Optimization monitoring distribution method for gas pipeline leakage detection in underground spaces

Highlights

• An optimized distribution model for gas pipeline leakage detection was developed.

• Total risk of underground space system was calculated using the micro-element method.

• Relationship between the number of monitoring points and total risk could be obtained.

• The rate of risk control of the proposed model is twice that of the conventional model.

Abstract

Gas leakage from buried gas pipelines in urban areas can lead to accidents involving fire and explosion when the gas gets concentrated into the adjacent underground spaces. Determining monitoring points for the gas leakage in the underground spaces can prevent the initiation of fire and explosion. In this regard, this study proposes an optimized distribution model which relies on risk prediction. It maps the fire and explosion risk in the underground spaces to the discrete target pipeline based on the effect predicted by the monitoring sensors. Moreover, the total risk in this system is calculated through the micro-element method to design an effective distribution optimization strategy. A case study is conducted to illustrate the effectiveness of the new approach and compare it with the risk-based distribution method and effective monitoring length method. The results show that determining the optimized distribution plan is difficult using the risk-based distribution method and effective monitoring length method because these methods may determine a large number of monitoring points or cannot determine the specific location of the monitoring point. The proposed optimization model enables to derive the relationship between the number of distribution points and the risk in the system. For the same number of monitoring points, the rate of risk control in the system of the proposed model is twice that of the conventional model. As the number of monitoring points decreases, the monitoring cost for the prevention of fire and explosion would be largely reduced.

Introduction

In recent years, consumption of natural gas has rapidly grown on a global level. Since 2010, the average consumption of natural gas has increased globally by 1.8% per year (Alvela et al., 2018). However, the current development of natural gas pipelines faces several challenges (Dong et al., 2017, Zou et al., 2018). In recent years, with the popularization of applications of natural gas, the scale of gas pipelines in urban regions has rapidly expanded. Due to the complexity of underground pipelines in urban regions, it is observed that with the leak in a gas pipeline, the leak rapidly spreads to adjacent underground spaces and causes large-scale explosions (Liaw, 2016, Qiu et al., 2018, Wang et al., 2017, Zhang and Zhang, 2014). The adjacent underground spaces are the adjacent confined spaces in the underground and other underground pipelines in the gas pipeline network such as sewage pipelines, water supply pipelines, underground supermarket, and underground parking lot. Zhang (2016). Recently, numerous explosions related to these pipelines have occurred worldwide (Du, 2007, Fu, 2009, Lu et al., 2020) that have resulted in severe consequences. Therefore, the prevention of fire and explosion related to gas pipelines in the underground region is paramount for urban safety (Zhang et al., 2016, Tang et al., 2018).

Currently, two methods for the detecting and monitoring of leaks in urban gas pipelines exist: Supervisory Control and Data Acquisition (SCADA) monitoring system (Liu et al., 2018, Zhang, 2012), pipeline valve leakage monitoring system (Li et al., 2017), and manual periodic inspection (Datta and Sarkar, 2016, Hasan and Iqbal, 2006). However, the SCADA system cannot detect small leaks in pipe networks (Li et al., 2019). On the other hand, manual regular inspection has the disadvantage of poor real-time performance.

Monitoring the adjacent spaces that are prone to accumulation of natural gas after the leakage in the buried gas pipeline (e.g., rainwater and sewage wells, gas valve wells, power wells) (Hou et al., 2017, Lu et al., 2018, Yan et al., 2015, Zhang, 2016) is important. Therefore, optimizing the layout of the monitoring points can effectively reduce the fire and explosion risk caused by a gas leakage, with an adequate warning concerning large-scale fire and explosion events (Tsinidis et al., 2019a).

The approach for optimizing the distribution of the monitoring points can be divided into classical mathematical methods, intelligent optimization algorithms, and system evaluation methods. The classic mathematical method relies on developing a model based on the distribution of the monitoring points and the subsequent use of logistics planning algorithms to solve the model for obtaining the optimal planning solution (Qi et al., 2019, Wang et al., 2005, Xu et al., 2010;). However, with improving the boundary conditions, the calculation volume index increases; thus, the calculation process becomes very complicated (Tsinidis et al., 2019b). The swarm intelligence algorithm has been developed and successfully applied to monitor the point distribution optimization. It includes genetic algorithms (Ayvaz and Elci, 2018, Caputo et al., 2015, Garnier et al., 2007), particle swarm optimization (Garnier et al., 2007), and harmonious search algorithms (Verma et al., 2010). These algorithms use the fitness function to eliminate older solutions with low fitness. Simultaneously, new solutions were provided for data processing through different mechanisms. The convergence speed of the algorithms depends on a variety of factors, such as the different approaches of encoding, conception, and threshold selection. Moreover, system evaluation methods are applied to specific distribution optimization problems by analyzing and screening from a system evaluation perspective, which is performed rather than planning the distribution of the monitoring points (Shan et al., 2014).

The underground adjacent spaces of the urban pipelines are complex and diverse due to unreasonable design (Bonnaud et al., 2018, Wang et al., 2020). Describing accurately and systematically the risks of the pipeline leakage and fire are challenging, and the explosion due to the leak locations is difficult to predict (Zhang et al., 2015, Ebrahimi-Moghadam et al., 2016). Therefore, in this study, we propose a novel optimization monitoring distribution method for gas pipeline leakage detection in underground spaces. The rest of this paper is organized as follows. Section 2 establishes a risk assessment method for underground space explosions based on the deployment of the monitoring points. An optimization method is proposed to eliminate the monitoring points with the weakest control effect sequentially. Section 3 considers a real scenario as an example to show the effectiveness of the proposed approach, which is compared with the effective monitoring length method (Zhang, 2016) and the risk-based distribution method (Hou et al., 2017).