为探讨多重防护机制下的液化天然气(Liquefied Natural Gas，LNG)动力船营运风险特征，提出了一种基于复杂性控制系统作用模式的风险态势分析方法。从系统控制过程视角，采用系统控制过程分析方法(System-Theoretic Process Analysis，STPA)确立LNG动力船系统安全控制的风险致因，揭示基于LNG泄漏三重防护机制的风险因子体系和风险形成机制;引入隐马尔可夫模型(Hidden Markov Model，HMM)，进行无监督学习并确立模型参数，推理多重防护机制下的LNG动力船风险转移特征;结合LNG动力船营运的具体场景，对三重防护机制分别进行仿真并统计风险态势。仿真结果表明，三重防护机制在降低LNG动力船营运风险态势方面具有因子诱导的针对性，在防漏因子、止漏因子和治漏因子诱导下LNG动力船分别趋向一般风险状态、较高风险状态和高风险状态。基于复杂性控制系统作用模式的风险态势分析可为LNG动力船营运的安全控制和风险管理提供管控模式和手段。
Methodology for risk analysis of LNG-fueled vessels under multiple protection mechanisms
To explore the navigation risk characteristics of Liquefied Natural Gas(LNG)-fueled vessels under multiple protective mechanisms,a risk situation analysis method based on the operational patterns of complex control systems is proposed.First,from the perspective of system control processes,the safety control path of LNG-fueled vessels navigation process was constructed and analyzed.The System-Theoretic Process Analysis(STPA)method was employed to analyze the safety control feedback path structure and process of LNG-fueled vessels navigation.By analyzing unsafe system control operations,key risk factors and their interrelationships can be identified,revealing the risk factors system and the mechanism of risk formation based on a triple protection mechanism against LNG leaks.Second,given the premise of stochastic processes,the Hidden Markov Model(HMM)was introduced to quantify risk by defining model risk states and observation states.Through unsupervised learning,the parameters of the HMM model were established to analyze the risk situation characteristics of LNG-fueled vessels under multiple protection mechanisms.Last and not least,combining specific operational scenarios of LNG-fueled vessels,simulations were conducted to assess the risk situations of each component of the triple protection mechanism.Simulation results indicate that the triple protection mechanism exhibits targeted effectiveness induced by specific factors in reducing operational risk scenarios for LNG-fueled vessels.Under the induction of leakage prevention factors,leak-stoppage factors,and leak-treatment factors,LNG-fueled vessels respectively trend towards general risk states,higher risk states,and high-risk states.With the reduction in risk observation levels of relevant factors under various protective conditions,there is a notable trend towards decreased overall risk levels for ships.This substantiates the effectiveness of the"leak-prevention-leak-stoppage-leak-treatment"triple protection mechanism in mitigating LNG leakage risks.The risk situation analysis based on the operational patterns of complex control systems provides a management mode and means for the safety control and risk management of LNG-fueled vessels operations.
safety systematologyLNG-fueled vesselsSystems Theory Accident Modeling and Processes(STAMP)Hidden Markov Model(HMM)risk situation
NETL
NSTL
万方数据
