Development of Machine Vision System to Track Movement of an Engine Timing Belt Tensioner Based on Deep Neural Network

Belt tensioner is one of the main components in timing mechanism of IC engines. This component has an angular motion to adjust the timing belt tension and ensure the continuous connection between crankshaft and camshafts. This angular motion is one of the important parameters in design of the belt tensioner, which is evaluated during validation tests. The present paper, with the aim of using in modern validation tests, presents a new method for measuring the kinetic characteristics of the belt tensioner. The proposed method is a machine vision system that uses a deep neural network to track the movement of the belt tensioner indicator and then obtain its motion characteristics including angular displacement, velocity and acceleration. To this end, the engine test was designed and performed so that the belt travels its entire stroke. Simultaneously the movement of the belt tensioner was captured with a camera. The results showed that the tensioner indicator was correctly tracked with about 80% accuracy during the whole test. The maximum angular displacement of this component reached 14 degrees at the end of the test. The results also showed that the belt tensioner under study moved with the maximum angular speed and acceleration of 1 rad/s and 61 rad/s2. In addition, it was found that the frequency of the belt tensioner movement reached about 10 Hz. The results showed that the proposed method could be an appropriate alternative to conventional methods to measure the tensioner movement.