تشخیص عیوب بالپین در سیستم تعلیق خودرو با استفاده از شبکه‌های عصبی LSTM

Due to the progress of science and technology, the industry's need to upgrade advanced equipment and devices is more important than ever. In the production of mechanical parts, diagnosis of production parts is a vital part of the production process, and separating healthy parts from defective parts before the final product is of great importance. Ballpin parts are a critical part of the vehicle suspension system. It is known that a fault in the vehicle's suspension system causes a lot of life and financial damage to the vehicle and its passengers while driving. Today, much research has been conducted for intelligent fault diagnosis of mechanical parts by modern methods, but no method has been provided for ball-pin diagnosis. This research proposes methods based on frequency analysis, neural networks, and an intelligent fault diagnosis method for ballpins. In the proposed method, acoustic signals are first obtained from the ball pin, and its frequency content is extracted. The extracted frequency content is then applied as input to an LSTM neural network. This neural network is trained to separate healthy and defective parts. The evaluation of the proposed intelligent method has been done using ball pins obtained from an industrial production line. The results of the experimental test of the proposed method show that the accuracy of this intelligent method is 100% and it can accurately detect ballpin faults.