Long short-term memory (LSTM) has been widely applied to real-time automated natural gas leak detection and localization. However, LSTM approach could not provide the interpretation that this leak position is localized instead of other positions. This study proposes a leakage detection and localization approach by integrating the attention mechanism (AM) with the LSTM network. In this hybrid network, a fully-connected neural network behaving as AM is first applied to assign initial weights to time-series data. LSTM is then used to discover the complex correlation between the weighted data and leakage positions. A labor-scale pipeline leakage experiment of an urban natural gas distribution network is conducted to construct the benchmark dataset. A comparison between the proposed approach and the state-of-the-arts is also performed. The results demonstrate our proposed approach exhibits higher accuracy with AUC = 0.99. Our proposed approach assigns a higher attention weight to the sensor close to the leakage position, indicating the variation of data from the sensor has a significant influence on leakage localization. It corresponds that the closer to the leakage position, the larger variation of monitoring pressure after leakage, which enhances the detection results’ trustiness. This study provides a transparent and robust alternative for real-time automatic pipeline leak detection and localization, which contributes to constructing a digital twin of emergency management of urban pipeline leakage.@en

Deep learning has been widely applied to automated leakage detection and location of natural gas pipe networks. Prevalent deep learning approaches do not consider the spatial dependency of sensors, which limits leakage detection performance. Graph deep learning is a promising alternative to prevailing approaches as it can model spatial dependency. However, the challenge of collecting real-world anomaly data for training limits the accuracy and robustness of currently used graph deep learning approaches. This study proposes a deep probabilistic graph neural network in which attention-based graph neural network is built to model spatial sensor dependency. Variational Bayesian inference is integrated to model the posterior distribution of sensor dependency so that the leakage can be localized. An urban natural gas pipe network experiment is employed to construct the benchmark dataset, in which normal time-series data is applied to develop our proposed model while anomaly leakage data is used for performance comparison between our model and other state-of-the-art models. The results demonstrate that our model exhibits competitive detection accuracy (AUC) = 0.9484, while the additional uncertainty interval provides more comprehensive leakage detection information compared to state-of-the-art deep learning models. In addition, our model's posterior distribution enhances the leakage localization with the accuracy of positioning (PAc) = 0.8, which is higher than that of other state-of-the-art graph deep learning models. This study provides a comprehensive and robust alternative for subsequent decision-making to mitigate natural gas leakage from pipe networks.@en

Urban gas pipelines usually have high structural vulnerability due to long service time. The locations across urban areas with high population density make the gas pipelines easily exposed to external activities. Recently, urban pipelines may also have been the target of terrorist attacks. Nevertheless, the intentional damage, i.e. terrorist attack, was seldom considered in previous risk analysis of urban gas pipelines. This work presents a dynamic risk analysis of external activities to urban gas pipelines, which integrates unintentional and intentional damage to pipelines in a unified framework. A Bayesian network mapping from the Bow-tie model is used to represent the evolution process of pipeline accidents initiating from intentional and unintentional hazards. The probabilities of basic events and safety barriers are estimated by adopting the Fuzzy set theory and hierarchical Bayesian analysis (HBA). The developed model enables assessment of the dynamic probabilities of consequences and identifies the most credible contributing factors to the risk, given observed evidence. It also captures both data and model uncertainties. Eventually, an industrial case is presented to illustrate the applicability and effectiveness of the developed methodology. It is observed that the proposed methodology helps to more accurately conduct risk assessment and management of urban natural gas pipelines.

@en

Subsea gas release is an industrial hazard that can impose fire hazards on offshore facilities near the gas surfacing area. However, risk assessment of the fire caused by subsea gas release is challenged due to inadequate recognition of the knowledge of subsea gas release mechanism and resulting hazards. At present, minimal researches involving risk assessment of offshore fire resulting from a subsea gas release were reported, and this paper is an extension of the previous works on subsea gas behavior. This paper focuses on modeling fire risk on offshore facilities due to subsea gas release. A numerical simulation is carried out using the Computational Fluid Dynamic technique of Fire Dynamics Simulator (FDS) to analyze fire propagation characteristics and assess the impact of fire on personnel and assets. A probit model is adopted to calculate the probabilities of injury or death caused by fire hazards. This study also investigates the effect of wind speed, gas release rate and the distance between gas pool and platform on fire impacts and casualty probabilities. The present study can support safety measure design to mitigate or avoid the impacts of offshore fire events from subsea gas release.

@en

An incident may propagate to an accident with different severity dependent on its propagation scenarios. Since the accident propagation is a two-way process, the current research is focusing on the one-way analysis. This paper aims to analyze the combined effect of multi-units sources and their interactions during the accident propagation. The bi-directional connectivity diagram (BDCD) is applied to visualize the interactions between multiple process units as hazardous sources. The deployed safety barriers interrupt the connection between the hazardous sources and thus minimize the influence of one BDCD node on another. Through which, the accident propagation is reduced. The proposed method can be suitable to the general accidents, and it is applied to a case study of the LNG terminal station to assess the potential consequences of explosion caused by the leakage, in which the cost of the safety barrier is also considered. The BDCD approach is found more effective than traditional single-hazardous source methods for analyzing the accident propagation of multi-units sources in the chemical plant and achieving intrinsic safety.

@en

Managed Pressure Drilling (MPD) system is widely used in the deepwater drilling operation. Reliability assessment plays a critical role in the MPD system in the management of drilling operation risk and the prevention of blowouts. However, the reliability assessment of the MPD system is challenged due to its sequential operations and multiple processes. Consequently, the present work proposes a sequence-based dynamic reliability assessment method, which focuses on the dynamic modeling of sequential operations for the MPD system by integrating GO-FLOW and dynamic Bayesian Network (DBN). GO-FLOW models are firstly used to define the time interaction between multiple phases for complex systems. A sequence-based mapping method is also proposed for the DBN to construct the reliability model of the MPD system throughout the entire drilling cycle. In the end, the case study analyzed by the proposed framework indicates that the reliability of the MPD system decreases with increasing drilling depth, and the reliability of “tripping in” is highest among four different phases, while the “drilling process” is the lowest. The method provides an important technique that can be implemented with online condition monitoring tools to assess and monitor the reliability of the MPD operation in real-time.

@en

Ensuring the operational safety of subsea pipelines is a growing challenge. It requires monitoring of failure causation factors and detailed safety analysis. This paper proposes an integrated framework for comprehensive safety analysis of subsea pipelines. This framework is developed based on an index-based risk evaluation system, which incorporates interdependency analysis of failure causal factors, hazard coupling analysis and risk grade evaluation in an integrated manner. The Decision-Making Trial and Evaluation Laboratory (DEMATEL) method is utilized to analyze the interdependencies among risk factors. The fuzzy reasoning algorithm is employed to calculate the risk grades of these factors in the evaluation system. This framework has also considered the effect of hazard coupling (combination of hazards) on subsea pipeline failure. The framework is tested on a case study. The case study demonstrates the practicality and usability of the proposed framework. This framework will serve an important tool for swiftly assess and manage the risk of subsea pipeline failures.

@en