Uncertainty Quantification
Formally characterizing and propagating uncertainty through technical and economic analyses.
Definition
Uncertainty quantification is the discipline of representing what is not precisely known — a failure rate, a repair duration, a commodity price, a production rate — as a distribution rather than a single number, and then propagating that distribution through an analysis so the resulting output carries an honest range instead of a false point estimate. The goal is not to eliminate uncertainty; it is to bound it, quantify it, and make its consequences visible to the decision being made.
Why a Single Number Understates Risk
A point estimate — "mean time to failure is 400 hours," "the project costs $12M" — presents a single value as if it were certain. In reality, each input used to build that estimate carries its own variability, and when those variabilities are not carried forward, the final answer looks more precise than the evidence supports.
Uncertainty quantification replaces that single number with a distribution — a range with an associated confidence level — and propagates it mathematically through the model, so the final output (a cost, an availability figure, a production forecast) reports a credible interval instead of a misleadingly exact figure.
Where It Applies in Industrial Decisions
Uncertainty quantification underlies probabilistic cost estimates (P50/P80/P90 ranges), reliability and availability forecasts built from variable failure and repair data, and production forecasts that depend on variable equipment performance. It is the mathematical foundation that makes Monte Carlo simulation meaningful — the simulation samples from the quantified uncertainty to produce its output distribution.
How This Connects to Knar Global's Work
Knar Global treats uncertainty as a quantity to be bounded and carried through every analysis, not smoothed away for the sake of a clean-looking number — reflected in probabilistic ranges rather than single-point projections across RAM models, cost estimates, and production forecasts.