习, 多目标预测, 快速模拟
KeywordsUrban waterlogging, coupled rainstorm-flood model, machine learning, multi- objective prediction, fast simulation
作者赖成光1,2, 廖耀星1, 王兆礼1,2,*, 陈晓宏3
AuthorLAI Chengguang1,2, LIAO Yaoxing1, WANG Zhaoli1,2,*, CHEN Xiaohong3
所在单位1. 华南理工大学土木与交通学院, 广州 510641
2. 人工智能与数字经济广东省实验室 (广州) , 广州 510330
3. 中山大学水资源与环境研究中心, 广州 510275
Company1. School of Civil Engineering and Transportation, South China University of Technology, Guangzhou 510641, China
2. Artificial Intelligence and Digital Economy Laboratory (Guangzhou) , Guangzhou 510330, China
3. Center for Water Resources and Environment Research, Sun Yat-sen University, Guangzhou 510275, China
浏览量1309
下载量433
基金项目国家重点研发计划项目(2021YFC3001000);
国家自然科学基金项目(U1911204; 51879107)
参考文献[1] IPCC. Climate Change 2021: The Physical Science Basis [M]. Cambridge: Cambridge University Press, 2021.
DOI:10.1017/9781009157896
[2] Hallegatte S, Green C, Nicholls R J, et al. Future flood losses in major coastal cities[J]. Nat Clim Change, 2013, 3(9): 802-806.
https://doi.org/10.1038/nclimate1979
[3] 徐宗学, 叶陈雷. 城市暴雨洪涝模拟: 原理、模型与展望[J]. 水利学报, 2021, 52(04): 381-392.
DOI:10.13243/j.cnki.slxb.20200515
[4] 王兆礼, 陈昱宏, 赖成光. 基于TELEMAC-2D 和 SWMM模型的城市内涝数值模拟[J]. 水资源保护, 2022, 38(01): 117-124.
DOI:10.3880/j.issn.1004-6933.2022.01.015
[5] Wu X, Wang Z, Guo S, et al. Scenario-based projections of future urban inundation within a coupled hydrodynamic model framework: A case study in Dongguan City, China [J]. J Hydrol, 2017, 547: 428-442.
https://doi.org/10.1016/j.jhydrol.2017.02.020
[6] 曾照洋, 赖成光, 王兆礼, 等. 基于WCA2D与SWMM模型的城市暴雨洪涝快速模拟[J]. 水科学进展, 2020, 31(01): 29-38.
DOI:10.14042/j.cnki.32.1309.2020.01.004
[7] 曾照洋, 王兆礼, 吴旭树, 等. 基于SWMM和LISFLOOD模型的暴雨内涝模拟研究[J]. 水力发电学报, 2017, 36(05): 68-77.
DOI:10.11660/slfdxb.20170508
[8] 刘媛媛, 刘业森, 郑敬伟, 等. BP神经网络和数值模型相结合的城市内涝预测方法研究[J]. 水利学报, 2022, 53(03): 284-295.
DOI:10.13243/j.cnki.slxb.20210611
[9] Teng J, Vaze J, Kim S, et al. Enhancing the capability of a simple, computationally efficient, conceptual flood inundation model in hydrologically complex terrain[J]. Water Resour Manage, 2019, 33(2): 831-845.
https://doi.org/10.1007/s11269-018-2146-7
[10] Chu H, Wu W, Wang Q J, et al. An ANN-based emulation modelling framework for flood inundation modelling: Application, challenges and future directions[J]. Environmental modelling & software: with environment data news, 2020, 124: 104587.
https://doi.org/10.1016/j.envsoft.2019.104587
[11] Xu T, Liang F. Machine learning for hydrologic sciences: An introductory overview[J]. WIREs Water, 2021, 8(5): e1533.
https://doi.org/10.1002/wat2.1533
[12] Hu C, Wu Q, Li H, et al. Deep learning with a long short-term memory networks approach for rainfall-runoff simulation[J]. Water, 2018, 10(11): 1543.
https://doi.org/10.3390/w10111543
[13] Panda R K, Pramanik N, Bala B. Simulation of river stage using artificial neural network and MIKE 11 hydrodynamic model[J]. Comput Geosci, 2010, 36(6): 735-745.
https://doi.org/10.1016/j.cageo.2009.07.012
[14] Lin G-F, Lin H-Y, Chou Y-C. Development of a realtime regional-inundation forecasting model for the inundation warning system[J]. J Hydroinf, 2013, 15(4): 1391-1407.
DOI:10.2166/hydro.2013.202
[15] Jhong B-C, Wang J-H, Lin G-F. An integrated twostage support vector machine approach to forecast inundation maps during typhoons[J]. J Hydrol, 2017, 547: 236-252.
https://doi.org/10.1016/j.jhydrol.2017.01.057
[16] Lin Q, Leandro J, Gerber S, et al. Multistep flood inundation forecasts with resilient backpropagation neural networks: Kulmbach case study[J]. Water, 2020, 12(12): 3568.
https://doi.org/10.3390/w12123568
[17] Kabir S, Patidar S, Xia X, et al. A deep convolutional neural network model for rapid prediction of fluvial flood inundation[J]. J Hydrol, 2020, 590: 125481.
https://doi.org/10.1016/j.jhydrol.2020.125481
[18] 刘晓明, 王新, 徐慧. 基于多目标随机森林的煤层厚度 同步预测方法[J]. 计算机工程与设计, 2021, 42(04): 1116-1121.
DOI:10.16208/j.issn1000-7024.2021.04.031
[19] Liang X, Yuetian L, Yifei X, et al. A data-driven shale gas production forecasting method based on the multiobjective random forest regression[J]. Journal of Petroleum Science and Engineering, 2021, 196: 107801.
https://doi.org/10.1016/j.petrol.2020.107801
[20] Liu W, Xu D, Tsang I W, et al. Metriclearning for multi-output tasks[J]. IEEE transactions on pattern analysis and machine intelligence, 2019, 41(2): 408-422.
DOI:10.1109/TPAMI.2018.2794976
[21] Rossman L A. Storm Water Management Model User's Manual Version 5. 1[M]. Washington DC: USEPA, 2015.
[22] Bates P D, De Roo A P J. A simple raster-based model for flood inundation simulation[J]. Journal of Hydrology, 2000, 236(1): 54-77.
https://doi.org/10.1016/S0022-1694(00)00278-X
[23] Zhang M, Xu M, Wang Z, et al. Assessment of the vulnerability of road networks to urban waterlogging based on a coupled hydrodynamic model[J]. J Hydrol, 2021, 603: 127105.
https://doi.org/10.1016/j.jhydrol.2021.127105
[24] Breiman L. Random forests[J]. Machine Learning, 2001, 45(1): 5-32.
https://doi.org/10.1023/A:1010933404324
[25] Liao Y, Wang Z, Xiong J, et al. Dimming in the pearl river delta of China and the major influencing factors [J]. Clim Res, 2021, 82: 161-176.
https://doi.org/10.3354/cr01626
[26] Kocev D, Vens C, Struyf J, et al. Ensembles of Multi- Objective Decision Trees[M]. Berlin, Heidelberg: Springer Berlin Heidelberg, 2007.
https://doi.org/10.1007/978-3-540-74958-5_61
[27] Saha D, Alluri P, Gan A. A random forests approach to prioritize Highway Safety Manual(HSM)variables for data collection: Random Forests to Prioritize HSM Variables[J]. J Adv Transp, 2016, 50(4): 522-540.
https://doi.org/10.1002/atr.1358
[28] Borchani H, Varando G, Bielza C, et al. A survey on multi-output regression[J]. WIREs Data Min Knowl Discovery, 2015, 5(5): 216-233.
https://doi.org/10.1002/widm.1157
[29] Chen T, Guestrin C. XGBoost: A Scalable Tree Boosting System[C]. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco, California, USA; Association for Computing Machinery. 2016: 785-794.
https://doi.org/10.1145/2939672.2939785
[30] 周聂, 侯精明, 陈光照, 等. 基于机器学习的山洪灾害快速预报方法[J]. 水资源保护, 2022, 38(02): 32-40, 111.
DOI:10.3880/j.issn.1004-6933.2022.02.005
[31] Zeng Z, Wang Z, Lai C. Simulation performance evaluation and uncertainty analysis on a coupled inundation model combining SWMM and WCA2D[J]. Int J Disaster Risk Sci, 2022: 13(3): 448-464.
https://doi.org/10.1007/s13753-022-00416-3
[32] 赖成光, 廖耀星, 王兆礼. 不同暴雨雨型条件下的城市内涝响应特征分析[J/OL]. 水资源保护: 1-14[2022- 08-09].
https://kns-cnki-net.webvpn.scut.edu.cn/kcms/detail/32.1356.TV.20220801.1424.002.html
引用本文赖成光, 廖耀星, 王兆礼, 等. 基于多目标机器学习模型的城市内涝快速模拟研究[J]. 中国水科学前沿, 2023, 1(1): 1-16.
CitationLAI Chengguang, LIAO Yaoxing, WANG Zhaoli, et al. Fast simulation of urban waterlogging based on multi-objective machine learning models[J]. Frontiers of Chinese Water Sciences, 2023, 1(1): 1-16.