Transient dynamic characteristics of the rolling-lobe air spring based on physics-data hybrid driven approach

Abstract

The dynamic characteristics of air springs are crucial in their production, manufacturing, and intelligent control. However, most existing research and methods focus primarily on studying properties such as hysteresis and dynamic stiffness under steady-state conditions, with limited attention to their transient responses under actual working conditions. This paper introduces a transient characteristic mapping technology that comprehensively considers the dynamic stiffness mechanism model generated by the air spring’s internal gas and rubber diaphragm. The study establishes a nonlinear relationship between the transmission force of the air spring and the measured transient physical quantities. Based on this physical model, a physics-data hybrid-driven framework is developed, which considers the loading sequence and the discrete nature of multi-dimensional physical quantities in the loss function. Experimental results validate the accuracy of the proposed theory, with the error in transient dynamic characteristics controlled within 3.04% for internal pressure and 1.92% for transmission force over the full test set. This provides valuable theoretical support for developing air suspensions under real-world working conditions.