To address the issue of limited interpretability in current deep learning-based bearing fault diagnosis methods, this study proposes a convolutional neural network model incorporating spectral physical constraints. The model integrates data-driven learning with physical constraints by embedding the spectral features of bearing faults as prior knowledge into the network architecture during training. This guides the model to adaptively learn spectral features closely associated with fault mechanisms. Experimental results on both public datasets and self-collected data show that the proposed model not only maintains high diagnostic accuracy but also provides intuitive and credible justification for fault classification through the visualization of spectral responses at the network output layer. This enhances the spectral interpretability of the model’s decisions, achieving a transition from a “black box” to a “white box” and effectively improving the reliability of deep diagnostic models.

Eight-track vehicles feature large grounding areas, low ground contact pressure, and high load-bearing capacity, and are widely adopted for heavy-duty engineering transportation and large special-purpose operation platforms. Accurate characterization of their steady-state steering forces is essential for tracked-chassis parameter matching and structural optimization. Nevertheless, existing steering dynamic models mainly focus on double-track, four-track, and six-track configurations, ignoring slip distribution differences induced by the parallel dual-track and multi-segment grounding structure of eight-track vehicles. To fill this gap, this paper establishes a steady-state steering dynamic model for large-scale eight-track vehicles. Lateral force, traction force, and steering resistance moment of each track group are derived via grounding micro-element integration, and vehicle static equilibrium equations are constructed. The damped Newton iteration algorithm is employed to solve the transcendental equations numerically, followed by key-parameter sensitivity analysis. A RecurDyn-based virtual prototype is further developed for multi-body dynamic simulation verification. Quantitative comparison shows that the maximum relative error of track driving force is within the engineering acceptable range, validating the correctness and engineering applicability of the presented model. Track gauge and grounding length are identified as dominant factors for steering driving force, whereas track width and the spacing between homolateral parallel tracks exert limited influence. This model can provide theoretical support for multi-track vehicle chassis optimization.