34 / 2023-05-15 10:36:37

Dynamic probabilistic risk assessment of emergency response for intelligent coal mining face system, case study: Gas overrun scenario

Dynamic probabilistic risk assessment,Human-system interaction,emergency response,Intelligent coal mine,Dynamic Bayesian network

Introduction: The safe mining of intelligent coal mine (ICM) directly affects the safe, efficient, green, and sustainable development of the coal industry. Most miners’ task responsibilities may change from field operation to remote monitoring and control, and intelligent coal mining operations depend more on the human-system interaction. Moreover, ICM construction introduces new challenges, such as the increased cyber security threats, possible communication interruption, and data reliability. The objective existence of safety risks is an important feature of coal industry, work safety is always the top priority. Specifically, emergency response/decision-making level is particularly important for ensuring the work safety of ICMs, which urgently needs an accurate risk assessment method to assess the risk level of ICM emergency response. Method: Combining the ICM system structure and emergency response process of human-system interaction based on the information, decision, and action in crew context (IDAC) model, this study proposes a dynamic probabilistic risk assessment (PRA) method based on dynamic Bayesian network (DBN), mainly including the event sequence diagram (ESD) from the cognitive perspective, human-system concurrent task analysis (CoTA) from success perspective and fault tree analysis (FTA) model from failure perspective, dependency model of performance influencing factor (PIF) from organizational impact perspective, and the “ESD-FTA-PIFs” coupled BN model. To address the uncertainty of risk factors (e.g., human factors and system equipment) and achieve quantitative analysis, the coupled BN model is mapped to the DBN model to perform dynamic PRA combining with existing databases, fuzzy set theory, Dempster-Shafer (D-S) evidence theory, and time-variant model. Finally, the practicality and advantages of the proposed method are demonstrated by a real case study in the gas overrun scenario. Results: The results of dynamic PRA of various emergency response events show that the risk probabilities of each event range from 0.01 to 0.08, and the risk probabilities increase as the time step increases. The risk probabilities of Event D (0.0582-0.0752) and E (0.0524-0.0671) in the diagnosis and decision-making stage are relatively high. These are caused by the job characteristics of diagnosis and decision-making process (e.g., high task complexity, more system components, and relatively long required time), which is also reflects the actual situation of emergency response in ICM operations. The influence analysis results of human and system equipment show that equipment related to diagnosis and decision making, action execution, and communication have significant impacts on the risk of events and system states. The CFMs with significant impacts on events or systems are mainly related to human decision-making and monitoring, which is consistent with the results of dynamic risk assessment and further verifies the correctness of the constructed DBN model. The PIFs of S4, S10, S11, S15, S17, S18, and S25 have more significant impacts on the risk level, which demonstrates that ICM operations have higher requirements for the abilities of decision-making and interactive operations. Conclusions: Compared with the conventional BN model which only performs probabilistic inference once, the DBN model can perform temporal dynamic inference to achieve the prediction of event risk probability. The comprehensive integrated DBN model of “ESD-FTA-PIFs” not only enables bidirectional risk analysis of emergency response event, but also enables the systematic traceability of risk factors. Practical applications: The proposed method can be a useful tool to assess the dynamic risk of ICM emergency response and provide decision supports for safety and emergency management, and it can provide method and scientific references for risk assessment during emergency response of other safety-critical industries.