Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ASME B31G-2012 (2017) — Manual for Determining the Remaining Strength of Corroded Pipelines
🛢️ ASME B31G-2012 (2017) occupies a distinctive position in the ASME B31 Code for Pressure Piping. It is not a design code but a manual, and it answers one very practical question: given a steel pipeline that has already suffered corrosion metal loss, can it continue to operate safely at the intended pressure, or must it be derated, repaired, or replaced? First published in 1984, the document has become one of the most widely cited technical references in pipeline integrity management worldwide, and it is adopted by reference in ASME B31.4, B31.8, B31.11, and B31.12.
The 2012 edition refined the structure introduced in 2009 into an explicit three-level assessment framework: Level 1 is a rapid screening using tabulated allowable corrosion lengths, Level 2 is the analytical modified B31G (0.85dL) calculation, and Level 3 covers detailed evaluation including the RSTRENG effective area method, numerical analysis, and full-scale test validation. This gradation lets an operator allocate assessment effort in proportion to defect severity — neither missing risk nor spending heavily on repairs that are not required. The edition was formally reaffirmed by ASME in 2017 without technical change.
💡 The engineering philosophy of B31G is that a certain amount of metal loss can be tolerated. Corrosion thinning is not synonymous with pipeline failure; the real question is whether the remaining wall can carry the hoop pressure load at flow-stress level. The typical output is a maximum allowable operating pressure, not a simple pass/fail verdict.
1. Scope and Acceptable Applications 📌
Per Section 1, B31G applies to pipelines and piping systems within the scope of the transportation pipeline codes that form part of the ASME B31 Code for Pressure Piping, namely: ASME B31.4 (Pipeline Transportation Systems for Liquid Hydrocarbons and Other Liquids), ASME B31.8 (Gas Transmission and Distribution Piping Systems), ASME B31.11 (Slurry Transportation Piping Systems), and ASME B31.12, Part PL (Hydrogen Piping and Pipelines). Where the term “pipeline” is used, it may also be read to apply to piping or pipe conforming to the acceptable applications.
Paragraph 1.2 lists the acceptable applications, which must be checked one by one before any assessment begins:
- Location: metal loss in pipelines belowground, aboveground, or offshore.
- Cause: metal loss due to external or internal corrosion.
- Grinding: metal loss produced by grinding used to completely remove mechanical damage, cracks, arc burns, manufacturing defects, or other surface defects.
- Components: metal loss in field bends, induction bends, and elbows.
- Adjacent welds: metal loss incidentally affecting longitudinal, helical, or circumferential electric welds of sound quality and ductile character, provided workmanship flaws are not close enough to interact.
- Depth: metal loss of any depth, except that measurement accuracy and effective corrosion rate deserve particular consideration where depth exceeds 80% of actual wall thickness.
- Stress and loading: pipe operating at any allowable design hoop stress level, and at temperatures above ambient within the range recognized by the governing standard, with internal pressure as the primary loading.
Paragraph 1.3 states the exclusions with equal clarity: crack-like defects or mechanical surface damage not completely removed to a smooth contour by grinding; metal loss in indentations or buckles producing radial distortion greater than 6% of pipe outside diameter (unless a Level 3 assessment per para. 2.4 is performed); grooving, selective, or preferential corrosion; and cases outside the ductile fracture initiation characteristics of the material (again, unless a Level 3 assessment is used).
⚠️ The most common misuse is treating B31G as a universal defect assessment tool. B31G evaluates volumetric metal loss only — not cracks, not geometric distortion (dents, buckles, ovalization), and not pipe subject to significant axial compression, bending, or cyclic fatigue. Crack-like defects belong to fracture-mechanics methods such as API 579-1/ASME FFS-1.
2. The Three-Level Assessment Structure 🔧
The technical core of B31G is Section 2, Evaluation Methods. The three levels are not competing options but a progression of decreasing conservatism and increasing data demand. If one level shows the defect is unacceptable, a more detailed level may be used to re-evaluate it.
| Level | Basis | Data Required | Conservatism | Typical Use |
|---|---|---|---|---|
| Level 1 | Section 3 tables of allowable corrosion length (L values) | Pipe size, wall thickness, maximum pit depth, corrosion length, hoop stress ratio | Most conservative; often under-predicts capacity by 20%–50% | Routine screening, bulk classification of ILI defect lists |
| Level 2 | Modified B31G (0.85dL) analytical equation | As above, plus a reliable axial projected length and flow stress | Moderately conservative; valid for d/t ≤ 0.80 | The workhorse method for most corrosion defects |
| Level 3 | RSTRENG effective area, finite element analysis, full-scale test validation | High-resolution profile data (river-bottom profile), material properties, measured geometry | Least conservative; closest to true failure pressure | Deep defects, complex or interacting defects, disputed cases |
2.1 The Level 2 equation
The modified B31G (0.85dL) method idealizes the corroded region as a rectangular defect of length L and depth d, applies an 0.85 depth-area reduction, and uses SMYS + 68.95 MPa (10 ksi) as flow stress. The predicted failure pressure is:
Pf = 2·Sflow·t / D × [ (1 − 0.85d/t) / (1 − 0.85d/(t·M)) ]
where the Folias bulging factor M is computed in two branches: for L²/(D·t) ≤ 50, M = [1 + 0.6275·L²/(D·t) − 0.003375·(L²/(D·t))²]1/2; for L²/(D·t) > 50, M = 0.032·L²/(D·t) + 3.3. Here Pf is the predicted failure pressure, Sflow the flow stress, t the measured wall thickness (or nominal minus mill tolerance), D the outside diameter, d the maximum corrosion depth, and L the axial length of the corroded region.
Compared with the original 1984 B31G method, two changes matter. The original used 1.1 × SMYS as flow stress, approximated the corroded area as a parabola-plus-rectangle combination (equivalent to 2/3·d), and used M = [1 + 0.8·L²/(D·t)]1/2. The modified method raises the flow stress, uses a 0.85d rectangular area, and adopts the more accurate M fit. Field experience shows the modified approach predicts roughly 10%–20% higher capacity for moderate-length defects, avoiding many unnecessary repairs.
✅ Acceptance logic: compare the predicted failure pressure Pf with the target pressure established by the governing code (normally MAOP or design pressure), applying the safety factor that code requires. If Pf cannot support current operation, the options are pressure reduction, repair, pipe replacement, or a shortened re-inspection interval.
2.2 Interaction between adjacent corrosion pits
Real corrosion is rarely a single isolated pit. Paragraph 1.12 and Figure 1.12-1 define pit interaction: when two areas of metal loss are separated, axially or circumferentially, by less than the specified threshold, the sound metal between them cannot act as an effective load bridge, and the areas must be treated as a single composite defect with the combined length and maximum depth. This rule has an outsized effect on results — treating a dense colony of interacting pits as independent single pits is the most common source of overestimated remaining strength in ILI data assessment.
Paragraph 1.13 and Figure 1.13-1 address helically oriented corrosion. Because helical corrosion follows the plate rolling direction, its axial projection can be much shorter than its actual path length; taking only the axial length underestimates severity, so the manual requires converting the helix geometry into an equivalent length.
3. Tables of Allowable Length of Corrosion (Section 3) 📏
Section 3 provides a complete set of directly usable tables covering NPS 2 through NPS 60, each with a metric counterpart (for example, Table 3-1 and Table 3-1M cover NPS 2–NPS 6 and OD 60–168 mm respectively). Each table gives the maximum allowable corrosion length L for a given pipe size, wall thickness, depth ratio d/t, and hoop stress ratio. If the measured axial length does not exceed the tabulated value, the defect is acceptable without further calculation.
| Table (U.S. Customary / Metric) | Size Coverage | Inputs | Output |
|---|---|---|---|
| 3-1 / 3-1M | NPS 2 to < NPS 6 (OD 60–168 mm) | Wall thickness, d/t, hoop stress ratio | Allowable length L |
| 3-2 / 3-2M | NPS 6 to < NPS 10 (OD 168–273 mm) | As above | Allowable length L |
| 3-3 / 3-3M | NPS 10 to < NPS 16 (OD 273–406 mm) | As above | Allowable length L |
| 3-4 / 3-4M | NPS 16 to < NPS 20 (OD 406–508 mm) | As above | Allowable length L |
| 3-5 / 3-5M | NPS 20 to < NPS 24 (OD 508–610 mm) | As above | Allowable length L |
| 3-6 to 3-12 / 3-6M to 3-12M | NPS 24 to NPS 60 (OD 610–1 524 mm) | As above | Allowable length L |
💡 The critical lookup variable is the stress ratio. The same defect has a very different allowable length at different hoop stress levels: the higher the ratio (for example 0.72 SMYS near class locations or compressor stations), the shorter the allowable length. Always establish the actual stress ratio from the governing code — B31.8 class location and design factor — before entering the tables.
4. Level 3 Assessment and the RSTRENG Effective Area Method 🔬
Level 3 is required whenever a defect falls outside the validity of Levels 1 and 2 — d/t greater than 0.80, highly irregular profiles, multiple interacting pits, locations at bends or tees, or the presence of axial loads. Level 3 is not a table lookup or a single closed-form equation; it permits any adequately validated method. Common routes include:
- Effective area method (RSTRENG): the corroded region is discretized from high-resolution profile data into axial increments of metal loss, and the remaining load-bearing area is integrated along the true river-bottom profile. Experience shows predicted capacity can be another 10%–20% higher than the 0.85dL method.
- Nonlinear finite element analysis: used for complex geometry — elbows, tees, reducers, cased crossings — and for combined loading, solving directly for elastic-plastic limit capacity.
- Full-scale test validation: burst testing of pipe segments with the same size, grade, and defect morphology, calibrating the model against measured failure pressure.
The manual stresses that, at every level, input data quality governs the credibility of the conclusion. Depth accuracy of high-resolution magnetic flux leakage (MFL) in-line inspection data, the judgement of where a corrosion length starts and ends, and the determination of an effective (not average) corrosion rate are usually the dominant uncertainty contributors. Paragraph 1.4 specifically requires that measurement accuracy and effective corrosion rate be considered when metal loss exceeds 80% of wall thickness.
🚨 Never extrapolate remaining life from an average corrosion rate in place of a strength assessment. Corrosion is typically locally accelerated, and average rates badly underestimate the growth of the deepest pit. The correct practice is to pair the same defect across successive inspections, derive its measured growth rate, take a conservative upper bound, and then set the re-inspection interval.
5. Engineering Practice Essentials 🛠️
- 🔍 Close the data loop: in-line inspection → defect list and clustering → interaction screening → graded assessment → repair decision → re-inspection interval. Document every step; a B31G result is only meaningful inside a complete integrity management process.
- 📊 Wall thickness input: use measured wall thickness. Where none exists, take nominal thickness minus the permitted mill under-tolerance (for example the −12.5% allowed by API 5L). Never use nominal thickness unadjusted.
- 🔧 Repair selection: grinding (shallow defects with adequate remaining wall), Type B sleeves, composite wraps, clamp-type repair sleeves, or pipe replacement — chosen from the assessed derating requirement and whether the defect is still growing.
- ⚖️ Derating: where the assessment supports operation only at reduced pressure, implement the temporary pressure reduction per the governing code while accelerating inspection and planning permanent repair.
- 📋 Documentation: the assessment report should record defect location, dimensions, measurement method, assessment level and equations, parameter sources, conclusions, and margin — for regulatory review and future comparison.
6. Frequently Asked Questions ❓
❓ Q: How does B31G relate to API 579-1/ASME FFS-1?
A: Both assess remaining strength of damaged equipment, but they serve different purposes. B31G focuses on volumetric corrosion metal loss in pipelines, is simpler, and dovetails with the B31.4/B31.8 framework. API 579-1/ASME FFS-1 is broader, covering cracks, creep, fire damage, pitting, and other damage modes. The industry habit is to run the B31G graded assessment first and move to the relevant API 579 part when B31G’s validity limits are exceeded.
A: Both assess remaining strength of damaged equipment, but they serve different purposes. B31G focuses on volumetric corrosion metal loss in pipelines, is simpler, and dovetails with the B31.4/B31.8 framework. API 579-1/ASME FFS-1 is broader, covering cracks, creep, fire damage, pitting, and other damage modes. The industry habit is to run the B31G graded assessment first and move to the relevant API 579 part when B31G’s validity limits are exceeded.
❓ Q: Original B31G, modified B31G (0.85dL), or RSTRENG — which should I use?
A: In descending order of conservatism: original B31G > 0.85dL > RSTRENG effective area. The original suits early, low-quality data; 0.85dL is today’s industry baseline for defects with d/t ≤ 0.80; RSTRENG is more economical where profile resolution is high and unnecessary repairs must be minimized. All three must remain inside B31G’s stated applicability.
A: In descending order of conservatism: original B31G > 0.85dL > RSTRENG effective area. The original suits early, low-quality data; 0.85dL is today’s industry baseline for defects with d/t ≤ 0.80; RSTRENG is more economical where profile resolution is high and unnecessary repairs must be minimized. All three must remain inside B31G’s stated applicability.
❓ Q: How far apart must two pits be before they can be assessed separately?
A: Figure 1.12-1 gives the axial and circumferential interaction criteria. The engineering interpretation is that when the sound metal between two pits is too short to bridge the load — on the order of a few wall thicknesses — they must be combined. Because the criterion depends on diameter, wall thickness, and depth, check each case against Figure 1.12-1 rather than applying a single fixed number.
A: Figure 1.12-1 gives the axial and circumferential interaction criteria. The engineering interpretation is that when the sound metal between two pits is too short to bridge the load — on the order of a few wall thicknesses — they must be combined. Because the criterion depends on diameter, wall thickness, and depth, check each case against Figure 1.12-1 rather than applying a single fixed number.
❓ Q: Once a defect passes assessment, can inspection be discontinued?
A: No. B31G assesses remaining strength at a point in time, while corrosion continues. Every assessment must be paired with a remaining-life calculation and a re-inspection interval: extrapolate the measured growth rate of that defect to the time when failure pressure falls to operating pressure, then shorten the interval by the safety margin the code requires.
A: No. B31G assesses remaining strength at a point in time, while corrosion continues. Every assessment must be paired with a remaining-life calculation and a re-inspection interval: extrapolate the measured growth rate of that defect to the time when failure pressure falls to operating pressure, then shorten the interval by the safety margin the code requires.
