Study On Explosion Risk Analysis And Mitigation Design Of Hydrogen Production Plant

Hydrogen is a colorless and odorless gas.Due to its environmental protection,pollution-free,high combustion calorific value,and renewable advantages,hydrogen is one of the clean energy sources to alleviate the current energy shortage in the future.At the same time,hydrogen,as a flammable and explosive gas,has a wide range of explosions and is prone to explosion accidents,causing casualties and property losses.However,most of the current analysis and research on hydrogen explosion accidents focus on hydrogen refueling stations with a relatively open hydrogen storage environment,and there is less research on hydrogen production devices that obtain hydrogen.However,the module equipment of the hydrogen production plant is congested,and hydrogen leakage is prone to gas cloud accumulation,which may lead to explosion accidents.Therefore,it is particularly important to carry out hydrogen explosion potential risk analysis and explosion disaster mitigation design for hydrogen production devices.This paper studies the hydrogen production device of a typical hydrogen production plant,and has obtained the following research results:(1)Hydrogen leakage and diffusion combustible gas cloud prediction modelBased on the Bayesian Regularization Network(BRANN),using FLACS software to simulate and calculate the Q9 and Q6 data,construct a diffusion-BRANN prediction model for predicting the volume of gas clouds.It is used to predict the volume of combustible gas cloud in different scenarios.The model is used to predict the combustible gas cloud,and the predicted results are compared with the simulation results of the FLACS software to verify the accuracy and generalization ability of the model.The model is used to provide a large amount of accurate non-simulated prediction data of the combustible gas cloud for the hydrogen production device.Simplify the risk analysis process of the hydrogen production unit and reduce the time and cost of risk analysis.(2)Hydrogen gas cloud explosion overpressure prediction modelBased on the Bayesian Regularization Network(BRANN),a large amount of data in typical explosion scenarios generated by FLACS software simulation calculations are used to construct an explosion-BRANN prediction model that predicts explosion overpressure.It is used to predict the explosion overpressure in different gas cloud scenarios.And use the model to predict the explosion overpressure of the combustible gas cloud,and provide a large amount of non-simulated prediction data of the explosion overpressure of the combustible gas cloud for the hydrogen production device.(3)Risk analysis of explosion of hydrogen production plantBy constructing a parameter probability model related to hydrogen leakage and diffusion,coupled with the OLF ignition probability model,a parameter probability model in different scenarios related to hydrogen diffusion-explosion is established.Combining the diffusion-BRANN prediction model and the explosion-BRANN prediction model,the risk analysis of the device is carried out from the perspective of probability and consequence,and an explosion risk analysis(ERA)method for high-congestion hydrogen production plants is proposed.The performance of this method is verified by analyzing the sensitivity of the parameters beyond the frequency curve and comparing it with the traditional ERA method.(4)Optimal design of water spray parameters for hydrogen production plantThe explosion model of the hydrogen production device under different water spray scenarios is constructed.Based on the numerical results of the explosion overpressure under different water spray scenarios,a preliminary design method of water spray parameters based on risk analysis is proposed.And use the BRANN prediction model to predict the pressure reduction effect of different water spray scenes,and combine with the ABC optimization algorithm to obtain the optimal water spray parameters.