Construction of a Surrogate Model for Train Air Springs in Changing Temperature Environment

As an important part of the train’s secondary suspension system, air springs can support the body weight and absorb the vibration from track, ensuring ride comfort and running safety of trains. The material parameters of rubber and cord change with actual operating temperature, which will change the dynamic performance of air springs, and further affect the stability of train operation. In this paper, an air spring surrogate model with temperature-varying characteristics is proposed for both forward analysis and parameter inversion. Firstly, a dynamic finite element model of an air spring considering different temperatures is established based on ABAQUS. Secondly, using optimization algorithms and the radial basis function (RBF) neural network, a finite element surrogate model is constructed between the material parameters, the airbag’s initial pressure, the external excitation and the air spring’s characteristic parameters. Finally, based on the constructed surrogate model, the dynamic characteristics of the air spring at different temperatures are forwardly analyzed, and the rubber hardness of the airbag and emergency spring and the elastic modulus of the cord material are further inverted at specific temperatures. This work can provide a basis for trains dynamics analysis under changing temperature environment.