Design,Implementation,and Analysis Of A High-Throughput Wear Debris Image Sensor Based On Magnetic Deposition And CMOS Imaging

It's inevitable for rotating elements,such as bearings and gears,to be worn down during the operation of the machine,which has always been the main cause of the failure.There is a constant desire to online monitor the wear state of the machine so that the abnormal wear state can be warned timely,and several sensors for detecting the wear-generated particles in lubricating oil have been developed,and some success has been achieved.Still,a few shortcomings are restricting the ability of sensors to detect online,which are:1)the competition between the sensitivity and the throughput,2)the chain-form distribution of wear particles,and 3)the size distribution involving the information of micro-sized particles.For this,a novel wear debris sensor employing magnetic deposition and imaging methods is designed and implemented in this thesis,to overcome the above defects and improve the capability for detecting wear particles,especially for particles within 10-50?m.This sensor consists of 1)a cage-shaped electromagnet that can generate a ring magnetic field,2)a large-dimension flow channel,3)an imaging module that has two CMOS.It can deposit the wear debris suspending in lubricating oil onto the surface of the oil reservoir and then image it.Thereinto,the large-dimension flow channel can increase the throughput of the sensor;the ring magnetic field can gather small particles together so that the sensor can perceive them much easier as well as extract the amount;the way of imaging separated wear particles allows the sensor to acquire shape and size of the single particles.It has been proven that this sensor has a maximum throughput above 100ml/min in the test.With the flux of 80±5ml/min,it can perceive the concentration change of ferromagnetic wear particles(1-220?m)in oil sample with concentration of 5-250ppm,the resolvable limit is 5ppm.Moreover,this sensor can get the images of which ferromagnetic wear particles exist in the separated form rather than the chain form,solving the problem that cannot identify the single particles due to the chain-form distribution.The results of image processing have shown that the image of separated wear particles is capable of getting the detailed contour features and shape features of particles above 26-54?m,by which four types of wear particles(circular,regular,irregular,and elongated)can be told apart.Besides,the size distribution of 10-170?m ferromagnetic wear particles can be calculated with the image of separated wear particles.