Research on a Safety Dynamic Early Warning Model for Hazardous Chemical Tankers under Multi-Parameter Spatio-Temporal Conditions
-
摘要: 针对现有危化品运输罐车预警系统依赖固定阈值、难以刻画多源时空参数动态耦合关系而导致误报与漏报的问题，研究了基于长短期记忆自编码器与多元状态估计技术（long short-term memory autoencoder-multi-variate state estimation technique，LSTM-AE-MSET）的时空多参数安全动态预警模型。该模型以危化品罐车运行安全动态预警为目标，融合温度、压力、液位、倾角和振动等具有时空属性的多源传感器数据，并引入GNSS、IMU和API获取的道路空间信息，利用LSTM-AE提取时序特征与表征变量间的长期依赖关系，将高维输入映射至低维潜在空间；基于历史健康状态数据构建MSET记忆矩阵，通过状态估计与残差分析实现预警阈值的动态调整。采用田纳西-伊斯曼过程（Tennessee Eastman process，TEP）数据集验证模型有效性，数据集包含41个过程测量变量、12个操纵变量与21种故障模式，涵盖阶跃扰动、随机变化、阀门黏滞等典型异常。模型在TEP数据集上完成训练后迁移至罐车运行场景，用于识别不同时空条件下的温度突变、压力漂移、液位振荡等异常。结果表明，LSTM-AE-MSET模型的平均决定系数(R2) 值为0.621 8，均方误差值为4.301 4；其中位数分别达到0.998 5与0.000 1，表明模型在多数正常运行工况下具有优异的拟合能力；异常检测召回率由57%提升至88%，F1: 分数从60%提升至74%。与单一MSET方法相比，LSTM-AE-MSET模型整体准确率提高2%，漏报率降低72%。本文方法能够增强危化品运输罐车异常状态的识别能力，为复杂运行环境下的实时动态预警提供技术支持。
-
关键词:
- 道路交通安全 /
- 危险化学品运输罐车 /
- 安全动态预警 /
- LSTM-AE-MSET模型
Abstract: Existing early-warning systems for tank trucks transporting hazardous chemicals commonly rely on fixed thresholds and fail to characterize the dynamic coupling among multi-source spatio-temporal parameters, which leads to false and missed alarms. To address this limitation, a spatio-temporal multi-parameter dynamic safety ear ly-warning model was proposed based on a long short-term memory autoencoder and the multivariate state estimation technique (LSTM-AE-MSET). The model integrates multi-source sensor data with spatio-temporal attributes, including temperature, pressure, liquid level, inclination, and vibration, and incorporates road spatial information acquired via the global navigation satellite system (GNSS), the inertial measurement unit (IMU), and map application programming interfaces (API). The LSTM-AE extracts temporal features, captures long-term dependencies among variables, and maps high-dimensional inputs into a low-dimensional latent space; an MSET memory matrix is con-structed based on historical healthy-state data, and the early-warning thresholds are dynamically adjusted through state estimation and residual analysis. The Tennessee Eastman Process (TEP) dataset is used to evaluate the model. The dataset comprises 41 process measurement variables, 12 manipulated variables, and 21 fault modes, covering typical anomalies such as step disturbances, random variations, and valve stiction. After being trained on the TEP dataset, the model is transferred to a tank-truck operating scenario to identify anomalies such as abrupt temperature changes, pressure drift, and liquid-level oscillations under different spatio-temporal conditions. The results show that the model achieves an average coefficient of determination (R2) of 0.621 8 and a mean squared error of 4.301 4, with medians of 0.998 5 and 0.000 1, respectively, indicating excellent fitting performance under most normal operating conditions. Compared with the standalone MSET method, the proposed model increases the anomaly-detection recall from 57% to 88% and the F1-score from 60% to 74%, improves the overall accuracy by 2 percentage points, and reduces the missed-alarm rate by 72%. The proposed method enhances the identification of abnormal op erating states in hazardous-chemical tank trucks and provides technical support for real-time dynamic early warning under complex operating conditions. -
表 1 部分特征验证结果
Table 1. Partial feature verification results
序号 特征 R2分数 1 01EAOTDCSA012_06 0.964 2 2 01LAC31CS021 0.986 5 3 01LAK32CT221 0.985 3 4 01LAK31CT222 0.983 6 5 01LAK30CT223 0.978 0 6 01LAK31CY021 0.962 6 7 01LAK31CY022 0.923 2 8 01LAG31CT221 0.980 7 9 01LAG32CT222 0.981 3 10 01LAG33CT223 0.982 6 
下载: 导出CSV
表 2 原始数据样例
Table 2. Sample raw data
时间 经度/(°) 纬度/(°) 温度传感器1/℃ 温度传感器2/℃ 压力传感器1/kPa 压力传感器2/kPa 当前容积/L 11/18/2023 18:11:47 113.860 523 22.874 985 -5.8 -6.9 440.19 384.23 42.88 11/18/2023 18:12:47 113.860 613 22.874 761 -5.9 -7.0 440.77 385.97 42.87 11/18/2023 18:13:47 113.860 613 22.874 761 -5.9 -7.1 440.25 374.27 42.81 11/18/2023 18:14:47 113.860 591 22.874 798 -6.0 -7.1 439.95 379.20 42.84 11/18/2023 18:16:47 113.860 591 22.874 798 -6.0 -7.2 440.53 386.95 42.79 11/18/2023 18:17:47 113.860 700 22.874 745 -6.0 -7.3 440.77 371.63 42.78 11/18/2023 18:18:47 113.860 700 22.874 745 -6.1 -7.4 440.13 377.11 42.77 11/18/2023 18:20:47 113.860 690 22.874 678 -6.3 -7.4 440.47 383.20 42.72 11/18/2023 18:21:47 113.860 690 22.874 678 -6.4 -7.6 440.62 374.30 42.69
下载: 导出CSV
表 3 迁移数据特征评估结果
Table 3. Evaluation results of migrated data characteristics
平均$R^{2}$ 平均MSE $R^{2}$中位数 MSE中位数 0.6218 4.3014 0.9985 0.0001
下载: 导出CSV
表 4 模型对比结果
Table 4. Model comparison results
模型 类别 精确率 召回率 F1分数 误报数 漏报数 LSTM-AE-MSET 正常 0.97 0.90 0.93 3 380 829 异常 0.64 0.88 0.74 MSET 正常 0.91 0.93 0.92 2 246 2 953 异常 0.63 0.57 0.60
下载: 导出CSV
-
[1] YANG W, ROY N, ZHANG B. Multi-objective transportation route optimization for hazardous materials based on GIS[J]. Journal of Loss Prevention in the Process Industries, 2023, 81: 104954. doi: 10.1016/j.jlp.2022.104954 [2] GUO J, MA K J. Risk analysis for hazardous chemical vehicle-bridge transportation system: a dynamic Bayesian network model incorporating vehicle dynamics[J]. Reliability Engineering & System Safety, 2024, 242: 109732. [3] LIU Y Y, ZHANG M G, XU Z, et al. Risk warning model of road transportation of hazardous chemical tanker based on D-S evidence theory[C]. The 3rd International Conference on Big Data, Artificial Intelligence and Risk Management(ICBAR '23), New York, USA: Association for Computing Machinery, 2023. [4] LIU L P, WANG R, TIAN T, et al. Hazardous materials transportation route optimization considering dynamic wind speeds and risk equity[J]. Journal of Loss Prevention in the Process Industries, 2025, 98: 105748. doi: 10.1016/j.jlp.2025.105748 [5] 李诚信, 柳本民, 廖晨非, 等. 基于BKPE安全场模型的复杂天气状态下的行车风险识别方法[J]. 交通信息与安全, 2025, 43(2): 177-186. doi: 10.3963/j.jssn.1674-4861.2025.02.018LI C X, LIU B M, LIAO C F, et al. A method for traffic risk identification under complex weather conditions based on the BKPE security field model[J]. Journal of Transport Information and Safety, 2025, 43(2): 177-186. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2025.02.018 [6] 柯星安, 赵丹, 王秋鸿, 等. 基于模型群的重型货车交通安全风险因素分析[J]. 交通信息与安全, 2024, 42(4): 72-80. doi: 10.3963/j.jssn.1674-4861.2024.04.008KE X A, ZHAO D, WANG Q H, et al. An analysis of risk factors of traffic safety for heavy trucks based on model group[J]. Journal of Transport Information and Safety, 2024, 42(4): 72-80. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2024.04.008 [7] 毛伟成. 危化品运输车辆在线监测管理系统研究[D]. 成都: 西南交通大学, 2019.MAO W C. Research on online monitoring management system for hazardous chemical transportation vehicles[D]. Chengdu: Southwest Jiaotong University, 2019. (in Chinese) [8] 程思哲. 基于物联网和数据挖掘技术的危化品预警平台的设计与实现[D]. 南京: 南京邮电大学, 2022.CHENG S Z. Design and implementation of hazardous chemical warning platform based on internet of things and data mining technology[D]. Nanjing: Nanjing University of Posts and Telecommunications, 2022. (in Chinese) [9] 张恒清. 危化品车辆监测系统设计及多传感器融合状态判决研究[D]. 秦皇岛: 燕山大学, 2023.ZHANG H Q. Design of monitoring system for hazardous chemical vehicles and research on multi-sensor fusion state decision[D]. Qinhuangdao: Yanshan University, 2023. (in Chinese) [10] WANG B, LI W, KHATTAK Z H. Anomaly detection in connected and autonomous vehicle trajectories using LSTM autoencoder and Gaussian mixture model[J]. Electronics, 2024, 13 (7): 1251. doi: 10.3390/electronics13071251 [11] WANG Z, DAHOUDA M K, HWANG H, et al. Explanatory LSTM-AE-based anomaly detection for time series data in marine transportation[J]. IEEE Access, 2025, 13: 23195-23208. doi: 10.1109/ACCESS.2025.3535695 [12] CHEN X, FAN Y, LIN X, et al. Adversarial data augmentation enhanced LSTM autoencoder for anomaly detection in industrial pipeline networks[J]. Journal of Signal Processing Systems, 2025, 97: 197-207. doi: 10.1007/s11265-025-01962-x [13] 卢东艳, 刘浩, 牛祥华, 等. 基于深度学习的危化品罐车多模态信息智能监测预警模型研究[J]. 化工安全与环境, 2025, 38(9): 10-13.LU D Y, LIU H, NIU X H, et al. Research on intelligent monitoring and early warning model for multimodal information of hazardous chemical tank trucks based on deep learning[J]. Chemical Safety & Environment, 2025, 38(9): 10-13. (in Chinese) [14] CUI Z Y, KE R M, PU Z Y, et al. Stacked bidirectional and unidirectional LSTM recurrent neural network for forecasting network-wide traffic state with missing values[J]. Transportation Research Part C: Emerging Technologies, 2020, 118: 102674. doi: 10.1016/j.trc.2020.102674 [15] 刘涛, 刘吉臻, 吕游, 等. 基于多元状态估计和偏离度的电厂风机故障预警[J]. 动力工程学报, 2016, 36(6): 454-460.LIU T, LIU J Z, LYU Y, et al. Fault early warning of power plant fans based on multivariate state estimation and deviation degree[J]. Journal of Chinese Society of Power Engineering, 2016, 36(6): 454-460. (in Chinese) [16] 孙建平, 赵立慧, 孙博扬, 等. 基于数据驱动与多变量状态估计的汽动引风机故障预警[J]. 热力发电, 2020, 49(1): 78-84.SUN J P, ZHAO L H, SUN B Y, et al. Fault early warning of steam-induced draft fans based on data-driven and multivariate state estimation[J]. Thermal Power Generation, 2020, 49 (1): 78-84. (in Chinese) [17] 费圣城. 基于多元状态估计和模糊综合评价的矿用主通风机健康评估研究[D]. 北京: 中国矿业大学, 2024.FEI S C. Research on health assessment of mine main ventilators based on multivariate state estimation and fuzzy comprehensive evaluation[D]. Beijing: China University of Mining and Technology, 2024. (in Chinese) [18] 浦汉军, 谢小鹏, 周受钦, 等. 基于GPRS的天然气压缩机物联网监控系统[J]. 现代电子技术, 2015, 38(2): 92-95.PU H J, XIE X P, ZHOU S Q, et al. IoT monitoring system for natural gas compressors based on GPRS[J]. Modern Electronics Technique, 2015, 38(2): 92-95. (in Chinese) [19] LYU J, ZHOU S, ZHAO X. Knowledge automation based smart platform for the safe and efficient management of hazardous chemicals supply chain[J]. Basic & Clinical Pharmacology & Toxicology, 2020, 126: 100. [20] LI Y Z, DAI W, ZHU L D, et al. A novel fault early warning method for mechanical equipment based on improved MSET and CCPR[J]. Measurement, 2023, 218: 113224. doi: 10.1016/j.measurement.2023.113224 [21] 贺超群, 马社强. 高速公路隧道区域纵向风险驾驶行为时空特征[J]. 交通信息与安全, 2024, 42(4): 53-61. doi: 10.3963/j.jssn.1674-4861.2024.04.006HE C Q, MA S Q. Spatiotemporal characteristics of longitudinal risky driving behaviors in highway tunnel sections[J]. Journal of Transport Information and Safety, 2024, 42(4): 53-61. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2024.04.006 [22] 张开亮, 茅大钧, 蒋欢春. 基于多元状态估计及模糊综合评价分析的汽轮机故障预警[J]. 电力科学与工程, 2024, 40 (5): 55-63.ZHANG K L, MAO D J, JIANG H C. Fault early warning of steam turbine based on multivariate state estimation and fuzzy comprehensive evaluation analysis[J]. Electric Power Science and Engineering, 2024, 40(5): 55-63. (in Chinese) [23] 孙海丽, 黄炎, 韩兰胜, 等. 基于无监督时空状态估计的信息物理系统细粒度异常诊断[J]. 通信学报, 2025, 46(7): 45-59.SUN H L, HUANG Y, HAN L S, et al. Fine-grained anomaly diagnosis for cyber-physical systems based on unsupervised spatiotemporal state estimation[J]. Journal on Communications, 2025, 46(7): 45-59. (in Chinese) -
点击查看大图
图(9) / 表(4)
计量
- 文章访问数: 19
- HTML全文浏览量: 16
- PDF下载量: 0
- 被引次数: 0