Study On Flow Characteristics Of High Temperature Buoyant Jet And Optimizing Side Suction Hood Based On Bp Neural Network

In industrial buildings,various pollutants（particulate matter,sulfur dioxide and carbon monoxide）and waste heat,which are often emitted during the production process such as molten steel and casting,are expressed in the form of high-temperature dust-containing flue gas.The physical properties are:high temperature and high concentration.Because the high temperature flue gas temperature is high and the particle size of most dust particles in the airflow is less than 1.0 um,the high temperature dust flue gas is simplified to high temperature buoyant jet for treatment in this paper.The local ventilation system is effective for controlling the above-mentioned high-temperature pollution airflow,and is suitable for a place where a high-temperature pollution source exists.At present,research on buoyant flow is mostly focused on the study of flow field distribution and empirical formula,but there are few studies on the vortex structure of high-temperature buoyant jet.A deeper understanding of the vortex structure of the high-temperature buoyant jet can lay a theoretical foundation for the design of the local exhaust system under the action of high-temperature buoyant jet.Taking the flow field characteristics of high-temperature buoyant jet as the main research object,the following contents are studied and analyzed by the combination of experiment and simulation:Firstly,the large eddy simulation is used to simulate the buoyant flow at different initial temperatures（high temperature buoyant flow and lower temperature）.The vortex structure（vortex ring,spiral vortex,etc.）and the evolution process of the temperature field formed by the buoyant flow are analyzed.The flow characteristics of the high-temperature buoyant jet will affect the capture of the buoyant flow by the exhaust hood.Secondly,on the basis of the calculation formula of the limit flow ratio of the existing side hood,the effects of the horizontal width,installation height,buoyant jet and environmental temperature difference on the limit flow ratio are studied,and the corresponding correction formula is given;the BP neural network combined with CFD is used for the optimization of the side suction hood,and the MATLAB is used for programming to establish a BP neural network model suitable for the side hood capture efficiency prediction.The main conclusions are as follows:（1）For the buoyant jet of Ar₀=0.12（T₀=30℃,V₀=1m/s）,the temperature distribution takes a mushroom-like form over time and has an axisymmetric flow structure.For the buoyant flow under Ar₀≥4.83（T₀≥400℃,V₀=1m/s）,the temperature distribution of the mushroom-like axisymmetric structure at the beginning of the motion.As the buoyant jet moves upward,the temperature distribution exhibits a left-right asymmetrical shape,the shape of the buoyant jet will swing to the left and right,and a contraction section appears in the temperature cloud image.The larger the Ar₀,the lower the position of the contraction section of the buoyant jet.（2）The buoyant jet shows the streamline characteristics that the reverse rotating vortex pair mainly moves downstream of the fluid.The reverse rotating vortex pair surrounds both sides of the main body of the buoyant jet.When the temperature of the buoyant jet is low,it mainly occurs at the tail of the buoyant jet motion.With the increase of Archimedes number（up to Ar₀=4.83）,there will also be reverse rotating pairs near the exit of the buoyant jet.As the fluid moves downstream,the two sides of the buoyant jet will be entrained to form multiple vortices.（3）Large-scale axisymmetric vortex rings are formed in the flow field of buoyant jet,which dominate the flow field.With the increase of Archimedes number（the temperature of buoyant jets increases）,the number of vortex rings increases,the intensity of vortices increases,and the initial position is closer to the exit of buoyant jets.For the buoyant jet of Ar₀≥4.83,with the increase of flow height,the fluid in the flow field stretches and grows,and the vortex rings fall off.The vortex structure of the buoyant jet presents the form of spiral vortices and vortex rings.（4）The existing formula for calculating the limit flow ratio of side suction hood does not take into account the influence of horizontal width of side suction hood,and the influence of installation height of side suction hood is not accurately described.It is only applicable to the case of contaminated air flow and ambient temperature difference of 0℃.It is found that when the installation height of the side suction hood is constant,the limit flow ratio increases with the increase of the horizontal width of the side hood.The higher the installation height of the side hood,the greater the change of the limit flow ratio with the horizontal width;The maximum flow error ratio formula can achieve a maximum error of 67%between the calculation result and the numerical simulation of the side suction hood at different dimensionless installation heights,and the installation height influence term of the side hood limit flow ratio ratio needs to be corrected;With the increase of the installation height of the side suction hood,the slope of the limit flow ratio curve increases with the temperature.By analyzing the variation of the limit flow ratio of the side suction hood with different influencing factors,the formula for calculating the limit flow ratio of the upper side suction hood of the high temperature buoyant jet is amended and supplemented,and the fitting formula for the limit flow ratio of the high temperature buoyant jet is obtained.（5）The neural network algorithm combined with CFD is used to predict the efficiency of side suction hood capture.The mapping relationship between the influencing factors of the side suction hood and the side suction hood trapping efficiency is established by BP neural network.It is verified that the BP neural network model can be used for the prediction of pollutant trapping efficiency by the side suction hood（applicable conditions:H/D=0～0.75,L/D=1～1.6,Δt=100～1200°C,U₀=1～2.5m/s）.According to the intensity of a certain pollution source,the simulation of the neural network model is carried out,and the change of the prediction value of the capture efficiency of the side suction hood under different influencing factors is analyzed,which can provide suggestions for the optimization of the side suction hood.