| As an important by-product of iron and steel production,blast furnace(BF)slag has extremely high temperature and contains a large amount of waste heat can be recovered.At present,water quenching is the main method of BF slag treatment,however this technology not only wastes a lot of heat and water,but also causes environmental pollution.Therefore,dry granulation heat recovery technologies of BF slag have become the focus of current research,centrifugal granulation has become the most feasible dry waste heat recovery technology for BF slag because of its advantages such as high quality of waste heat recovery,good granulation effect and strong device reliability.Unfortunately,this technology is not mature at present.The technical bottleneck lies in the difficulty in measuring and adjusting the flowrate of BF slag,under the condition that the performances of granulation and heat recovery are directly related to the flowrate.Therefore,this paper studies the flow rate measurement and regulation of high temperature molten BF slag,in order to provide theoretical guidance for the application of the centrifugal granulation and heat recovery system.Based on image identification and Python language,this paper develops the flow rate measurement program of BF slag under Anaconda environment,then studies on the application of low temperature simulated working medium and high temperature slag respectively.At first,the feasibility of image identification method was proved,and the characteristics of slag liquid column image were analyzed.Based on this,an algorithm for slag liquid column flow rate was designed.The influences of cooling conditions,shooting aperture,initial temperature of molten slag on the measurement accuracy of were studied.Then,the effects of shooting parameters and data processing methods on the measurement accuracy were systematically studied by using glycerine solution,which has the similar properties to molten BF slag.Finally,the influence factors of flow rate were studied with glycerin solution.The influence rules of liquid level,flow area,shape of stopper rod,viscosity and other parameters on the flow rate were clarified,the flow rate was simulated by neural network method and the slag flow regulation system was designed and studied.The main conclusions are as follows:(1)Molten BF slag liquid column has clear edge,obvious and stable color characteristics,so it is very suitable for flow rate measurement of image identification method;Images of high-temperature slag liquid column need to be preprocessed first,Canny operator is more suitable for the edge recognition;higher initial temperature,higher cooling intensity and appropriate shooting parameter can improve the measurement accuracy;The error of the BF slag flow rate measurement by program is within 15%.(2)For glycerin solution,exposure time should be selected reasonably,on the one hand,too short exposure time will create dark pictures,which lost the details,on the other hand,too long exposure time will cause over exposure phenomenon;When the frame rate is 400 fps,the measurement error is only 0.74%,if the frame rate is lower than that,the information will be easily lost,and if the frame rate is higher than that,the rounding error will occur;In order to obtain more image details,the object distance should be as short as possible without affecting the normal use and working life of the camera.When the object distance is 0.4m,the measurement error is only 1.2%;Region of interest(ROI)also has an important impact on measurement accuracy.In order to avoid areas with sharp changes in diameter and flow velocity,the ROI should be 1/3~2/3 pipe diameter in length,and at 4 times pipe diameter from the nozzle.(3)Flow rate of glycerine solution increases with the increase of liquid level,besides,when the liquid level is less than 10 times the nozzle diameter,the flow coefficient will change with the liquid level;Stopper rod is a powerful flow rate regulating method,which affects both the flow area and the flow coefficient.The stopper rod with small head Angle is suitable for precise regulation of flow rate,and it is difficult to regulate flow area when the flow area ratio is lower than 40%;The flow rate of molten slag is greatly affected by the viscosity,so the BF slag should be well insulated to prevent the flow rate fluctuation greatly from the viscosity change in centrifugal granulation process;BP neural network with 11 nodes in hidden layer has strong ability of simulation for molten slag flow rate,so it is suitable for the research of molten slag flow rate regulation.(4)BF slag flow regulation system with a daily processing capacity of 100 tons was studied,the structure parameters and system adjustment parameters of the tunic was obtained.The user image interface(GUI)for the slag flow regulation system was designed,which lays a foundation for the subsequent development of the system. |
PDF Full Text Request