After a train of trains completes daily operation, a system maintenance check will be conducted to ensure the next safe operation of the train.
On December 6, the Intelligent Maintenance System for Rail Vehicles ("Intelligent Maintenance Robot") that debuted at Nanjing 2017 World Smart Manufacturing Conference was the first to realize automated and intelligent detection, which greatly improved the efficiency of train fault identification and detection.
The intelligent maintenance system for rail vehicles was developed by Qingdao Sifang Vehicle Research Institute Co., Ltd. (Sifang Sifang) .The control system consists of two parts: the control center and the image acquisition robot. The control center collects the collected images and schedules the entire inspection process The image acquisition robot collects the detection information of all parts of the train. It mainly consists of the intelligent delivery platform, multi-degree-of-freedom flexible robotic arms, cameras, sensors and surveillance cameras.
The system is the first device in the world that uses highly flexible robots and machine vision fusion technology to perform visual inspections of various parts of a vehicle in full compliance with high-speed train maintenance standards or urban rail train detection standards.
When the train storage for maintenance, start the image acquisition robot for self-test, the robot arrived at the front of the vehicle uniform head position, the robot automatically detects the bottom of the vehicle images collected at the end of the picture; when the robot reaches the tail fixed position, the robot extension Out of the robot arm and reverse the direction of the robot laser positioning to the bogie position, the arm in accordance with the predetermined design into the narrow space at the bottom of the bogie and the vehicle, began to capture the key parts of the train pictures at the same time in the robot arm with the robot forward acquisition The picture, the robot automatically open the area camera, a more detailed bottom HD shooting.
Compared with the traditional manual operation, the intelligent maintenance system can increase over 40% of the inspection range, reduce the missed inspection rate and improve the maintenance efficiency. The train collection information is displayed on the control center display in real time. After the collection, the robot automatically returns to Initial position and shut down.
As the old saying goes, 'There are people Sambo, essence of God', the system also has three magic weapons to ensure that the inspection of the efficient, high-quality, high-definition image acquisition is clear, accurate fault identification and a variety of driver-compatible.
The image acquisition robot can capture high-definition images of key parts of the bogie in all directions. The multi-degree-of-freedom flexible arm carries the image acquisition device to collect the top, side and bottom high-definition images of the bogie. Meanwhile, the line scan camera collects high-definition images of the entire body bottom, The array camera collects high-definition images of the bottom of the bogies to ensure full coverage of the vehicle to be examined.
Fault recognition is accurate
By adopting advanced image feature recognition technology and machine depth learning technology, the accuracy of fault recognition is greatly improved.The collected high-definition images are transmitted to the control center in real time through a wireless network and the image is automatically detected by an image recognition algorithm, for example, Parts are missing, loose bolts, foreign body, fracture, size wear, the system can promptly feedback to the maintenance personnel to verify information.
Compatible with a variety of drive modes
According to the different conditions of maintenance work, the image acquisition robot can work in two ways: hanging or laying the rail on the ground, and is compatible with various working conditions. The maintenance mode can also be customized according to user requirements.