Safety Analysis Of Hydrogen Fuel Cell Vehicle Leakage And Explosion Accident In An Underground Parking Garage

Currently,fossil fuels like coal,natural gas,and oil,still occupy the main position of world energy consumption.Greenhouse gases are continuously emitted into the atmosphere in the process of using large amounts of traditional fossil energy sources,leading to serious damage to the environment on which human beings rely for survival.The frequent occurrence of extreme weather in recent years is representative of this problem.For a better tomorrow and to slow down the process of global climate change,mankind has been seeking a low-carbon or even carbon-free energy source to replace fossil fuels.Hydrogen energy may be the optimal solution to this problem.A growing number of countries have reached this consensus and have begun to focus on the development of the hydrogen energy industry.Hydrogen energy is now beginning increasingly in the transport sector.However,the safety of hydrogen as an energy carrier has always been a major concern for the engineering community.The wide flammability range and low ignition energy of hydrogen require special attention to its safety.The diffusion and explosion of hydrogen gas after leakage is an important direction of hydrogen safety research,which is important for the development and improvement of hydrogen safety guidelines.Therefore,in this paper,a series of experiments,numerical simulations,and quantitative risk assessments are carried out for a hydrogen fuel cell vehicle in an underground parking garage.In the present work,a scale-down experimental platform was built to simulate the hydrogen leak and dispersion accidents in a large underground garage based on similarity principles.Helium gas,which has the closest physical properties to hydrogen gas,was used as a substitute for hydrogen to ensure the safety of the experiments.A series of experiments were conducted by varying the leak location,leak duration,and leak flow rate,and recording the helium concentration over time using 12 sensors placed on top of the model.The experimental results were processed and compared using the dimensionless analysis method.The results show a clear "three-stage"distribution of hydrogen concentration inside the semi-enclosed space.After the sensor detects a change in concentration,there is a clear rise in concentration,followed by a gradual stabilization of the helium concentration,and a clear decrease in concentration at each monitoring point when the leak stops.After dimensionless processing of parameters such as geometric length,leakage time,and concentration,it can be found that the trend of concentration changes in each group of conditions shows a clear consistency.This paper also combines machine learning technologies to develop an algorithm for locating the leak source in an underground garage.The concentration data obtained from the model test were divided into a training set and a test set in the ratio of 8:2 after noise reduction and normalization.In this paper,a neural network model,a deep residual network model,and a K-DTW model were used for leak localization.The neural network can only use 12 sensor data at a specific time as an input,which may ignore some temporal features,and the final accuracy of the algorithm was 80.88%.The deep residual network model and the K-DTW model made full use of the chronological concentration information and the prediction accuracy was higher.The results showed that the above algorithms can be used for leak source localization.The hydrogen fire event is one of the hydrogen safety accidents that can directly cause great harm to people.In this paper,an actual size model of the underground garage was established using the FLACS software,and the consequences of the accident such as over-temperature and over-pressure that may be caused by different leakage directions and leakage locations were compared by numerical simulations.The simulation results showed that only jet flames were formed for an upward release and the overpressure phenomenon was not obvious.However,the ignited downward jet resulted in an explosion and the car body was surrounded by flames to form a wide range of high-temperature areas.There were obvious over-temperature and over-pressure phenomena inside the car when the leak occurs inside the vehicle,which might cause fatal injuries to people inside the car.The numerical simulation results were then combined with quantitative analysis methods to calculate the probability of fatal injury to people using the numerical simulation results as inputs.The accident impact range calculated based on the numerical simulation showed good agreement with the accident range derived from the quantitative analysis.