A Sensing And Classification Method For Ball Mill Fill Level Based On FSR And SVM

This paper presents a new classification method of Ball Mill Fill Level based on ball's compressive force signal by technically combining force sensor resistor and SVM neural network.A ball shaped equipment is invented to receive the real-time signal of compressive force on its surface,and transmit these data to computer through MCU inside of it.SVM neural network works on these data,namely training samples,and then predict the current Ball Mill Fill Level status.DEM simulations provide reasonably realistic predictions of ball motion and force condition based on contact parameters measured by single-particle tests,and wireless sensing is used for its convenience to send out the Ball mill's force data.This approach provides the theoretical and experimental prototype for a more effective and convenient method of Ball Mill Fill Level classification,and the experimental results show that the proposed method is feasible.Based on the different state of motion,the steel ball surface under compression stress is different,so we can continuously monitor the classification of coal mill current coal loading capacity of the numerical signal to identify.In order to verify the theory,a ball mill steel ball coal model is established.In the ball mill and the design,use Solidworks to build 3D model file and imported into EDEM software,through the EDEM unique "particle factory" function of steel ball and coal mixture model,the physical parameters and the corresponding import,software simulation.In the experiment part,a ball size equipment was designed by Auto CAD and completed by 3D printing.After covering on the part of the surface of FSR and circuit boards in the shell internal,pressure data was sent into computer.In data processing,SVM analysis the FSR pressure sensor data as the recognition results in the formation of a sample,after training the classification recognition.The results proved significantly.