Design And Study Of Flotation Intelligent Dosing System

At present,the medicament addition is mainly artificial in the flotation system,which will inevitably cause gross error,result in the waste of the medicament and cannot reach the precise dosage.The existing automatic dosing system is mainly based on metering pump.The cost of the metering pump is high,and the pulse flow rate is generated,which causes that the return valve ball in the medium cannot reset in time,and it affects the accuracy of the adding agent seriously.In order to resolve this problem and achieve precise control of the small flow of medicament,the paper selects the high precision micro flow control pump--peristaltic pump as the medicament added actuator on the basis of a large number of experimental data.On this basis,a dosing tank is designed,of which the structure includes medicine chest,impurity removal device,shelter cabinet and mounting bracket.This paper makes the removal process more efficient through the design and optimization of the case removing device.Through the structural design of the box,the feeding pressure is reduced,the stability of the material flow is ensured,and the pulse problem of the medicament is solved,which provides a guarantee for the safe,reliable and effective work of the peristaltic pump.And the control system includes the automatic impurity removal system,liquid level automatic detection and adding system,and the flow control system of the drug executing agency,which can ensure the automation of the case adding.In this paper,according to the control process of the dosing system,the off-line data is collected as the basic training data of the model,and the optimal dosage of the dry slime is taken as the correction data in real time.The on-line variable is used as the input data of the model,and the amount of the collector and the dosage of the foaming agent is used as the output data of the model.In order to build the control process of the dosing model,we need to design the following related experiments to support it.The paper selects the coal sample of GUOBEI as the experimental background to carry out the step release test and optimal drug system exploration test.The test shows that flotation feed is an easy floating coal quality,and when the ratio of the kerosene and SEC is 4,the yield and ash comprehensive index is the best.Based on the PyCharm software,different algorithm models were constructed,and the off-line data of floating concentration,floating ash,floating tail ash,collector and foaming agent flow were used as model input,and the training and fitting effect of the model were compared.The paper selects the best fitting Lasso model to deploy the dosing model,in which the paper takes online floating concentration,floating ash and tailings ash as input variables to predict dosage of collector and foaming agent.Therefore,some performance tests based on the image recognition are carried out in this paper.In order to verify the accuracy of on-line detection device for the input variables needed to add the model,this paper performs the performance test on the hardware devices such as concentration meter.The results show that the relative error of the online detection device can be controlled within 5%,and the detection result is reliable.Because there is no equipment to detect the ash content of ore slurry online,this paper takes the extraction of the grayscale characteristic value of the picture by collecting the pictures of flotation tailings,and establishes the mathematical model to achieve the soft measurement of the ash content of flotation tailings.The model fitting R value is 0.956,and the fitting effect is good.For the ash content of floating pulp,the relationship model between floating ash and original coal ash is established,and the formula of the model is derived.Finally,two aspects of the application effect of the dosing system are explored,including control accuracy and adaptability to flow change.The test of control precision of adding medicine shows that the relative error of the measured data is more than 1% without correction.When the value of flow correction width is 50,small amount of data for the measurement of relative error is less than 1%.When the flow exceeds 500ml/min,the relative error suddenly increased to 3.80%,which indicates that the dosing device is more accurate for the small flow,and the large flow dosing error will increase.The flow changes adaptability test shows that after calibrating the flow of 450ml/min,the relative error of 325ml/min-450ml/min interval is less than 1%,of which the correction width is 125,and the relative error of 450ml/min-475ml/min interval is small,but it still exceeds 1%,of which the correction width is less than 25.Therefore,for the small flow,the adaptability of the system to the flow change is better.For the large flow,the adaptability of the system to the flow change is poor.