چكيده به لاتين
Nowadays, timing belts of cars are one of the essential parts of most cars. In this research, while introducing, classifying and rooting out the appearance defects in belts, desirable studies and activities in three different areas such as statistical studies, intelligent timing belt`s diagnosis and also simulation (analytical) fields have been performed.
In the statistical field, more than 330 timing belts were collected in the field. Statistics show that the defects that appear for replacement belts over 75 thousand kilometers are almost twice as much as the range of 60-45 thousand. In addition, it was found that most car owners change their car belts within 60,000 kilometers. In the field of diagnosis, an intelligent system for belt diagnosis, suitable for technical inspection centers as well as at the factory level, was proposed. In total, two diagnostic programs based on the concepts of deep learning and convolutional neural network were written and proposed. The sensitivity and productivity of the performance of these two programs in detecting defects was over 98% and showed more efficiency compared to other programs. In the analytical field, various simulations were performed about the defects of the timing belt and various parameters such as belt material, type of tooth profile, number of tensile members, etc. were investigated. The results showed that EPDM belts (suggested in this study) and HTD tooth profiles had the best performance.
