Reference
Glossary
Precise definitions of the technical terms, standards and methods used across Wood Intelligence.
- Service life
- The period during which a timber element performs its required function without unforeseen maintenance. In Wood Intelligence it is modelled as a function of climate load, species resistance, orientation, exposure and surface treatment.
- Durability class (EN 350)
- A five-class scale from DC 1 (very durable) to DC 5 (not durable) based on laboratory testing against wood-destroying organisms. It anchors the reference service life in the model.
- Use class (EN 335)
- The exposure scenario that defines the biological hazard. Above-ground exterior cladding is typically use class 3, while ground contact is use class 4 or 5.
- Köppen climate classification
- A global climate zoning system used to group locations by temperature and precipitation regime. Wood Intelligence uses the main class (A–E) to scale climate uncertainty.
- P10 / P50 / P90
- Percentiles from the Monte Carlo distribution. P50 is the median estimate; P10 and P90 bound the lower and upper plausible range. They are empirical, not a fitted normal curve.
- Monte Carlo propagation
- A method that repeatedly evaluates the model with inputs perturbed within their declared uncertainty. It produces the reported percentiles and shows which inputs drive the spread.
- Confidence band
- A qualitative label (High / Moderate / Indicative) that reflects how tightly the inputs constrain the result. It is not a statistical confidence interval and should not replace project-specific testing.
- Calibration factor (WI-CAL)
- A published, zone-specific multiplier that corrects the uncalibrated model against the WI-VAL empirical benchmark. It keeps the physics and the correction auditable as separate layers.
- Silvering
- The visible greying of exposed wood caused by UV photodegradation of surface lignin. The time to uniform silvering is reported in months and depends mainly on UV load.
- Trace / calculation record
- A complete, versioned export of one evaluation: inputs, every intermediate value, the method specification and the outputs. It allows independent recomputation without access to the codebase.