Inspection Intervals, Detectability & Probability of Detection
How NDT capability, flaw-growth rate, sizing uncertainty and critical defect size are combined to set defensible inspection intervals rather than arbitrary calendar periods.
Inspection Interval Is an Engineering Calculation
An inspection interval should be related to how quickly a relevant defect can progress from a reliably detectable condition to an unacceptable condition. Calendar convenience alone is not a technical basis. The assessment therefore couples inspection capability with degradation mechanics, future duty and the consequence of missed or mis-sized defects.
Detection Threshold vs Probability of Detection
The smallest indication ever seen in a trial is not the same as a reliably detectable flaw. Probability of detection describes the likelihood that a flaw of a given size will be found under defined inspection conditions. Real capability depends on technique, access, surface condition, operator, calibration, orientation and geometry. Integrity calculations should use a detection basis supported by the actual inspection system.
Sizing Error Is Different from Detection
An inspection can detect a flaw but still estimate its dimensions inaccurately. Depth sizing is often particularly important for fracture and remaining-ligament assessment. Detection probability and sizing uncertainty should therefore be treated separately. A crack-growth calculation starting from the nominal reported size may be non-conservative if the method has a known positive or negative sizing bias.
The Detectable-to-Critical Growth Window
The central damage-tolerance quantity is the time or number of cycles required for a flaw to grow from the inspection basis to the terminal flaw size. The scheduled interval must then include an appropriate margin within that window, accounting for uncertainty and the inspection philosophy. A highly capable inspection technique can increase the available growth window by reliably detecting smaller flaws.
Missed Detection and Repeated Inspections
Repeated inspections can reduce the probability that a continuously growing defect remains undetected, but only if the inspections are sufficiently independent and the degradation remains within the assumed growth model. Simply multiplying probabilities can be misleading when the same access limitation, geometry or systematic blind spot affects every inspection.
Coverage and Inspectability
An inspection plan is only as good as its coverage of the locations where damage can occur. Weld caps, attachments, supports, insulation, internal surfaces and complex geometry can restrict access or create dead zones. The structural analysis should identify the critical locations and orientations, while the NDT plan should demonstrate that those locations are actually inspectable.
Uncertain Growth Rate
Material scatter, environment and load uncertainty can produce a wide range of crack-growth rates. The interval should be based on a conservative or probabilistically justified growth model rather than the mean rate alone. Where operating data are available, the authorised interval can be tied to accumulated cycles or transients instead of elapsed time.
Condition Monitoring and Online Data
Monitoring pressure, temperature, strain, vibration or acoustic response can provide evidence about load exposure or degradation rate, but it rarely removes the need for direct inspection unless the monitoring system is specifically validated for defect detection. Its strongest role is often to confirm that the assumed future duty and load spectrum remain valid between inspections.
Inspection Findings Feed Back into the Model
Each inspection should be treated as new evidence. No growth, measured growth, new indications or improved sizing can all justify reassessment. The model should be updated using controlled configuration and service-history records rather than treating the original life calculation as fixed for the remainder of asset life.
Setting a Defensible Interval
- Critical defect size is mechanics-based — Defined by fracture, collapse, leakage or another explicit limit state.
- Inspection capability is demonstrated — POD and sizing performance correspond to the actual technique and geometry.
- Growth law matches the mechanism — Fatigue, corrosion fatigue, SCC or creep mechanisms are represented as applicable.
- Future duty is bounded — Cycle count, transient severity and environment are linked to the authorised operating envelope.
- Reassessment trigger is defined — Unexpected inspection findings or duty changes prompt review rather than waiting for the next planned interval.
Reporting the Basis
The integrity record should state the detectable flaw basis, assumed sizing uncertainty, critical flaw size, growth model, future duty, calculated growth window, chosen inspection interval and margin. This creates a traceable chain from NDT capability to continued-service decision.
Inspection intervals should be derived from detectable-to-critical growth, not selected first and justified afterwards.
Engineering judgement — governing sensitivities
For Inspection Intervals, Detectability & Probability of Detection, the most useful review question is not simply whether the solver has produced a plausible contour or scalar result, but whether the model preserves linking inspection performance to the flaw population and failure mechanism being managed. A nominal POD curve is not portable between techniques, access conditions, flaw orientations or operator/procedure states without justification. This is where apparently small modelling choices can change the engineering conclusion. The analyst should identify the variables that can move the governing response, separate physical uncertainty from deliberate conservatism, and show that the selected modelling fidelity is proportionate to the decision being supported. Where the response is close to an acceptance boundary, sensitivity cases should bracket credible changes rather than apply arbitrary percentage perturbations.
Verification evidence for the engineering record
A defensible Inspection Intervals, Detectability & Probability of Detection assessment should leave an evidence trail that another engineer can independently interrogate. At minimum, review POD evidence, sizing error, false-call treatment, coverage, interval sensitivity and the probability of a flaw growing from the detectable range to the critical range before the next inspection. Numerical convergence should be demonstrated on the response quantity that drives the decision, not only on generic mesh or solver metrics. The report should distinguish verified numerical behaviour from validation against test or service evidence, record any extrapolation beyond the supporting data, and state which assumption would most likely change the conclusion. This turns the analysis from a plausible calculation into an auditable engineering substantiation.