Wear particle classification considering particle overlapping

Abstract

Fatigue, oxide and spherical particles are the three most common kinds of wear particles in rotate vector (RV) reducer and are often bonded together in ferrography images. However, conventional methods of wear debris identification cannot classify wear particles when the wear particles stick together. To this end, this study presents a novel model of wear debris classification to cope with this problem. Firstly, the Inception-v3 model is utilised to extract the features of wear particles automatically. Then, a new neural network with three classifiers is devised to simultaneously determine whether there are fatigue particles, oxide particles and spherical particles in the ferrography image. Finally, a large number of particle samples obtained from experiments are used to test the proposed method. The results demonstrate that the proposed model can simultaneously recognise the different types wear particles even the wear particles are overlapped. Moreover, separation of wear particles and selection of features are not required in the proposed method, which provides an objective and convenient approach for the recognition of wear particles.