How do you validate a wheel-rail contact model without full-scale tests?

I work on wheel-rail contact and wear modelling, and full-scale rig tests are usually out of reach for my group.

I rely on scaled rig data and published field measurements, but reviewers often push back on how well these transfer to the model.

How do you justify a contact or wear model when full-scale validation isn't possible, and what evidence do you find convincing?