Numerical Combustion Simulation Of Burner Of Flame Cleaner Based On Big Data

The flame cleaning process of the slab adopts high temperature and high speed flame spraying to the surface,oxidation,ablation,melting and dynamic removal of defects to make the slab surface smooth and clean.It is difficult to accurately adjust the structure or process parameters on site due to the variety of steel grades and sizes,the variety of defects,and the large fuel consumption of the billets.Therefore,the structural parameters and process parameters of the flame cleaning machine are optimized through numerical simulation calculation to improve the flame cleaning quality,and the neural network prediction model for the amount of debris removal is established based on the big data on site,which has both the significance of energy conservation and emission reduction and the engineering application value of intelligent production.Based on the production data of the flame cleaner,this thesis analyzes the influence of different structural parameters and process elements on the flow,combustion and heat transfer behavior of propane and oxygen burners of the flame cleaner,and establishes a BP neural network prediction model to predict the amount of chip removal on the upper surface.The numerical simulation of a single group of propane and oxygen burners shows that the combustion heat release makes the jet expand,and the hot vortex is more completely mixed than the cold vortex;The jet Kangda effect can accelerate the combustion process and produce positive effects;However,the formation of cold air flow interlayer after air entrainment has a negative impact.The combustion temperature can be increased by properly increasing the oxygen fuel volume ratio,but the temperature will drop when the oxygen fuel ratio is too high.When the volume flow ratio of O₂ to C₃H₈ is 3.4,the average flame temperature is the highest.The numerical simulation of multiple propane burners shows that the distance diameter ratio mainly affects the flame cleaning area and jet merging behavior.The distance diameter ratio increases,and the cleaning area expands;However,when the flame temperature drops,the flame cannot completely cover the slab surface;The smaller the distance diameter ratio is,the better the secondary jet merging and uniform temperature distribution are.The impact angle mainly affects the travel of flame to the slab surface.With the increase of impact angle,the high temperature zone is closer to the slab surface,and the heat transfer between the flame and the slab surface is strengthened,but the temperature distribution is uneven.The impact height affects the flame stability and surface shear force distribution of the burner.With the increase of impact height,the combustion zone is stable,the flame volume is larger,and the temperature is more uniform;However,the slab edge temperature decreases.The impact height shall be increased appropriately to make the shear force evenly distributed and avoid groove marks.But the impact height is too high,which makes the shear force attenuate too much and reduces the cleaning efficiency.The best working condition is determined by simulation calculation:distance diameter ratio S/d=6,impact angle 35°,impact height H/d=36.To explore the non-linear mapping relationship between the process parameters of the flame cleaning machine and the characteristics of the billet itself and the amount of chip removal,a neural network prediction model of the amount of chip removal of the flame cleaning machine is established with the cleaning time,the oxygen pressure of the upper surface cleaning,the propane pressure of the upper surface cleaning,the slab cutting width,and the slab cutting length as inputs,and the amount of chip removal on the upper surface as output.Most of the relative errors of the prediction model are within 5.5%,and the average relative error is 4.2%,which has certain reliability.