When static reserve and ductility tell different stories.

The engineering question

Would a strength-focused result tell the whole story? In this example, static reserve remains above its screening threshold while elongation reserve falls below a separate provisional sensitivity threshold.

The stored result at 437 days

SYN-03 represents CTGAN_0001, a synthetic input evaluated by frozen paper models. The two results are kept separate; Copilot cannot average away or clear the ductility flag.

From regional map to evidence

Region B1 is a fictional map label. Selecting it in the dashboard opens the stored result and provenance. The heat layer interpolates eight synthetic endpoint severities; it does not infer unobserved site conditions or estimate failure probability.

What the curves and shading show

Stored queries cover 90 to 437 days. The shaded range comes from 1,000 chloride-input perturbations per case, summarised as P2.5-P97.5 and smoothed for display. It describes input response, not a calibrated prediction interval. Tree-model responses can be stepped and asymmetric.

What an engineer would need next

Dated water records, verified support identity and inspection history, a reviewed geological/RRI assignment and applicable company criteria are needed for a real assessment. This case supplies none of those histories. It does not prescribe a replacement date or prove maintenance savings.

All supports within the displayed region share one installation batch as a presentation assumption. The example is not a measured population of installed supports.

Explore the method and discuss a pilot

Read the method and governance design, explore technical questions or discuss the evidence available in your operation.