Full walkthroughs of how HeaterIntegrity's calculation engine was checked against API's own published worked examples, item by item — inputs, method, and results.
API 579-2 / ASME PTB-14-2023 is the companion volume to API 579-1 / ASME FFS-1, the fitness-for-service standard HeaterIntegrity implements — it works through fully published, independently checkable worked examples of that standard's own methods. We use it as our reference for validating HeaterIntegrity's calculation engine, example by example, below.
A liquid knock-out vessel, normally operating below the creep range, experienced a temperature excursion during a relief event.
Per the API worked example (Annex 2C), the operating stress at this condition is computed directly:
| Component | Operating stress |
|---|---|
| Cylindrical shell | 9,091 psi (9.091 ksi) |
| Elliptical head | 9,049 psi (9.049 ksi) |
Reading API 579-1 Figure 10.4 (Level 1 screening curve, Carbon Steel 60 ksi) at 975°F, the API example states the acceptable creep life is over 1,000 hours — 2,231.8 hours (shell) and 2,293.3 hours (head), using logarithmic interpolation. Since the actual exposure was only 20 hours, the example concludes both components are fit for service.
The check ran directly against HeaterIntegrity's own design-based check — the same logic the app's Preliminary Assessment page uses — not a re-derivation from memory or a spreadsheet copy.
This inverts the direction of the API's own Step 5, which reads time → finds stress, to solve stress → time instead, so the comparison is apples-to-apples: both numbers answer the same question — how long can this component run at this stress and temperature before hitting the screening curve?
HeaterIntegrity's design-based check reproduces the API's own worked answer to within about 1.2%, for both components.
| Component | Applied stress | API's published max time (log-interpolated) |
HeaterIntegrity's computed max time | Difference |
|---|---|---|---|---|
| Cylindrical shell | 9.091 ksi | 2,231.8 h | 2,204.4 h | −1.23% |
| Elliptical head | 9.049 ksi | 2,293.3 h | 2,267.5 h | −1.13% |
Both figures are far above the 20-hour actual exposure, so both components are well within their permissible time — consistent with the API example's own conclusion.
The design-based check — screening-curve allowable stresses plus log-linear time interpolation — against API 579-2's own published worked example, independently, to within about 1.2%.
Level 2 methods are checked separately — see Example Problem 3, below.
A fired crude heater tube experienced a temperature excursion for a short duration before a scheduled turnaround. The check: how much creep damage occurred, and are the tubes fit for another run?
For each condition, the standard computes the corroded wall thickness, the resulting hoop stress, and the permissible time at temperature from the Level 1 screening curve — then checks each condition's actual service time against that permissible time.
Run the same way as Example Problem 1 — directly against HeaterIntegrity's real code, not a hand re-derivation:
HeaterIntegrity's design-based check reproduces the standard's own worked answer to within about 1–2%, across all three conditions:
| Condition | Temp | Hoop stress | Standard's permissible time | HeaterIntegrity | Difference |
|---|---|---|---|---|---|
| Past | 1130°F | 3,543 psi | 256,739 h | 252,847 h | −1.5% |
| Excursion | 1200°F | 3,546 psi | 20,450 h | 20,673 h | +1.1% |
| Future | 1130°F | 3,974 psi | 141,920 h | 138,926 h | −2.1% |
Every condition's actual service time is well within its permissible time — consistent with the standard's own conclusion for the Past and Excursion conditions.
Corrosion-projected wall thickness, hoop stress, and the design-based check — this time for an actual heater tube (not a pressure vessel, as in Example Problem 1) and a Cr-Mo alloy material, across three different operating conditions rather than one.
The Level 2 engine is checked next — see Example Problem 3, below.
Same tube as Example Problem 2, continued: the standard's own text picks up from there ("the tubes have already failed a Level 1 Assessment") and re-runs the full load history as a Level 2 assessment — the app's real MPC Omega and Larson-Miller creep-life engine, not just the design-based check.
Run directly against the app's real creep-life engine — the same functions behind every Remaining Life figure the app produces:
Before running any numbers, the material-property constants behind both methods — in the app's own data files — were checked directly against the standard's own published constants for this material, and matched exactly.
The app's Larson-Miller engine reproduces the standard's own worked answer to within 0.01% — essentially exact. Its Omega engine matches to within about 1%, consistent with the ~1–2% seen throughout this validation:
| Method | HeaterIntegrity — total damage | Standard's total damage | Difference |
|---|---|---|---|
| MPC Omega | 0.436 | 0.440 | −0.8% |
| Larson-Miller, minimum properties | 0.981 | 0.980 | +0.1% |
| Larson-Miller, average properties | 0.378 | 0.377 | +0.1% |
Both methods agree the component is well below the allowable damage of 1.0 — matching the standard's own conclusion that the tubes are fit for the additional planned run.
HeaterIntegrity's real Level 2 creep-life engine — both the MPC Omega and Larson-Miller methods, including the material-property data behind each — against a full multi-condition, multi-increment worked example for a real Cr-Mo heater tube. This closes the loop left open after Example Problem 2, which checked the design-based path only.
Next: a Level 3 crack-like-flaw problem — checked wherever it exercises the app's own engine — see Example Problem 4, below.
A longitudinal crack-like flaw in an SA-240 Type 316/316H cylindrical vessel. This example is a full Level 3 fracture-mechanics assessment — a Failure Assessment Diagram (FAD), stress-intensity factors, crack growth integration — none of which HeaterIntegrity performs, since it doesn't do crack-like-flaw assessments. What's checked here is everything in the example that is a direct input into the app's engine: stress components, the plastic-collapse check, and the ordinary Omega-method creep damage on the component before any crack is introduced.
Run directly against the app's real stress and creep-life functions, same as every example before it:
The standard applies two adjustment factors to this step simultaneously: Δsr = −0.5 (bottom of the creep-strain scatter band) and Δcd = −0.3 (a creep-ductility adjustment for austenitic, ductile-behavior materials). HeaterIntegrity's "User Defined" Creep Properties option (MPC Omega method only) takes both directly, per API 579-1/FFS-1 Annex 10B.1 — the results below use that option, set to the standard's own values.
Stress components match the standard to within 0.003% — essentially exact. The plastic-collapse check reaches the same conclusion as the standard. With Δsr and Δcd set to the standard's own values, the Omega engine reproduces the standard's own published creep life and damage to within 0.004%:
| Method | HeaterIntegrity | Standard's Result | Difference |
|---|---|---|---|
| MPC Omega — creep life (Δsr=−0.5, Δcd=−0.3) | 2,615,108 h | 2,615,000 h | +0.004% |
| MPC Omega — total damage | 0.01696 | 0.01696 | ~0% |
Stress components on a third geometry/loading case; the plastic-collapse check, on its own methodology; and, using the app's "User Defined" Creep Properties option (MPC Omega method only, any Δsr −0.5 to +0.5 / Δcd −0.3 to +0.3, per API 579-1/FFS-1 Annex 10B.1) — the Omega engine reproducing the standard's own published creep life and damage to within 0.004%.