Improved Disturbance Observer adhesion control with surface recognition module for Heavy Trains

Abstract Adhesion control between wheel and rail is a critical factor in the performance, safety and energy efficiency of electric rail traction systems. This paper proposes an adaptive fuzzy based approach for optimal adhesion control in electric heavy duty rail traction. A feedforward fuzzy logic network is trained using simulated adhesion data derived from a dynamic traction model incorporating wheel- rail creep forces and environmental parameters. The proposed controller is integrated into a vector controlled traction drive model in MATLAB/Simulink. Simulation results demonstrates improved adhesion utilization. A six-axle locomotive dynamical model and empirical adhesion characteristic are used to evaluate the methods in single- and multi-axle scenarios. Simulation results demonstrate that the proposed hybrid method substantially improves adhesion utilization, while maintaining robustness across dry, wet, and low-adhesion surfaces. The focus of this paper is on analyzing the adhesion coefficient. Keywords Heavy-duty locomotive, adhesion control, disturbance observer (DOB), fuzzy logic, support vector machine (SVM), adhesion coefficient.