| PUBLIC SAFETY |
|
|
|
|
|
Spatiotemporal rapid prediction model of urban rainstorm waterlogging based on machine learning |
| DAI Xin, HUANG Hong, JI Xinyu, WANG Wei |
| Institute of Public Safety Research, Department of Engineering Physics, Tsinghua University, Beijing 100084, China |
|
|
|
|
Abstract [Objective] Rapid prediction of rainstorm waterlogging is crucial for disaster prevention and reduction. However, the traditional numerical models for simulating and predicting large-scale and complex subsurface conditions are complicated and time-consuming; moreover, the time-efficiency requirement of rainstorm waterlogging prediction is difficult to meet. To address these shortages of the numerical models, this study constructs a spatiotemporal prediction model of urban rainstorm waterlogging based on machine learning methods to rapidly predict waterlogging extent and water depth changes.[Methods] This study constructs a rapid prediction model of urban rainstorm waterlogging based on a hydrodynamics model and machine learning algorithms. First, a hydrodynamic model is constructed based on InfoWorks integrated catchment management (InfoWorks ICM) for rainstorm waterlogging in the study area with the parameter rate determination and model validation to realize the high-precision simulation of urban rainstorm waterlogging. On this basis, a rainfall scenario-driven hydraulics model is designed to further obtain rainstorm waterlogging simulation results. These results are used as the base dataset for machine learning. Second, the spatial characteristics data of rainstorm waterlogging are obtained from three aspects: rainfall situation, subsurface information, and the drainage capacity of the pipe network, which, together with the grid simulation results, comprise the dataset. The spatial prediction models are based on random forest, extreme gradient boosting (XGBoost), and K-nearest neighbor algorithms. Finally, the simulation results of waterlogging points are used to generate rainstorm waterlogging time series data. The rainfall, cumulative rainfall, and water depth of the first four moments (every 5 min) are used as the input for a long short-term memory (LSTM) neural network to predict the present water depth of the flooding point. The two models collaborate to achieve rapid spatial and temporal predictions of urban rainstorm waterlogging.[Results] For spatial predictions, the random forest model has the best fitting performance regarding evaluation indexes such as the mean square error, the mean absolute error, and the coefficient of determination (R2). When a rainstorm scenario with an 80-year event and a 2.5 h rainfall calendar prediction set is used, the prediction results concur with the risk map of urban waterlogging in Beijing. Compared with the simulation results of InfoWorks ICM, the prediction accuracy of the predicted inundation extent reaches 99.51%, and the average prediction error of waterlogging depth does not exceed 5.00% by the random forest model. For temporal predictions, the trend of the water depth change of the LSTM neural network model is more consistent with the simulation results of InfoWorks ICM, the R2 of four typical inundation points are above 0.900, the average absolute error of water depth prediction at the peak moment is 1.9cm, and the average relative error is 4.0%.[Conclusions] When addressing sudden rainstorms, the rapid prediction model based on machine learning algorithms built in this study can generate accurate prediction results of flooding extent and water depth in seconds by simply updating the forecast rainfall data in the model input. The model computational speed is greatly improved compared to the hydrodynamics-based numerical model, which can help plan waterlogging mitigation and relief measures.
|
| Keywords
rainstorm waterlogging
rapid spatiotemporal prediction
machine learning
random forest
long short-term memory (LSTM) neural network
|
|
Issue Date: 12 May 2023
|
|
|
[1] 舒章康,李文鑫,张建云,等.中国极端降水和高温历史变化及未来趋势[J].中国工程科学, 2022, 24(5):116-125. SHU Z K, LI, ZHANG J Y, et al. Historical changes and future trends of extreme precipitation and high temperature in China[J]. Strategic Study of CAE, 2022, 24(5):116-125.(in Chinese)
[2] 孔锋. 2012年北京"7·21 "特大暴雨洪涝灾害应对及启示[J].中国减灾, 2022(9):42-45. KONG F. Response to the" 7·21"heavy rain and flood disaster in Beijing in 2012 and its implications[J]. Disaster Reduction in China, 2022(9):42-45.(in Chinese)
[3] 黄晶,李梦晗,康晋乐,等.基于社交媒体的暴雨灾情信息实时挖掘与分析:以2019年"4·11深圳暴雨"为例[J].水利经济, 2021, 39(2):86-94, 98. HUANG J, LI M H, KANG J L, et al. Mining and analysis of rainstorm disaster information based on social media:Case study of Shenzhen rainstorm on April 11, 2019[J]. Journal of Economics of Water Resources, 2021, 39(2):86-94, 98.(in Chinese)
[4] 国务院灾害调查组.河南郑州"7·20"特大暴雨灾害调查报告[R].北京:国务院灾害调查组, 2022. Disaster Investigation Team of the State Council. Investigation report of "7·20" heavy rain disaster in Zhengzhou, Henan Province[R]. Beijing:Disaster Investigation Team of the State Council, 2022.(in Chinese)
[5] 夏军,王惠筠,甘瑶瑶,等.中国暴雨洪涝预报方法的研究进展[J].暴雨灾害, 2019, 38(5):416-421. XIA J, WANG H J, GAN Y Y, et al. Research progress in forecasting methods of rainstorm and flood disaster in China[J]. Torrential Rain and Disasters, 2019, 38(5):416-421.(in Chinese)
[6] 黄国如,王欣,黄维.基于InfoWorks ICM模型的城市暴雨内涝模拟[J].水电能源科学, 2017, 35(2):66-70, 60. HUANG G R, WANG X, HUANG W. Simulation of rainstorm water logging in urban area based on InfoWorks ICM model[J]. Water Resources and Power, 2017, 35(2):66-70, 60.(in Chinese)
[7] 金潮森,邸苏闯,于磊,等.北京中心城区内涝风险区快速识别技术研究[J].北京规划建设, 2022(4):9-13. JIN C S, DI S C, YU L, et al. Research on rapid identification technology of waterlogging risk area in central urban area of Beijing[J]. Beijing Planning Review, 2022(4):9-13.(in Chinese)
[8] 曾鹏.机器学习技术在城市洪涝快速预报中的应用研究[D].北京:中国水利水电科学研究院, 2020. ZENG P. Application research of machine learning technology in rapid forecast of urban flood[D]. Beijing:China Institute of Water Resources and Hydropower Research, 2020.(in Chinese)
[9] YAN J, JIN J M, CHEN F R, et al. Urban flash flood forecast using support vector machine and numerical simulation[J]. Journal of Hydroinformatics, 2018, 20(1):221-231.
[10] 周易宏.基于深度学习的城市洪涝积水点淹没过程预报预警研究[D].郑州:郑州大学, 2021. ZHOU Y H. Study on prediction and early warning of urban flood waterlogging process based on deep learning[D]. Zhengzhou:Zhengzhou University, 2021.(in Chinese)
[11] WANG H F, ZHAO Y J, ZHOU Y H, et al. Prediction of urban water accumulation points and water accumulation process based on machine learning[J]. Earth Science Informatics, 2021, 14(4):2317-2328.
[12] KHATRI S, KOKANE P, KUMAR V, et al. Prediction of waterlogged zones under heavy rainfall conditions using machine learning and GIS tools:A case study of Mumbai[J]. GeoJournal, 2022, 87(4):1-15.
[13] 李海宏,吴吉东,王强,等.基于机器学习方法的上海市暴雨内涝灾情预测模型研究[J].自然灾害学报, 2021, 30(1):191-200. LI H H, WU J D, WANG Q, et al. A study on rainstorm waterlogging disaster prediction models in Shanghai based on machine learning[J]. Journal of Natural Disasters, 2021, 30(1):191-200.(in Chinese)
[14] TANG X Z, SHU Y Q, LIAN Y Q, et al. A spatial assessment of urban waterlogging risk based on a weighted naïve Bayes classifier[J]. Science of the Total Environment, 2018, 630:264-274.
[15] BERKHAHN S, FUCHS L, NEUWEILER I. An ensemble neural network model for real-time prediction of urban floods[J]. Journal of Hydrology, 2019, 575:743-754.
[16] LE X H, HO H V, LEE G, et al. Application of long short-term memory (LSTM) neural network for flood forecasting[J]. Water, 2019, 11(7):1387.
[17] 侯天宇,梁好,霍凯,等.基于智能物联网技术的天津城市积水监测预警系统[J].气象研究与应用, 2021, 42(1):85-89. HOU T Y, LIANG H, HUO K, et al. Tianjin urban water logging monitoring and early warning system based on intelligent Internet of things technology[J]. Journal of Meteorological Research and Application, 2021, 42(1):85-89.(in Chinese)
[18] 吕红燕,冯倩.随机森林算法研究综述[J].河北省科学院学报, 2019, 36(3):37-41. Lü H Y, FENG Q. A review of random forests algorithm[J]. Journal of the Hebei Academy of Sciences, 2019, 36(3):37-41.(in Chinese)
[19] 彭岩,马铃,张文静,等.基于集成学习的风险预测模型研究与应用[J].计算机工程与设计, 2022, 43(4):956-961. PENG Y, MA L, ZHANG W J, et al. Research and application of risk forecast model based on ensemble learning[J]. Computer Engineering and Design, 2022, 43(4):956-961.(in Chinese)
[20] 潘鑫鑫,侯精明,陈光照,等.基于K近邻和水动力模型的城市内涝快速预报[J/OL].(2022-08-26)[2022-11-24].水资源保护, 2022. http://kns.cnki.net/kcms/detail/32.1356.TV.20220825.1710.002.html. PAN, HOU J M, CHEN G Z, et al. Rapid prediction of urban inundation based on the K-nearest neighbor method and the hydrodynamic model[J/OL].(2022-08-26)[2022-11-24]. Water Resources Conservation, 2022. http://kns.cnki.net/kcms/detail/32.1356.TV.20220825.1710.002.html.(in Chinese)
[21] 杨丽,吴雨茜,王俊丽,等.循环神经网络研究综述[J].计算机应用, 2018, 38(S2):1-6, 26. YANG L, WU Y Q, WANG J L, et al. Research on recurrent neural network[J]. Journal of Computer Applications, 2018, 38(S2):1-6, 26.(in Chinese)
[22] 陈鹏,张继权,孙滢悦,等.暴雨内涝灾害模拟研究[J].科技导报, 2017, 35(21):89-94. CHEN P, ZHANG J Q, SUN Y Y, et al. Simulation study of rainstorm water logging disaster[J]. Science&Technology Review, 2017, 35(21):89-94.(in Chinese)
[23] 北京市规划和国土资源管理委员会,北京市质量技术监督局.城镇雨水系统规划设计暴雨径流计算标准:DB11/T 969-2016[S].北京:中国标准出版社, 2016. Beijing Municipal Commission of Planning and Land Resources Administration, Beijing Municipal Bureau of Quality and Technical Supervision. Standard of rainstorm runoff calculation for urban storm drainage system planning and design:DB11/T 969-2016[S]. Beijing:Standards Press of China, 2016.(in Chinese)
[24] 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(3):284-295. LIU Y Y, LIU Y S, ZHENG J W, et al. Intelligent rapid prediction method of urban flooding based on BP neural network and numerical simulation model[J]. Journal of Hydraulic Engineering, 2022, 53(3):284-295.(in Chinese)
[25] 北京市水科学技术研究院.北京发布首份城市积水内涝风险地图[J].北京水务, 2022(5):I0010. Beijing Institute of Water Science and Technology. Beijing releases the first urban waterlogging risk map[J]. Beijing Water, 2022(5):I0010.(in Chinese)
[26] 吕佳.梯度提升回归树算法研究及改进:一种混合建模方法与应用[D].上海:上海交通大学, 2017. Lü J. Study on gradient boosting decision tree and its improvement:An hybrid model and its application[D]. Shanghai:Shanghai Jiao Tong University, 2017.(in Chinese) |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
2023,
Vol. 63
