The vertical stiffness of railway tracks is crucial for ensuring safe and efficient rail transport. Rail-pad dynamic stiffness is a key component influencing track performance. Determining the dynamic stiffness of rail pads poses a challenge because it depends not only on the material and geometry of the rail pad but also on the testing conditions, due to the non-linear material response. To address this issue, a methodology is proposed in this paper to estimate dynamic stiffness using static stiffness measurements. This approach enables the prediction of dynamic stiffness for different situations from a single laboratory test. This study further examines whether this correlation remains valid for different types of rail pads, even when their mechanical behavior has been degraded by temperature, wear, or chemical agents. Experiments were conducted under varying temperatures and on rail pads that underwent mechanical and chemical degradation. The analysis assesses the validity of the static-to-dynamic stiffness correlation under degraded conditions and investigates the influence of each testing condition on the ability to estimate dynamic stiffness from static stiffness and operational parameters. The findings provide insights into the reliability of this predictive model and highlight the impact of degradation mechanisms on the dynamic behavior of rail pads. This research enhances the understanding of rail pad performance and offers a practical approach for evaluating dynamic stiffness. By considering all of the variables used in the analysis, the approach achieves R² values of up to 0.99, which carries significant implications for track design and maintenance.

Full-factorial characterization of rail-pad dynamic stiffness is costly because each operating combination of toe load, load amplitude, and frequency must be tested under every experimental condition. This study evaluates whether data-driven reconstruction can support reduced testing campaigns by deliberately withholding operating combinations and benchmarking a grid-aware conditional generative adversarial network (cGAN) against Gaussian process (GP) kriging and other reconstruction methods. The hidden-cell protocol was applied to 360 measurements from an EPDM pad under ten experimental conditions and independently replicated on two additional pads. At the primary 25% masking level, the cGAN and pooled GP achieved mean repetition-level median absolute percentage errors of 4.8% and 1.0%, respectively, although unrestricted random masking left only 39 of 90 withheld observations interpolative on average. In S1-R, mask selection was conditioned a priori on preserving nominal geometric support, using only the nominal design coordinates and retaining all eight vertices. All 90 withheld observations were then evaluated at their achieved coordinates; the median across repetition-level median errors was 4.38% for the cGAN and 1.23% for pooled GP, with 80.89% and 99.78% of predictions within 10% error, respectively. The GP also provided better-calibrated uncertainty (95% coverage: 96.2% versus 84.8%), and its superior accuracy was reproduced on both additional pads. Overall, the results support the retrospective feasibility of a 25% reduction in tested operating combinations under a nominal support-preserving design, with pooled GP kriging providing the most accurate and robust reconstruction.