Research on the prediction method of fatigue stress response of the bogie end structure based on data − driven approaches

Abstract

As the mileage of railway vehicles continues to increase, the fatigue failure problem of the suspension components at the end of the bogie frame has become increasingly prominent. Clearly defining the fatigue failure mechanism of the suspension components at the end of the bogie frame and predicting their fatigue life have become crucial issues that urgently need to be addressed in vehicle maintenance and operation. When evaluating structural fatigue damage, the accuracy of stress measurement points has a decisive effect on the evaluation results of fatigue damage. However, ensuring the accuracy of structural stress measurement points and conducting long-term monitoring poses challenges. In this paper, the simulation technology is first used to identify the high-risk areas of structural fatigue damage. Subsequently, vibration line tests and stress tests are carried out on the relevant structures to verify the causes of structural failure. Then, a deep-learning algorithm is adopted to develop a method for detecting the structural stress of suspension components based on the acceleration data at the end of the bogie frame. This method is used to evaluate the fatigue damage during long-term operation and solve the problem that it is difficult to diagnose the faults of the suspension components at the end of the bogie frame. This method trains a deep-learning model with the historical data of vibration acceleration and stress, establishes the corresponding relationship between vibration acceleration and stress, and realizes the indirect detection of structural stress.