Retarder multi-node fault classification method of RBF neural network

Abstract

The invention discloses a retarder multi-node fault classification method of an RBF neural network. The method comprises the steps: 1, taking the peak-to-peak value, kurtosis, frequency domain index,fluctuation entropy and temperature mean value of a retarder motion curve as elements to construct an RBF neural network input feature vector; 2, constructing three working states of the retarder intoRBF neural network output quantities; 3, constructing an RBF neural network structure, training the neural network according to the training data set, and constructing a retarder working state classifier; and 4, constructing a decision fusion method based on a Bayesian approximation method evidence theory, and further evaluating the three-dimensional feature vector output by the RBF neural network to complete the diagnosis of the working state of the retarder. According to the method, multi-node heterogeneous data are fused based on the RBF neural network, diagnosis of the working state of the retarder is completed, and the method has wide application prospect.

Description

technical field [0001] The invention belongs to the field of fault diagnosis, and specifically relates to a multi-node fault classification method of a deceleration top of an RBF neural network. The method learns an input matrix of a deceleration top through an RBF neural network, and realizes fault classification of all nodes through a classifier obtained through training. . Background technique [0002] Railway vehicle deceleration jack (deceleration jack) is a kind of speed regulating equipment widely used in marshalling yards at home and abroad. The popularization and use has completely changed the operation mode of the marshalling yard, and basically eliminated the "turning brake" and "iron shoe" braking, which not only reduces personal casualties, but also reduces the labor intensity of the staff. Multiple retarder devices constitute the retarder speed regulation system, and the failure of a single device will reduce the safety connection rate. In field operations, i...

AI Technical Summary

Problems solved by technology

[0003] At present, there are mainly three detection methods for retarders installed on rails: manual inspection, indoor retarder comprehensive inspection equipment and retarder detection vehicles. Excavate useful information from it to form the law; the comprehensive function of the retarder detection equipment occupies a large space, and is generally installed in the laboratory. During the test, the retarder needs to be disassembled and reinstalled, and its dynamic detection performance is poor; the working condition of the mobile retarder The detection vehicle relies on manpower to push on the rails, and the position is detected by infrared and the reaction force test is performed. The detection speed of the equipment is slow, it consumes a lot of human resources, and the positioning is not accurate.

Images