Global economic growth has long been based on fossil resources, which has led to increasingly serious problems such as climate change and air pollution. As a clean secondary energy source, the hydrogen has proven to be important as an important preservation of the “carbon emission speak and carbon neutrality” due to its advantages of low carbon emissions and high conversion efficiency [1,2]. However, the inherent security risk of hydrogen remains an urgent challenge. His low detonator (0.02 MJ), wide [3,4] Put considerable security risks during production, storage, transport and application. Unknown hydrogen leaves can quickly escalate into catastrophic combustion or explosion events. For example, in May 2019, a hydrogen storage tank exploded, which exploded in a system in gangne, South Korea due to a micro-leid caused by improper high-pressure component cultivation. In a similar way, on August 4, 2021, an undiscovered hose fracture led to an explosion of the hydrogen tank in Shenyang, Province of Liaoning, China, to a hydrogen tank explosion due to insufficient maintenance. These incidents underline the crucial importance of the prompt detection of hydrogen leak in preventing larger accidents and underline the need for robust process management management practices.
Computer fluid dynamics (CFD) have currently become an essential instrument to simulate hydrogen lull disease and consequence analysis [5]. Ra et al. [6] developed a transient CFD model to simulate hydrogen injection in a combustion chamber. Duronio et al. [7] Compared to underexpanded hydrogen and methane jets using large vertebral simulations. Zhang et al. [8] Used detached vertebral simulations with Gasflow-MPI to catch the turbulent hydrogen dispersion in a steam generator. Asahara et al. [9] simulated the under-exposed hydrogen beam from a 0.2 mm opening at 82 MPa using a compressed multi-component-navier-stokes equation solner. Li et al. [10] Combined experiments with CFD for examining various hydrogen beam flow fields and concentration distributions. Luan et al. [11] New calibrated RANS closure coefficients using replacement modeling and SOBOL index analysis, improvement in CFD accuracy. These progress together improve the reliability of hydrogen security ratings.
The primary challenge in achieving a quick and precise evidence of hydrogen leaks lies in the optimal placement of gas detectors [12]. S. Defriatend and J. Kwan Seo [13,14] Integrated detector default probabilities and environmental variables in risk reviews and suggested a weighted probability base placement strategy. Lyu et al. [15] showed that in battery systems for electric vehicles, an optimized double detector configuration could shorten the detection time of the hydrogen by 60 S, which effectively suppresses the thermal runaway. Tsukada et al. [16] Validated an optimized detectorn network through CFD-controlled flag analysis and a small experimental leakage, which achieved timely leak detection. Despite this progress, the inherent complexity of the gas dispersion – through leakage conditions, fluid properties, geometric restrictions and environmental factors – counteracts an important challenge for the placement of the detector [17].
In response to this challenge, researchers have integrated mathematical optimization with CFD simulations. RW Lee [18] developed a detector model based on probabilistic risk coverage, while C. Gencer [19] proposed a maximum expected cover model. S. Jung [20,21] With mixed integer linear and non -linear programming to optimize plant layouts by integrating risks and economic factors. Bellegoni et al. [22] introduced a dual algorithm framework that combines CFD-based gas cloud quantification and economic risk assessment and shows its effectiveness in gasoline storage systems. Barbosa Alves et al. [23] Integrated spatial restrictions on the proximity to leak sources and equipment-heuristic and binary integrated programming models that show that the limited optimization exceeded the conventional 5 m distance in both feasibility and efficiency and achieved detector distances of ≤ 1 m. Zhang et al. [24] Validated a stochastic programming model (SP) by CFD and experimental tests, which confirmed the ability to improve detector performance.
Some scientists have proposed multi -objective optimization models. Idris et al. [25] Adopted a fuzzy multi-multi-linear programming approach with mixed intwerium in order to reduce the number of detectors and the associated risks in combustible gas scenarios. Colombo et al. [26] developed a multi-lens-optimization model for structural health monitoring in order to maximize classification accuracy and to minimize the total cost/risk. Sun et al. [27] the MCDT-SRE model, which minimizes the cumulative detection time, is taken into account, while the scenario probability and reliability is taken into account, and validated its efficiency in a diesel hydration unit. Cheng et al. [28] proposed a layout method based on two covers and reliability and improves the likelihood of leakage detections and system reliability in order to trigger alarms within 10 s. Kang et al. [29] Creates a multi-lens model that concentrated on the minimization of the detection period and the leak volume using offshore default data and achieved better surveillance performance. Zhou et al. [30] Creates a multi-lens optimization model that aims to minimize the cumulative probability of death and to optimize the cost-benefit ratio, to improve the efficiency of detectorn networks and the investment flexibility. A comparative taxonomy of the discontinization literature is shown in Table 1.
Although the optimization of the gas detector layout is naturally a multi-lens problem, it is often simplified to a single objective formulation. In order to close this gap, the present study suggests a new optimization method for the placement of the hydrogen detector, which integrates the sample of the probabilistic scenario, the modeling of multi-factor-losing and hybrid optimization algorithms and aims to achieve a balance between safety costs and identification efficiency. A comprehensive set of leak scenarios was constructed using a Latin hypercube sample that contained variables such as leakage, opening size, direction and environmental wind field. Based on temporary hydrogen concentration data, which were derived from CFD simulations at surveillance points under high-pressure solutions, a one-objective optimization model was developed to minimize the cumulative detection period while the scenario cover was taken into account. Building on this foundation and the following of the cost-benefit analysis frameworks by Health and Safety Executive (HSE), a multi-lens-optimization model was formulated in order to minimize both the cumulative recognition time and the cost-benefit ratio at the same time. The proposed multi-lens-optimization framework was validated by a case study at a hydrogen tank station.