Hydrogen CPS Digital-Twin IDS Dataset
收藏资源简介:
This dataset contains multivariate time-series measurements representing a hydrogen logistics cyber–physical system that integrates cryogenic storage/transfer operations, SCADA/PLC control, IIoT sensing, fleet logistics telemetry, and cybersecurity network indicators. It is designed to support research on joint cyber intrusion detection and physical safety monitoring under a digital-twin–enabled operating setting, where simulated twin states and real process states may diverge during abnormal events. Each row corresponds to one timestamped observation from a sensor/terminal/tanker context. The dataset includes (i) raw operational signals (pressure, temperatures, flow, vibration), (ii) control and authentication signals from industrial automation, (iii) network/security behavior indicators, (iv) digital-twin synchronization variables, and (v) engineered cross-domain risk indices. The dataset supports both multi-class intrusion classification and continuous leakage severity estimation. File format: CSVTime axis: fixed-interval sampling (see Sampling_Interval)Key identifiers for grouping studies: Terminal_ID (site/client), Tanker_ID (asset), Sensor_Node_ID (device) Targets (labels) Intrusion_Class (multi-class; imbalanced) 0 Normal 1 SCADA valve spoofing 2 Sensor malware 3 PLC firmware intrusion 4 Fleet route hijacking 5 Digital twin tampering Leakage_Rate (regression; kg/hr)Continuous leakage severity signal reflecting cryogenic losses and safety-critical anomalies. Feature description (by category) A) Indexing and entities Timestamp: observation time Terminal_ID: terminal/site identifier (useful as a federated “client”) Tanker_ID: tanker/vehicle identifier Sensor_Node_ID: IIoT sensor node identifier B) Cryogenic tank telemetry (process physics) LH2_Temperature (°C): liquid hydrogen temperature Vapor_Temperature (°C): vapor phase temperature Tank_Pressure (bar): tank pressure reading Pressure_Rise_Rate (bar/min): pressure change rate indicator Boil_Off_Rate (%/hr): boil-off proxy Liquid_Level (%): cryogenic tank fill level Heat_Ingress (proxy units): thermal ingress driver Wall_Stress (MPa): structural stress proxy linked to pressure/vibration Relief_Valve_Status (0/1): relief valve activation indicator Cryo_Pump_Status (0/1): cryogenic pump operational state C) Pipeline and terminal infrastructure (flow dynamics) Pipeline_Pressure (bar): pipeline pressure Flow_Rate (kg/s): hydrogen flow rate Compressor_Load (%): compressor utilization Transfer_Line_Temperature (°C): transfer line temperature Joint_Leakage_Index (%): leakage propensity indicator at joints Flow_Turbulence_Score (score): turbulence/instability indicator Terminal_Throughput (throughput units): terminal processing intensity Pump_Vibration_Level (vibration units): mechanical vibration proxy D) IIoT device telemetry (edge/network health) Sensor_Battery_Level (%): device battery estimate Signal_Strength_RSSI (dBm): radio signal strength Packet_Loss_Rate (%): packet loss indicator Sampling_Interval (minutes): configured sampling rate Edge_Buffer_Load (%): edge buffer utilization Sensor_Calibration_Error (score): calibration drift indicator Device_Tamper_Flag (0/1): device tamper indicator E) SCADA/PLC control and security (automation layer) Valve_Command (categorical): 0 Hold, 1 Open, 2 Close, 3 Partial Command_Source (categorical): 0 Auto, 1 Operator, 2 Remote API Execution_Latency (ms): command execution latency PLC_CPU_Load (%): PLC utilization Firmware_Integrity (0/1): firmware status (1 OK, 0 compromised) Override_Command (0/1): manual override indicator Shutdown_Trigger (0/1): shutdown event indicator Authentication_Status (0/1): authentication verified vs suspicious/failed F) Fleet and logistics telemetry (transport risk) Tanker_GPS_Latitude, Tanker_GPS_Longitude: asset position Route_Deviation_Index (%): route anomaly indicator Transit_Time (hours): travel time estimate Loading_Delay (minutes): loading delay Fleet_Density (count): vehicles in operating region Boil_Off_Transit_Loss (fraction): transit boil-off loss proxy Vibration_Exposure (units): vibration exposure during transit G) Cybersecurity network behavior (IDS signals) Login_Attempts (count) Failed_Logins (count) Access_Time_Anomaly (0/1): abnormal access timing IP_Entropy (score): IP distribution entropy Packet_Size_Variance (bytes² proxy): traffic variability Malware_Flag (0/1) Privilege_Escalation (0/1) Network_Latency_Spike (score): latency anomaly magnitude H) Digital-twin synchronization (twin vs real consistency) Twin_Pressure, Real_Pressure Twin_Temperature, Real_Temperature Twin_Real_Deviation: aggregated deviation between twin and real State_Error_Index: bounded twin-state inconsistency measure Simulated_Valve_Response, Observed_Valve_Response: actuation mismatch signals Twin_Integrity_Score (0–1): estimated twin trust/integrity I) Engineered cross-domain risk features (for learning robustness) Cryo_Instability_Index: cryogenic instability score Leak_Risk_Score: leakage risk composite Cyber_Physical_Risk: fused cyber + physical risk score Valve_Spoofing_Confidence: spoofing likelihood proxy Twin_Tamper_Score: twin tamper likelihood proxy Routing_Risk_Index: transport-side risk composite
本数据集包含多变量时间序列测量数据,表征一套集成了低温存储/转运作业、数据采集与监视控制系统(Supervisory Control And Data Acquisition, SCADA)/可编程逻辑控制器(Programmable Logic Controller, PLC)控制、工业物联网(Industrial Internet of Things, IIoT)传感、车队物流遥测以及网络安全网络指标的氢能物流信息物理系统(Cyber-Physical System, CPS)。本数据集旨在支持数字孪生(Digital Twin)赋能的运行场景下,联合网络入侵检测与物理安全监测的研究——该场景中,异常事件发生时,模拟孪生状态与实际过程状态可能出现偏离。 每一行对应一个来自传感器/终端/油罐车场景的带时间戳的观测样本。本数据集包含以下五类数据:(i) 原始作业信号(压力、温度、流量、振动);(ii) 工业自动化领域的控制与认证信号;(iii) 网络/安全行为指标;(iv) 数字孪生同步变量;(v) 人工构建的跨域风险指数。本数据集可支持多分类入侵分类任务与连续泄漏严重程度估计任务。 文件格式:CSV格式。时间轴:固定间隔采样(详见Sampling_Interval)。用于分组研究的关键标识符:Terminal_ID(站点/客户端)、Tanker_ID(油罐车资产)、Sensor_Node_ID(传感器节点)。 ### 标签(目标变量) 1. **Intrusion_Class(多分类任务,样本不均衡)**: - 0:正常 - 1:SCADA阀门欺骗 - 2:传感器恶意软件感染 - 3:PLC固件入侵 - 4:车队航线劫持 - 5:数字孪生(Digital Twin)篡改 2. **Leakage_Rate(回归任务,单位:kg/hr)**:反映低温损耗与安全关键异常的连续泄漏严重程度信号。 ### 特征分类说明 #### A) 索引与实体类 - Timestamp:观测时间 - Terminal_ID:终端/站点标识符(可作为联邦学习“客户端”使用) - Tanker_ID:油罐车/运输车辆标识符 - Sensor_Node_ID:工业物联网(IIoT)传感器节点标识符 #### B) 低温储罐遥测(过程物理类) - LH2_Temperature(单位:°C):液态氢温度 - Vapor_Temperature(单位:°C):气相温度 - Tank_Pressure(单位:bar):储罐压力读数 - Pressure_Rise_Rate(单位:bar/min):压力变化率指标 - Boil_Off_Rate(单位:%/hr):蒸发损耗代理指标 - Liquid_Level(单位:%):低温储罐充装液位 - Heat_Ingress(单位:代理单位):热侵入驱动因子 - Wall_Stress(单位:MPa):与压力/振动相关的结构应力代理指标 - Relief_Valve_Status(0/1):安全阀激活状态指示器(1为激活,0为未激活) - Cryo_Pump_Status(0/1):低温泵运行状态指示器 #### C) 管道与终端基础设施(流动动力学类) - Pipeline_Pressure(单位:bar):管道压力 - Flow_Rate(单位:kg/s):氢气流量 - Compressor_Load(单位:%):压缩机负载率 - Transfer_Line_Temperature(单位:°C):输送管线温度 - Joint_Leakage_Index(单位:%):管道接头泄漏倾向指标 - Flow_Turbulence_Score(单位:分):流动湍流/不稳定性指标 - Terminal_Throughput(单位:处理量单位):终端作业处理强度 - Pump_Vibration_Level(单位:振动单位):机械振动代理指标 #### D) 工业物联网设备遥测(边缘/网络健康类) - Sensor_Battery_Level(单位:%):设备电池剩余电量估计值 - Signal_Strength_RSSI(单位:dBm):无线电信号强度 - Packet_Loss_Rate(单位:%):数据包丢包率指标 - Sampling_Interval(单位:分钟):配置的采样间隔 - Edge_Buffer_Load(单位:%):边缘缓冲区负载率 - Sensor_Calibration_Error(单位:分):传感器校准漂移指标 - Device_Tamper_Flag(0/1):设备被篡改状态指示器 #### E) SCADA/PLC控制与安全(自动化层类) - Valve_Command(分类变量):0=保持(Hold)、1=开启(Open)、2=关闭(Close)、3=部分开启(Partial) - Command_Source(分类变量):0=自动(Auto)、1=操作员操作(Operator)、2=远程API调用(Remote API) - Execution_Latency(单位:ms):命令执行延迟 - PLC_CPU_Load(单位:%):可编程逻辑控制器CPU负载率 - Firmware_Integrity(0/1):固件状态(1=正常,0=已被篡改/入侵) - Override_Command(0/1):手动强制干预指示器 - Shutdown_Trigger(0/1):停机事件触发指示器 - Authentication_Status(0/1):认证状态(1=验证通过,0=可疑/认证失败) #### F) 车队与物流遥测(运输风险类) - Tanker_GPS_Latitude、Tanker_GPS_Longitude:油罐车资产位置坐标 - Route_Deviation_Index(单位:%):航线偏离异常指标 - Transit_Time(单位:小时):运输时长估计值 - Loading_Delay(单位:分钟):装货延迟时长 - Fleet_Density(单位:台/辆):作业区域内的车辆数量 - Boil_Off_Transit_Loss(单位:比例):运输过程中蒸发损耗代理指标 - Vibration_Exposure(单位:单位):运输过程中的振动暴露量 #### G) 网络安全网络行为(入侵检测系统信号类) - Login_Attempts(单位:次):登录尝试次数 - Failed_Logins(单位:次):登录失败次数 - Access_Time_Anomaly(0/1):访问时间异常指示器 - IP_Entropy(单位:分):IP地址分布熵值 - Packet_Size_Variance(单位:字节²代理):网络流量包大小变异程度 - Malware_Flag(0/1):恶意软件感染指示器 - Privilege_Escalation(0/1):权限提升事件指示器 - Network_Latency_Spike(单位:分):网络延迟异常幅度指标 #### H) 数字孪生同步(孪生与实际系统一致性类) - Twin_Pressure、Real_Pressure:数字孪生模拟压力、实际系统压力 - Twin_Temperature、Real_Temperature:数字孪生模拟温度、实际系统温度 - Twin_Real_Deviation:孪生系统与实际系统的总偏差值 - State_Error_Index:有界孪生系统状态不一致性度量指标 - Simulated_Valve_Response、Observed_Valve_Response:阀门驱动响应不匹配信号 - Twin_Integrity_Score(范围:0–1):数字孪生系统可信性/完整性估计得分 #### I) 人工构建的跨域风险特征(用于提升模型鲁棒性) - Cryo_Instability_Index:低温系统不稳定性得分 - Leak_Risk_Score:泄漏风险综合得分 - Cyber_Physical_Risk:融合网络与物理风险的综合风险得分 - Valve_Spoofing_Confidence:阀门欺骗可能性代理指标 - Twin_Tamper_Score:数字孪生篡改可能性代理指标 - Routing_Risk_Index:运输侧风险综合得分



