An XGBoost-Based Onsite PGV Prediction Model
-
摘要: 地震动峰值速度（Peak Ground Velocity，PGV）是常用于衡量地震动对建筑结构破坏潜力的参数之一，实时预测PGV大小是重大工程地震紧急处置中的关键技术.为进一步提升PGV预测准确性，提出一种基于极限梯度提升树（Extream Gradient Boosting，XGBoost）的现地PGV预测模型.该模型以台站观测到的P波前3 s的峰值加速度（Pa）、峰值速度（Pv）、峰值位移（Pd）、累计绝对速度（CAV）及卓越周期（Tpd）5种特征参数为输入，以该台站观测到的PGV为预测目标.选取日本K-NET台网记录的102次地震的6 918组加速度记录进行模型训练，89次地震的3 430组加速度记录测试模型的泛化能力.结果表明，相同数据集下，对比基于Pd的PGV预测模型和基于支持向量机的PGV预测模型，基于XGBoost的PGV预测模型的预测值与实测值更趋近1∶1比例关系，且预测误差标准差更小，预测残差均值更接近0，且在中国的实际地震震例上的运行结果良好.基于XGBoost的PGV预测模型可用于现地地震预警地震动峰值的预测.Abstract: Peak Ground Velocity (PGV) is one of the parameter commonly used to measure the damage potential of ground shaking to building structures, and real-time prediction of PGV is a key technology in emergency response to major engineering earthquakes. To further improve the accuracy of PGV prediction, in this paper it proposes an onsite PGV prediction model based on Extreme Gradient Boosting (XGBoost). The model takes five characteristic parameters, including peak acceleration (Pa), peak velocity (Pv), peak displacement (Pd), cumulative absolute velocity (CAV), and predominant period (Tpd) in the first 3 seconds of the P-wave observed at the station as inputs, and the PGV observed at the station as the prediction target. 6 918 sets of acceleration records from 102 earthquakes recorded by the K-NET station network in Japan were used for model training, and 3 430 sets of acceleration records from 89 earthquakes were used to test the generalization ability of the model. The results show that, within the same dataset, the PGV prediction model based on XGBoost has a predictive value that is closer to a 1:1 ratio with the actual measured values compared to the PGV prediction models based on Pd and support vector machines. Additionally, the standard deviation of the prediction errors is smaller, the mean of the prediction residuals is closer to zero, and the model performs well on actual earthquake cases in China. The PGV prediction model based on XGBoost can be used for the prediction of peak ground motion in local earthquake early warning systems.
-
Key words:
- earthquakes /
- early warning /
- onsite warning /
- XGBoost /
- machine learning /
- peak ground velocity /
- engineering geology
-
图 1 所选记录台站及震中分布
a.训练集；b.测试集
Fig. 1. Distribution of selected stations and earthquake epicenters
图 3 训练集与测试集震中距分布对比
Fig. 3. Comparison of epicentral distance distributions between training and test sets
图 4 XGBoost-PGV模型在数据集上的预测结果
a.训练集；b测试集
Fig. 4. Predicted results of XGBoost-PGV model on the dataset
图 5 XGBoost-PGV模型在数据集上的残差图
a.训练集残差与震中距的关系；b.训练集残差与震级的关系；c.测试集残差与震中距的关系；d.测试集残差与震级的关系
Fig. 5. Residual plots of XGBoost-PGV model on the dataset
图 6 基于SHAP值的特征全局重要性排序
Fig. 6. Global importance ranking of feature based on SHAP value
图 8 XGBoost-PGV模型和Pd-PGV模型在测试集上预测与实测PGV的比较
Fig. 8. Comparison of predicted PGV and measured PGV between XGBoost-PGV model and Pd-PGV model on the test set
图 9 XGBoost-PGV模型和SVM-PGV模型在测试集上预测与实测PGV的比较
Fig. 9. Comparison of predicted PGV and measured PGV between XGBoost-PGV model and SVM-PGV model on the test set
图 10 积石山$ M $S6.2地震的震中和台站位置
Fig. 10. Epicenter and station distribution of the Jishishan $ M $S6.2 earthquake
图 11 积石山$ M $S6.2地震的PGV预测结果
Fig. 11. PGV prediction of the Jishishan $ {M}_{\mathrm{S}} $6.2 earthquake
图 12 XGBoost-PGV模型在P波S波时间间隔不足3 s数据上的预测结果
a.训练集；b.测试集
Fig. 12. Predicted results of XGBoost-PGV model on the data with a P-wave to S-wave time interval of less than 3 s
图 13 XGBoost-PGV模型在浅地壳地震和俯冲带地震数据上的预测结果
a.浅地壳地震；b.俯冲带地震
Fig. 13. Predicted results of XGBoost-PGV model on the set of Shallow crustal earthquake and subduction earthquakes
图 14 测试集7.3级地震数据上的特征参数分布
a. Pa与震中距的关系；b. Pd与震中距的关系；c. CAV与震中距的关系；d. Pv与震中距的关系；e. Tpd与震中距的关系
Fig. 14. Distribution of feature parameter on the test set of the magnitude 7.3 earthquake
表 1 模型关键超参数及其取值范围
Table 1. Key hyper-parameters and their value ranges in the model
参数名称 简写 超参数取值范围 搜索步长 决策树的深度 max_depth [3, 9] 1 决策树的数量 n_estimators [50, 600] 1 子节点最小权重 min_child_weight [1, 6] 1 损失函数下降值 gamma [0, 1] 0.1 学习率 learning_rate [0.01, 0.5] 0.05 
下载: 导出CSV
表 2 模型超参数的取值
Table 2. Model hyper-parameter values
参数名称 简写 超参数取值 决策树的深度 max_depth 6 决策树的数量 n_estimators 512 子节点最小权重 min_child_weight 2 损失函数下降值 gamma 0.5 学习率 learning_rate 0.3
下载: 导出CSV
表 3 潜在地震破坏和PGV之间的关系
Table 3. The relationship between potential earthquake damage and PGV
潜在破坏类别 无 非常轻微 轻微 中等 中等/严重 严重 非常严重 PGV（cm/s） 0~3.4 3.4~8.1 8.1~16 16~31 31~60 60~116 > 116
下载: 导出CSV
-
Allen, R. M., Gasparini, P., Kamigaichi, O., et al., 2009. The Status of Earthquake Early Warning around the World: An Introductory Overview. Seismological Research Letters, 80(5): 682-693. https://doi.org/10.1785/gssrl.80.5.682 Brondi, P., Picozzi, M., Emolo, A., et al., 2015. Predicting the Macroseismic Intensity from Early Radiated P Wave Energy for On-Site Earthquake Early Warning in Italy. Journal of Geophysical Research: Solid Earth, 120(10): 7174-7189. https://doi.org/10.1002/2015jb012367 Chen, T. Q., Guestrin, C., Chen, T. Q., et al., 2016. XGBoost. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, 785-794. https://doi.org/10.1145/2939672.2939785 Colombelli, S., Caruso, A., Zollo, A., et al., 2015. A P Wave-Based, On-Site Method for Earthquake Early Warning. Geophysical Research Letters, 42(5): 1390-1398. https://doi.org/10.1002/2014gl063002 Faenza, L., Michelini, A., 2011. Regression Analysis of MCS Intensity and Ground Motion Spectral Accelerations (SAs) in Italy. Geophysical Journal International, 186(3): 1415-1430. https://doi.org/10.1111/j.1365-246X.2011.05125.x Hildyard, M. W., Rietbrock, A., 2010. TPD, a Damped Predominant Period Function with Improvements for Magnitude Estimation. Bulletin of the Seismological Society of America, 100(2): 684-698. https://doi.org/10.1785/0120080368 Hoshiba, M., Kamigaichi, O., Saito, M., et al., 2008. Earthquake Early Warning Starts Nationwide in Japan. EOS, Transactions American Geophysical Union, 89(8): 73-74. https://doi.org/10.1029/2008eo080001 Hoshiba, M., Ohtake, K., Iwakiri, K., et al., 2010. How Precisely can We Anticipate Seismic Intensities? A Study of Uncertainty of Anticipated Seismic Intensities for the Earthquake Early Warning Method in Japan. Earth, Planets and Space, 62(8): 611-620. https://doi.org/10.5047/eps.2010.07.013 Hua, S. B., Xu, C. Y., Zhou, J. C., et al., 2024. Source Characteristics and Disaster Mechanisms of the 2023 Gansu Jishishan Mw6.0 Earthquake. Chinese Journal of Geophysics, 67(7): 2625-2636 (in Chinese with English abstract). Huang, F. M., Ouyang, W. P., Jiang, S. H., et al., 2024. Landslide Susceptibility Prediction Considering Spatio-Temporal Division Principle of Training/Testing Datasets in Machine Learning Models. Earth Science, 49(5): 1607-1618 (in Chinese with English abstract). Huang, X. H., Li, Z. H., Deng, T., et al., 2023. Uranium Potential Evaluation of Zhuguangshan Granitic Pluton in South China Based on Machine Learning. Earth Science, 48(12): 4427-4440 (in Chinese with English abstract). Iervolino, I., Giorgio, M., Galasso, C., et al., 2009. Uncertainty in Early Warning Predictions of Engineering Ground Motion Parameters: What Really Matters?. Geophysical Research Letters, 36(5): 2008GL036644. https://doi.org/10.1029/2008gl036644 Jin, X., Zhang, H. C., Li, J., et al., 2012. Research on Continuous Location Method Used in Earthquake Early Warning System. Chinese Journal of Geophysics, 55(3): 925-936 (in Chinese with English abstract). Kanamori, H., 2015. Earthquake Hazard Mitigation and Real-Time Warnings of Tsunamis and Earthquakes. Pure and Applied Geophysics, 172(9): 2335-2341. https://doi.org/10.1007/s00024-014-0964-y Kohler, M. D., Cochran, E. S., Given, D., et al., 2018. Earthquake Early Warning ShakeAlert System: West Coast Wide Production Prototype. Seismological Research Letters, 89(1): 99-107. https://doi.org/10.1785/0220170140 Li, T. L., Jiang, P., Jin, Y. X., et al., 2024. Construction and Application of China Earthquake Early Warning Technology Testing Platform. Earthquake Research in China, 40(1): 69-84 (in Chinese with English abstract). doi: 10.3969/j.issn.1001-4683.2024.01.006 Li, Y. Z., Yang, C., 2024. Overall Design of Customized Software for the National Seismic about the National Seismic Intensity Rapid Reporting and Early Warning Project. Earthquake Research in China, 40(1): 85-96 (in Chinese with English abstract). doi: 10.3969/j.issn.1001-4683.2024.01.007 Liu, C., Li, X. J., Jing, B. B., et al., 2019. The Distance Segmentation Characters of PGV-Pd Relationship Parameters for Earthquake Early Warning. Chinese Journal of Geophysics, 62(4): 1413-1426 (in Chinese with English abstract). Liu, L., Shen, J. K., Zhang, L. X., 2023. A Machine Learning-Based Method for Rapid Prediction of Earthquake Damage in Brick Masonry Houses. Earth Science, 48(5): 1769-1779 (in Chinese with English abstract). Liu, Y. Q., Zhao, Q. X., Wang, Y. W., 2024. Peak Ground Acceleration Prediction for On-Site Earthquake Early Warning with Deep Learning. Scientific Reports, 14(1): 5485. https://doi.org/10.1038/s41598-024-56004-6 Lu, J. Q., Li, S. Y., Xie, Z. N., et al., 2021. Real-Time Evolutionary Earthquake Location Method and Its Uncertainty Analysis. Journal of Natural Disasters, 30(3): 52-62 (in Chinese with English abstract). Lundberg, S. M., Lee, S. I., 2017. A Unified Approach to Interpreting Model Predictions. Proceedings of the 31st International Conference on Neural Information Processing Systems, New York, 21889700. https://doi.org/10.48550/arXiv.1705.07874 Luo, H. Y., Xu, Q., Jiang, Y. N., et al., 2024. The Prediction Method of Large-Scale Land Subsidence Based on Multi-Temporal InSAR and Machine Learning. Earth Science, 49(5): 1736-1745 (in Chinese with English abstract). Ma, Q., 2008. Study and Application on Earthquake Early Warning (Dissertation). Institute of Engineering Mechanics, China Earthquake Administration, Harbin (in Chinese with English abstract). Murphy, S., Nielsen, S., 2009. Estimating Earthquake Magnitude with Early Arrivals: A Test Using Dynamic and Kinematic Models. Bulletin of the Seismological Society of America, 99(1): 1-23. https://doi.org/10.1785/0120070246 Nielsen, D., 2016. Tree Boosting with XGBoost-Why does XGBoost Win "Every" Machine Learning Competition? (Dissertation). Norwegian University of Science and Technology, Norway. Olson, E. L., Allen, R. M., 2005. The Deterministic Nature of Earthquake Rupture. Nature, 438(7065): 212-215. https://doi.org/10.1038/nature04214 Ouyang, L. B., Su, Z. J., Liu, J., et al., 2024. Construction of Decision-Making Platform Fused in China Earthquake Early Warning Network. Earthquake Research in China, 40(1): 24-39 (in Chinese with English abstract). doi: 10.3969/j.issn.1001-4683.2024.01.003 Peng, C. Y., Ma, Q., Jiang, P., et al., 2020. Performance of a Hybrid Demonstration Earthquake Early Warning System in the Sichuan-Yunnan Border Region. Seismological Research Letters, 91(2A): 835-846. https://doi.org/10.1785/0220190101 Picozzi, M., Zollo, A., Brondi, P., et al., 2015. Exploring the Feasibility of a Nationwide Earthquake Early Warning System in Italy. Journal of Geophysical Research: Solid Earth, 120(4): 2446-2465. https://doi.org/10.1002/2014jb011669 Song, J. D., Yu, C., Li, S. Y., 2021. Continuous Prediction of Onsite PGV for Earthquake Early Warning Based on Least Squares Support Vector Machine. Chinese Journal of Geophysics, 64(2): 555-568 (in Chinese with English abstract). Wald, D. J., Quitoriano, V., Heaton, T. H., et al., 1999. Relationships between Peak Ground Acceleration, Peak Ground Velocity, and Modified Mercalli Intensity in California. Earthquake Spectra, 15(3): 557-564. https://doi.org/10.1193/1.1586058 Wu, Y. M., Kanamori, H., 2005. Experiment on an Onsite Early Warning Method for the Taiwan Early Warning System. Bulletin of the Seismological Society of America, 95(1): 347-353. https://doi.org/10.1785/0120040097 Zhang, H. F., 2023. A Study of Real-Time PGV Prediction Method with Regional and Onsite Integration (Dissertation). Institute of Engineering Mechanics China Earthquake Administration, Harbin (in Chinese with English abstract). Zhang, J. Y., Xi, N., Xu, T. R., et al., 2024. The Center Technology Platform for National Earthquake Intensity Quick Reporting and Early Warning. Earthquake Research in China, 40(1): 54-68 (in Chinese with English abstract). doi: 10.3969/j.issn.1001-4683.2024.01.005 Zhang, Y. Y., Miao, C. L., Chen, J. L., et al., 2024. Design and Construction of the Earthquake Waveform Comprehensive Analysis System for National Seismic Intensity Rapid Reporting and Early Warning Project. Earthquake Research in China, 40(1): 40-53 (in Chinese with English abstract). doi: 10.3969/j.issn.1001-4683.2024.01.004 Zhao, J. X., Zhou, S. L., Gao, P. J., et al., 2015. An Earthquake Classification Scheme Adapted for Japan Determined by the Goodness of Fit for Ground-Motion Prediction Equations. The Bulletin of the Seismological Society of America, 105(5): 2750-2763. https://doi.org/10.1785/0120150013 Zhu, J. B., Song, J. D., Li, S. Y., 2021. Rapid Magnitude Estimation for Earthquake Early Warning Based on SVM. Journal of Vibration and Shock, 40(7): 126-134 (in Chinese with English abstract). 华思博, 徐晨雨, 周江诚, 等, 2024.2023年甘肃积石山MW6.0地震震源特征与灾害机理. 地球物理学报, 67(7): 2625-2636. 黄发明, 欧阳慰平, 蒋水华, 等, 2024. 考虑机器学习建模中训练/测试集时空划分原则的滑坡易发性预测建模. 地球科学, 49(5): 1607-1618. doi: 10.3799/dqkx.2022.357 黄鑫怀, 李增华, 邓腾, 等, 2023. 基于机器学习的华南诸广山花岗岩体铀矿潜力评价. 地球科学, 48(12): 4427-4440. doi: 10.3799/dqkx.2022.006 金星, 张红才, 李军, 等, 2012. 地震预警连续定位方法研究. 地球物理学报, 55(3): 925-936. 李同林, 江鹏, 晋云霞, 等, 2024. 中国地震预警技术测试平台建设与应用. 中国地震, 40(1): 69-84. 李雨泽, 杨陈, 2024. 国家地震烈度速报与预警工程定制软件总体设计. 中国地震, 40(1): 85-96. 刘辰, 李小军, 景冰冰, 等, 2019. 地震预警PGV-Pd关系参数的距离分段特征. 地球物理学报, 62(4): 1413-1426. 刘丽, 沈俊凯, 张令心, 2023. 基于机器学习的砖砌体房屋震害快速预测方法. 地球科学, 48(5): 1769-1779. doi: 10.3799/dqkx.2022.481 卢建旗, 李山有, 谢志南, 等, 2021. 实时演化地震定位方法及其不确定性分析. 自然灾害学报, 30(3): 52-62. 罗袆沅, 许强, 蒋亚楠, 等, 2024. 基于时序InSAR与机器学习的大范围地面沉降预测方法. 地球科学, 49(5): 1736-1745. doi: 10.3799/dqkx.2023.048 马强, 2008. 地震预警技术研究及应用(博士学位论文). 哈尔滨: 中国地震局工程力学研究所. 欧阳龙斌, 苏柱金, 刘军, 等, 2024. 国家地震烈度速报与预警工程融合决策平台建设. 中国地震, 40(1): 24-39. 宋晋东, 余聪, 李山有, 2021. 地震预警现地PGV连续预测的最小二乘支持向量机模型. 地球物理学报, 64(2): 555-568. 张海峰, 2023. 区域与现地融合的PGV实时预测方法研究(硕士学位论文). 哈尔滨: 中国地震局工程力学研究所. 张建勇, 席楠, 徐泰然, 等, 2024. 国家中心地震烈度速报与预警技术平台. 中国地震, 40(1): 54-68. 张莹莹, 苗春兰, 陈经纶, 等. 2024. 国家预警工程综合地震波形分析系统设计与实现. 中国地震, 40(1): 40-53. 朱景宝, 宋晋东, 李山有, 2021. 基于支持向量机的地震预警震级快速估算研究. 振动与冲击, 40(7): 126-134. -
点击查看大图
计量
- 文章访问数: 28
- HTML全文浏览量: 6
- PDF下载量: 5
- 被引次数: 0
