Automatic detection of failure of missing bolts at the bottom of motor car

The bolts on the bottom brake shoe of the motor car play a key role in the overall train braking system. The loss of bolts in the brake shoe will pose a serious threat to the safe braking and safe driving of the train. Taking bolt loss failure detection as an example, an online detection and identification algorithm for missing parts and components in high-speed trains is proposed, which provides a guiding method for fault diagnosis of key parts of high-speed trains.

Combining the characteristics of bolt geometric structure, a complete local binary model feature extraction algorithm based on Sobel gradient edge of the image is proposed, combined with the training and learning of the binary classifier, to complete the automatic detection of bolt loss failure. The results show that the proposed algorithm has strong robustness in identifying bolt loss failures in complex scenarios, and its detection efficiency and accuracy are also high, which can meet the needs of field applications.

With the rapid development of China’s railways, China’s high-speed railway technology has become increasingly mature. High-speed railway construction is not only in full swing at home, but also actively promoted abroad. “High-speed rail diplomacy” has become a new business card in my country’s diplomatic field. The rapid development of high-speed railways brings convenience to people’s travel, but also brings challenges to their own maintenance.

The long-term high-speed operation of the train and the large and diversified number of parts may cause various problems in the operation of the high-speed train. The traditional manual maintenance method is not only low in efficiency, but also greatly affected by subjective factors. It is difficult to adapt to the high-efficiency and high-quality maintenance requirements of high-speed trains. Therefore, it is an urgent need in the field of railway transportation to use the relevant knowledge of computer vision and image understanding to realize the online detection of train component failures. The bolts on the brake shoe at the bottom of the motor car are important structural parts that connect the brake caliper and the brake shoe.

The loss of the bolts will cause the train to fail to brake, and even cause major accidents such as train derailment. Aiming at the problem of bolt loss failure, this paper proposes a complete local binary pattern (CLBP) algorithm based on Sobel gradient preprocessing, combined with machine learning knowledge, to realize whether the bottom bolt of the motor car is lost.

Automatic identification of bolt loss failure

EMU is a type of high-speed railway, and its operation mainly has the following characteristics:

1) The train is fixed;

2) The structure of the same type of vehicle is the same;

3) The structure of the body parts is unified;

4) The train running line is fixed.

The position of the bolts at the two wheel axles on the train bogie is fixed, and the precise external trigger method of the wheel axle is used to start the camera to take pictures, which can obtain the image of the precise position of the bolts.