Simulation Study On Smoke Temperature Distribution And Flow Law In High-rise Stairwell

In the past decade,China’s urbanization rate has risen from 51.3% at the end of2011 to 64.72% at the end of 2021,and is experiencing the largest and fastest urbanization process in the world’s history.While this large-scale and high-rate urbanization has injected strong momentum into economic and social development,the contradiction between decreasing urban land resources and increasing urban population has become increasingly prominent,and urban High-rise buildings have become the mainstream,which also brings more fire hazards than before.In case of fire,due to the chimney effect,the fire smoke will spread rapidly in the vertical direction in high-rise buildings,and the stairwell is one of the channels for the smoke to spread,and the stairwell is the main path for the evacuation of people in case of fire,so it is especially important to pay attention to it.Most of the previous studies on smoke flow characteristics in stairwells have used full-size experiments,scaled-down experiments and numerical simulations,which cost a lot of labor and material resources,but the experimental data obtained are still limited.In this paper,based on the previous research,different methods are used to study the flue gas flow characteristics of stairwells,and the relevant experimental data can be obtained by changing the parameters within the working conditions,which provides more research data for the subsequent research on the flue gas flow characteristics in high-rise buildings.The specific work of the thesis is as follows.The influencing factors of flue gas temperature are studied,and the quantitative relationships of the dimensionless temperature of flue gas in the stairwell,the dimensionless power of the fire source and the dimensionless area of the smoke vent in the stairwell are obtained by using the dimensionless variables in combination with the dimensional analysis,and then the numerical model is established by the experimental data,and the rationality of the numerical model is analyzed and verified under the same working conditions.Later,using several sets of experimental data compiled from the numerical model under different working conditions,the theoretical values of the temperature in the stairwell were obtained by fitting the dimensionless equations for the flue gas temperature Ts,the fire source power Q and the smoke vent area Aa based on multiple linear regression,and the reliability of the fitted equations was verified.Later,based on the dimensionless relational equation as well as the fitted equation,the temperature distribution of the flue gas plume can be obtained by inputting known basic physical quantities under different working conditions.The analysis and study of the simulation results show that under the conditions of different fire power and smoke vent area in the same stairwell,the calculated results are compared with the simulation results based on the temperature distribution in the stairwell,using the hydrostatic equation,the thermal mass conservation equation,and the ideal gas equation of state to verify the calculation by calculating the difference between the neutral surface position of the stairwell and the pressure intensity inside and outside the opening under different fire power The results are compared with the simulation results to verify the feasibility of the calculation.The results show that the effect of fire power on the neutral surface position is small for the same opening.As the fire power increases,the neutral surface shifts downward to a small extent,while the pressure difference between the top and bottom of the stairwell becomes larger and larger,and the chimney effect is enhanced.In the case of constant power,with the increase of smoke vent area,the neutral surface shifts upward significantly,the pressure difference from the bottom opening to the neutral surface gradually increases,the pressure difference from the top opening to the neutral surface decreases,the overall pressure difference between the two openings increases,and the chimney effect is enhanced.Using the multiple sets of data obtained from the simulation,the traditional BP neural network model was optimized using genetic algorithm based on MATLAB software platform,the connection weights and thresholds of the network were optimized,and the corresponding neural network model training and prediction procedures were written.The errors relative to the measured sample values were calculated separately in comparison with the GAs,and it was concluded that both GAs and GA-BP neural networks can reliably predict fire smoke parameters,among which GA-BP neural networks are computed rapidly and have more research potential but higher application threshold,and GAs are computationally relatively complex but easy to be popularized.