Real-time Measurement Of Burner Coal Input Based On Video Signal

Real-time coal consumption measurement of utility boilers is always one of the most critical information in the process of boiler modeling.However,the actual coal quantity is measured before the pulverized coal mill.The time uncertainty of pulverizing time brings difficulties to the determination of real-coal input.Based on the combustion image taken at the burner nozzle,this paper extracts the pulverized coal area in the image using the depth learning technology,and establishes the relationship between the pulverized coal area and the pulverized coal quantity signal,thus realizes the real-time measurement of the pulverized coal quantity in the burner based on the video signal.In addition to the pulverized coal area and combustion area,there may be coking and occlusion in the flame image obtained from the scene.Traditional image processing methods can not effectively solve such problems.In this paper,a flame image segmentation model based on convolution neural network is proposed.The flame image can be divided into different parts,from which the pulverized coal area can be obtained.The pulverized coal area itself contains two characteristics:the size of the area and the concentration of pulverized coal in the area.Firstly,the area measurement method based on boundary tracking method and the pulverized coal concentration measurement method based on pixel gray value are given.After two characteristics of pulverized coal area are obtained,two main factors affecting the characteristics of pulverized coal area are analyzed:coal quantity and furnace negative pressure.Through theoretical analysis and observation of actual data,it can be found that given relatively stable furnace negative pressure,the area increases with the increase of the coal quantity,and so do the concentration;given relatively stable coal quantity,the area increases with the increase of furnace negative pressure,but the concentration decreases.Therefore,the characteristics of area and concentration of pulverized coal are fused,and a characteristic parameter is obtained,which only has a strong correlation with the coal quantity signal,but has a weak correlation with the negative pressure signal of the furnace.Testing under different working conditions,this parameter can characterize the coal quantity information,taking the error into consideration,thus realizing the real-time measurement of coal quantity of Burner Based on video signal.