Engineering & Developmenttest reductionUK
Cutting Physical Test Cycles with Engineering AI
5 June 2026
A powertrain team trained surrogate models on historical test data to predict outcomes, removing a meaningful share of physical rig tests from the validation plan.
Download the full case study
Branded PDF, watermarked to your account. Registration required.
Physical rig testing is slow and expensive, and much of it re-confirms behaviour the team has effectively seen before.
By training surrogate models on years of historical test data, engineers could predict outcomes for many new configurations with quantified confidence.
Low-uncertainty cases were signed off on the model's prediction; only genuinely novel or high-risk cases went to the rig, trimming the test plan without lowering confidence.
Vendors referenced
Sources