Methods for diagnosing engineering objects based on neural networks
Abstract
Methods for diagnosing engineering objects based on neural networks
Incoming article date: 21.08.2020Rail transport plays a key role in the transport of passengers and goods. The high demand for effective means of flaw detection of high-speed rails, which is an important component of railway safety, is justified by the constantly growing number of vehicles and their speed [2]. Failure of elements of the railway infrastructure entails huge losses. In this regard, the analysis and forecasting of the state of elements of the railway infrastructure is an urgent task. This article analyzes modern methods for diagnosing the state of a rail track from a mobile bogie or a conventional locomotive. A method for diagnostics of a rail track is proposed, based on registration of vibrations near the rolling surface of rails and processing these signals using neural networks.
Keywords: rail flaw detection, wheel surface, algorithm, IMM, MEMS, control system, inertial sensor, rail defect, contactless measurement