A fault diagnosis system for rail transit platform doors based on deep learning

Abstract

The invention discloses a rail transit platform door fault diagnosis system based on deep learning, which includes a rail transit platform door system, a door machine transmission system, a unit door control system and a fault diagnosis system. Sequence control is used to control the switch closure of the rail transit platform door system; the fault diagnosis system collects the feedback signals of the monitoring sensors in real time, and uses the server training to generate a fault diagnosis model, and performs real-time fault diagnosis on the operating status of the platform door through the fault diagnosis model . The present invention makes full use of the neural network that can automatically learn implicit features from large-capacity multi-modal data instead of artificially designed feature models.

Description

technical field [0001] The invention relates to a rail transit platform door fault diagnosis system based on deep learning, belonging to the field of rail transit platform door fault diagnosis. Background technique [0002] Rail transit platform doors are installed on the edge of the platform to isolate the track area from the platform waiting area, prevent passengers from falling off the track, reduce the number of drivers looking around, reduce the loss of hot and cold air on the platform, reduce the energy loss of the platform air-conditioning system, and reduce the noise generated by train operation , dust, etc., improve the safety and comfort index of passengers waiting for the bus, and have the functions of safety, energy saving, environmental protection, and beauty. In addition, the safety, reliability and quickness of operation and maintenance of rail transit platform door systems will directly affect the transportation safety and efficiency of urban rail transit. T...

AI Technical Summary

Problems solved by technology

[0007] Aiming at the inaccurate fault diagnosis and the waste of time and energy caused by manual data analysis in the prior art using some significant indicators manually set objectively for fault diagnosis and analysis, the present invention provides a depth-based The fault diagnosis system for rail transit platform doors learned, using the convolutional neural network (CNN) of deep learning to automatically extract the multi-dimensional features of the data collected by the monitoring sensors, realizes the deep mining of a large number of multi-modal data information, which can be very good Quickly and effectively respond to abnormal data, so that it can respond well to complex and changeable rail transit platform door systems

Images