E911/X11CrMoWVNb9-1-1(T911/P911)seamless steel pipe

I. General Principles and Basic Description of Documents
1.1 Purpose and Scope of the Document
The purpose here is simple: to serve as a practical, from-the-trenches guide for the specification, production, and application of E911 seamless steel pipe. This isn’t a textbook for students; it’s a field manual for engineers, inspectors, and procurement specialists who are tired of boilerplate language.
It applies specifically to seamless pipes and tubes manufactured from the martensitic steel grade E911 (X11CrMoWVNb9-1-1), primarily used in the power generation and petrochemical industries. We are talking about the critical components—the main steam lines, the reheaters, the superheaters—where the pressure is high and the margin for error is zero. This material is the workhorse of the modern ultra-supercritical power plant, and treating it like common carbon steel will cost you millions in downtime.
1.2 Grade Comparison and Standard Basis
In the real world, nomenclature can be a minefield. You order P91, but the mill certificate says 10Cr9Mo1VNbN. You get E911, and the drawing calls for X10CrMoVNb9-1. It’s the same family, but the devil’s in the details. Here’s the breakdown from my dog-eared reference book:
-
ASME/ASTM Designation:T91 (Tube), P91 (Pipe). The ‘T/P’ is critical.
-
EN Designation:X11CrMoWVNb9-1-1 (Material Number 1.4905). Note the ‘W’ (Tungsten) – that’s the primary differentiator that gives E911 its superior high-temperature edge over standard P91.
-
Other Common Names:E911 (from the European standard), 9Cr-1Mo-V-Nb-N modified. Some old-timers still call it “the modified 9Cr.”
The core standards we live by are:
-
ASME SA-335/SA-335M:This is the bible for seamless ferritic alloy-steel pipe for high-temperature service.
-
EN 10216-2:The European counterpart for seamless steel tubes for pressure purposes with specified elevated temperature properties.
I always keep a copy of VdTÜV 511/2 in my bag too. It’s a German standard, but its supplementary requirements for creep rupture testing are often more stringent and give you a better picture of long-term performance.

1.3 Definitions of Terms and Abbreviations
In the mill and on site, we don’t always use the fancy terms. Here’s the real-world translation:
-
UTS:Ultimate Tensile Strength. In the field, we just call it “tensile.” “What’d the tensile come back at?”
-
YS:Yield Strength. That’s the “yield point.” The line it crosses before it won’t snap back.
-
PWHT:Post-Weld Heat Treatment. Or as we often call it, “the big bake.” Get this wrong, and you’re welding butter.
-
δ-ferrite:The enemy. A metallurgical phase that kills toughness. We talk about it in hushed tones.
-
Creep:The slow, painful stretching of metal under stress and heat. It’s why we’re all here.
-
MTR:Mill Test Report. The piece of paper that either proves you’re good or gives you a headache. Never lose it.
II. Core Technology Requirements for Materials
2.1 Chemical composition control (smelting analysis + finished product analysis)
This is where the magic, or the tragedy, begins. The chemistry is the recipe. I’ve seen heats come in with perfect chemistry on paper, but the final product is brittle because of a tramp element or an imbalance in the residuals. For E911, the precision required is higher than a surgeon’s.
Table 2.1-1: Chemical Composition Requirements for E911 (Weight %)

| Element | ASME SA-335 (P91) | EN 10216-2 (E911/X11CrMoWVNb9-1-1) | Why It Matters (The Field Engineer’s View) |
|---|---|---|---|
| Carbon (C) | 0.08 – 0.12 | 0.09 – 0.13 | The backbone. Too low, and you lose strength. Too high, and you’re welding with a bag of headaches. I aim for the middle, around 0.10-0.11%. |
| Manganese (Mn) | 0.30 – 0.60 | 0.30 – 0.60 | Deoxidizer and strength helper. We watch it closely with Sulfur. |
| Phosphorus (P) | ≤ 0.020 | ≤ 0.020 | The impurity. We fight to keep it as low as possible. 0.015% max is my unofficial rule. |
| Sulfur (S) | ≤ 0.010 | ≤ 0.010 | Another impurity. It causes hot shortness. We run desulfurization hard in the melt shop. |
| Silicon (Si) | 0.20 – 0.50 | 0.10 – 0.50 | Deoxidizer. Helps with high-temp oxidation resistance. |
| Chromium (Cr) | 8.00 – 9.50 | 8.50 – 9.50 | The oxidation boss. Forms the protective scale. On the low side, you scale. On the high side, you promote delta ferrite. We aim for 8.8-9.1%. |
| Molybdenum (Mo) | 0.85 – 1.05 | 0.90 – 1.10 | Solid solution strengthener. It’s like rebar in concrete at high temp. |
| Vanadium (V) | 0.18 – 0.25 | 0.18 – 0.25 | Forms fine carbides/nitrides for precipitation strengthening. We call these the “workhorses.” |
| Niobium (Nb) | 0.06 – 0.10 | 0.06 – 0.10 | Also forms stable carbides. Fine-tunes the grain structure. |
| Nitrogen (N) | 0.030 – 0.070 | 0.040 – 0.090 | Critical for forming those V/Nb carbonitrides. We manage it tightly with V and Nb. A common rookie mistake is having a V/N ratio off. |
| Nickel (Ni) | ≤ 0.40 | 0.10 – 0.40 | Helps toughness, but too much lowers the Ac1 temperature, complicating PWHT. |
| Aluminum (Al) | ≤ 0.02 | ≤ 0.02 | A deoxidizer. But any Al over 0.02% will form AlN and tie up nitrogen, robbing the V and Nb. We use it as a tracer for bad deoxidation practice. |
| Tungsten (W) | Not Specified | 0.90 – 1.10 | The E911 Signature!This is what gives it the edge. Tungsten provides additional solid solution strengthening and slows down the coarsening of the M23C6 carbides. It’s the secret sauce for higher creep strength. |
| Boron (B) | Not Specified | 0.0005 – 0.0050 | A trace addition, but potent. It segregates to grain boundaries, strengthening them and improving creep ductility. We measure it in parts per million (ppm). |
Personal Take:I was on a job in South Korea where a mill was having terrible impact test failures on P91. They kept fiddling with the heat treat. I asked for the melt data and saw their Nitrogen was consistently at 0.025%, just on the low edge of spec, and their Vanadium was on the high end at 0.24%. The V/N ratio was 9.6, way too high. You need enough N to form all those fine VN particles. We convinced them to target 0.05% N. Problem solved overnight.
2.2 Heat Treatment Process Specifications
You can have the perfect chemistry, but if you mess up the heat treatment, you have a very expensive, very heavy paperweight. For 9Cr steels, the heat treatment is a three-act play: Normalize, Quench, Temper.
-
Normalizing (Austenitizing):Heat to 1040°C – 1090°C (1900°F – 1995°F). Hold for a minimum of 30 minutes. The goal here is to dissolve all the primary carbides and get everything into a solid solution. Too low, and not all the V and Nb go into solution, robbing your final strength. Too high (over 1100°C), and you get runaway grain growth and, you guessed it, delta ferrite. I’ve seen pipes normalized at 1120°C; the grain size was like coarse gravel, and the creep life was shot.
-
Quenching (Cooling):This is the “transformation” step. It must be rapid enough to cool the entire wall thickness to below the martensite start (Ms) temperature before ferrite or bainite can form. For heavy-wall P91 pipe, this means a full water quench is often mandatory. Air cooling is only for thin walls. If you cool too slow, you get bainite, which has lower creep strength. The pipe comes out of the quench as hard, brittle, untempered martensite.
-
Tempering: Heat to 730°C – 780°C (1350°F – 1435°F). This is where we “take the edge off.” We precipitate those fine V/Nb carbides inside the martensite laths, which gives us our strength. And we temper the martensite itself to improve toughness and ductility. The tempering temperature is critical. Too low, and you’re brittle. Too high, and you’re approaching the lower critical temperature (Ac1), where you start to re-austenitize, forming fresh, untempered martensite on cooling. That’s a recipe for a brittle, low-toughness structure known as “overtempering.”
2.3 Mechanical Performance Indicators
The proof is in the pulling, the hitting, and the long-term stretching.
Table 2.3-1: Mechanical Properties at Room Temperature

| Property | ASME SA-335 P91 | EN 10216-2 E911 | Field Acceptance Criteria |
|---|---|---|---|
| Tensile Strength (Rm) | ≥ 585 MPa (85 ksi) | 620 – 850 MPa | The EN range is tighter. I’m wary of anything over 800 MPa in the as-received condition—it can signal undertempering. |
| Yield Strength (Rp0.2) | ≥ 415 MPa (60 ksi) | ≥ 450 MPa | EN has a higher bar. |
| Elongation (A) | ≥ 20% (for full wall) | ≥ 19% (longitudinal) | A measure of ductility. Low elongation means trouble. |
| Hardness (HBW) | ≤ 250 HBW (common spec) | 200 – 270 HBW | This is your quick field check. If you can’t get a hardness test below 250 HB (ASME) or within the EN band, stop everything. |
| Impact Toughness (CVN) | 27 J minimum @ RT (often specified) | 40 J average @ 20°C (for longitudinal) | This is the toughest (pun intended) spec to meet. Low toughness usually points to heat treat or chemistry issues. I’ve seen P91 with 200+ J, and I’ve seen it with 10 J. The difference is process control. |
2.4 Surface Quality and Internal Quality
-
Surface:Every inch of the pipe must be free from laps, cracks, seams, and other imperfections. We specify that any repair by grinding must result in a wall thickness still within the minus tolerance. Welding repair on the base pipe is a huge red flag and generally not allowed without specific approval. It tells me their process was out of control.
-
Internal:We are looking for laminations, cracks, and large non-metallic inclusions. This is where NDT comes in.
III. Dimensions and Weight Specifications
3.1 Dimensional parameters and tolerances
We don’t just order pipe; we order a specific geometry. The tolerances are tighter for these high-value alloys than for carbon steel. You can’t just have a “nominal” schedule. It’s all in the decimal points.
-
Outside Diameter (OD):For NPS 4 and over, ASME B36.10 gives a tolerance of +1/8 in., -1/32 in. for most schedules. For heavy-wall pipe, we often negotiate tighter, say +1.6 mm / -0.8 mm.
-
Wall Thickness (WT):Typically ±12.5% of the nominal wall. But if you’re designing a header with a specific minimum wall for creep life, you must order to that minimum, not a nominal with a minus tolerance.
-
Length:Usually a range, with a specific tolerance on the ends for squareness. A poor end cut can ruin a weld setup.
3.2 Theoretical Weight Calculation
Theoretical Weight (kg/m) = (OD – WT) * WT * 0.0246615 * Density Factor.
For steel, the density factor is roughly 1. For 9Cr steels, the density is around 7.78 g/cm³, slightly less than plain carbon’s 7.85. So, the exact formula for ordering is:
Weight (kg/m) = (OD – WT) * WT * 0.0246615 * (7.78/7.85)
This matters because you pay for the theoretical weight. If the mill runs their pipe heavy on the wall (within tolerance), your tonnage goes up, and so does your bill. I’ve seen procurement fights over a 2% weight variance on a 200-ton order.
IV. Production Process and Control
4.1 Production Process Flow
Let’s walk the floor. For seamless pipe of this grade, the majority is made by the Mannesmann plug mill process or the hot extrusion process.
-
Steelmaking:It starts in the Electric Arc Furnace (EAF) with strict control of scrap selection. Then it goes to a Ladle Furnace (LF) for fine-tuning the chemistry—adding those critical V, Nb, Ti, B. Finally, Vacuum Degassing (VD) to remove hydrogen and oxygen. This is the most critical step for cleanliness.
-
Ingot or Billet Casting:Usually continuous casting into a round billet. The casting must be done under inert gas shrouding to prevent re-oxidation. Surface conditioning of the billet (grinding) is mandatory to remove any surface defects that would turn into seams in the pipe.
-
Heating & Piercing:The billet is slowly and uniformly heated in a rotary hearth furnace. Then it’s pierced in a Mannesmann piercer to create a hollow shell.
-
Elongating (Plug Mill):The hollow is rolled over a mandrel bar to achieve the desired wall thickness and OD.
-
Sizing & Straightening:The pipe is sized to final dimensions and then straightened. This is a cold working step that can introduce residual stress if not controlled.
-
Heat Treatment (N+Q+T):As described in section 2.2. The pipe is normalized, quenched (often with an external and internal water quench system), and tempered in a continuous roller hearth furnace.
-
Finishing & Inspection:Cutting, de-burring, non-destructive testing (UT, Eddy Current), visual inspection, and dimensional checks.
4.2 Key Points for Controlling Critical Processes
Personal Anecdote: In the late 90s, I was at a mill in Germany that was one of the first to produce heavy-wall P91 for a project in the UK. They kept failing the UT inspection on the first few pipes. Internal cracks. The mill manager was pulling his hair out. We traced it back to the water quench. martensite, causing quench cracks. The solution was to slow the cooling rate slightly through the Ms temperature by adjusting the water flow and using a polymer quenchant. It was a hard lesson in physics.
-
Quench Control for Heavy Walls:For WT over 40mm (1.5 inches), the cooling rate is your biggest challenge. You need to cool the inside and outside fast enough to avoid bainite. This often requires dedicated internal and external quench systems. We monitor quench water temperature, flow rate, and pipe temperature during quench using pyrometers.
-
Straightening Stress:Cold straightening after tempering is necessary, but it imparts residual stress. If you straighten too aggressively, you can exceed the yield point locally. We always measure straightness, but we also sample for residual stress measurement if the pipe is for a critical application. You don’t want your perfectly heat-treated pipe to warp during the first high-temperature service.
-
Grain Size Control:We aim for a fine, uniform grain size (ASTM 7 or finer). This is controlled by the normalizing temperature and time. Coarse grain means low toughness. We do metallographic checks on every heat.
V. Inspection & Acceptance Specifications
5.1 Inspection Category Classification
Let me break this down the way we do it on the shop floor—three distinct categories: Mandatory, Supplementary, and For-Cause. I’ve seen too many procurement specs that just check every box on the list. That’s not quality control; that’s just burning money. You need to know where to put your attention.
Table 5.1-1: Inspection Category Matrix

| Inspection Item | Method/Standard | Frequency | Acceptance Level | Field Notes |
|---|---|---|---|---|
| Category A: Mandatory (Every Heat/Lot) | ||||
| Chemical Analysis (Ladle) | ASTM E415 / ISO 14284 | 1 per heat | Table 2.1-1 | This is your fingerprint. Keep it. |
| Chemical Analysis (Product) | ASTM E415 / ISO 14284 | 1 per 200 pipes | Table 2.1-1 | Check for segregation. I’ve seen centerline segregation kill toughness. |
| Tensile Test @ RT | ASTM E8 / ISO 6892-1 | 2 per heat/lot | Table 2.3-1 | If the yield is too high, suspect undertempering. |
| Hardness Test | ASTM E10 / ISO 6506-1 | 2 pipes per lot | 180-250 HBW (my range) | Your quick field check. I reject anything over 260 HBW on the spot. |
| Flattening Test | ASTM A370 / ISO 8492 | 2 per heat/lot | No cracks | Simple but tells you if the pipe is brittle. |
| Hydrostatic Test | ASTM A999 / ISO 10332 | 100% | No leakage | Standard. But for heavy wall, we often waive this and rely on UT. |
| Ultrasonic Examination | ASTM E213 / ISO 10893-10 | 100% | Quality Grade U3 | Catches internal lamination. Non-negotiable. |
| Dimensional Inspection | Calipers, Micrometers | 100% | ASME B36.10 | Wall thickness is where mills try to cheat. Watch the minus tolerance. |
| Visual Inspection | Unaided eye | 100% | No laps, cracks, seams | If you see a repair weld on the base pipe, stop. Reject it. |
| Category B: Supplementary (When Specified) | ||||
| Impact Toughness (CVN) | ASTM E23 / ISO 148-1 | 3 specimens per set | ≥ 40J avg @ 20°C | This is where good heat treatment proves itself. |
| Elevated Temp Tensile | ASTM E21 / ISO 6892-2 | 1 per heat | By design curve | For design data. We plot it against the ASME allowable stress. |
| Creep Rupture Test | ASTM E139 / ISO 204 | 1 per heat (rare) | ≥ 100,000h life | The gold standard. Takes a year to get results. |
| Metallography | ASTM E3, E407 / ISO 4967 | 1 per heat | Grain size ≥ ASTM 7 | I want to see tempered martensite, no δ-ferrite. |
| Micro-cleanliness | ASTM E45 / ISO 4967 | 1 per heat | Thin series ≤ 2.0 | Inclusions kill creep life. Period. |
| Category C: For-Cause (Troubleshooting) | ||||
| Residual Stress Measurement | XRD or Hole-drilling | As needed | ≤ 80 MPa | If pipes warp during machining, check this. |
| Hydrogen Analysis | LECO / Inert Gas Fusion | As needed | ≤ 2 ppm | For sour service or heavy wall. |
| SEM/EDS Analysis | Fractography | As needed | N/A | When something breaks and you need to know why. |
5.2 The Critical Curves: What the Data Actually Tells You
I don’t just look at numbers on a page. I plot them. Every time. A single data point can lie, but a curve—a curve tells you the story. Here are the three curves I keep taped to the wall of my office.
5.2.1 The Impact Transition Curve (Ductile-to-Brittle Transition)
This is the first thing I ask for when toughness is a concern. For martensitic steels like E911, the DBTT (Ductile-to-Brittle Transition Temperature) should be well below room temperature. If it’s not, your heat treatment is wrong, or your grain size is too coarse.
CHARPY IMPACT ENERGY (J) vs. TEMPERATURE (°C)
===============================================================================
300 |
| *
| * * * Fully Ductile
200 | * Region (Upper Shelf)
| *
| * (Target: >100J @ RT)
100 | *
| * * = Good Heat Treat
| * o = Poor Heat Treat
50 | * (DBTT too high!)
| * o
| o o
0 |__o____o___o___o___o___o___o___o___o___o___o___o___o___o___o___o___
-80 -60 -40 -20 0 20 40 60 80 100 120 140 160
Temperature (°C)
===============================================================================
Legend:
* - Properly heat-treated E911 (DBTT ~ -40°C, Upper shelf ~220J)
o - Improper heat treatment (DBTT ~ +20°C, Upper shelf ~80J)
Critical Observation:
The 'o' curve shows DBTT at +20°C. At room temperature, this material
is still in the transition zone. One cold morning, or a slight notch,
and it fractures. I rejected a whole heat in Texas in 2003 for this.
What I look for:
-
Upper Shelf Energy: Should be > 100J, preferably > 150J. Low upper shelf means dirty steel or wrong temper.
-
DBTT: Should be below -20°C, ideally -40°C or lower. If it’s near 0°C, you’re living dangerously.
-
Transition Width: A sharp, steep transition indicates uniform microstructure. A drawn-out transition suggests mixed grain sizes.
Personal Story: 2005, a job in Alabama. P91 headers failing impact tests at 15J at room temperature. Mill certificate said “meets spec.” I asked for the full transition curve. They hadn’t run one. We ran it. DBTT was at +30°C! The material was fully brittle at operating temperature. The culprit? Normalizing temperature was too low—primary carbides hadn’t dissolved, so the matrix was lean and the grain boundaries were weak. We had to re-normalize the whole lot. Cost them six weeks.
5.2.2 The Creep Rupture Curve (Larson-Miller Parameter)
This is the truth-teller for high-temperature materials. You can’t wait 100,000 hours for test results, so we use the Larson-Miller Parameter (LMP) to extrapolate.
The Formula:
LMP = T (C + log t) x 10^-3
Where:
T = Temperature (Kelvin)
t = Rupture Time (hours)
C = Material Constant (typically 20-22 for 9Cr steels)
For E911, with its tungsten addition, the LMP curve shifts to the right compared to standard P91. That means higher stress for the same life, or longer life for the same stress.
STRESS (MPa) vs. LARSON-MILLER PARAMETER (C=20)
===============================================================================
200 |
| E911 (with Tungsten)
180 | * * * * * * * * * * * * * * * * * * * * * * * *
| * * * * * * * * * * * *
160 | * P91 (Standard)
| * * * * * * * * * * *
140 | *
| *
120 | *
| *
100 | *
| *
80 | *
| *
60 | *
| *
40 | *
| *
20 | *
| *
0 +----+----+----+----+----+----+----+----+----+----+----+----+----+
18 19 20 21 22 23 24 25 26 27 28 29 30
Larson-Miller Parameter (x10^-3)
===============================================================================
Operating Point Example:
600°C (873K), 100,000 hours
LMP = 873 x (20 + log 100,000) x 10^-3
log 100,000 = 5
LMP = 873 x (25) x 10^-3 = 21.825 x 10^-3
At LMP = 21.8:
P91 allowable stress ≈ 65 MPa
E911 allowable stress ≈ 82 MPa
That's a 26% improvement. The tungsten matters.
What the curve tells me:
-
The Scatter Band: I don’t just want the average line. I want to see the individual data points. Wide scatter means poor process control.
-
The Extrapolation: We’re projecting from 10,000-hour tests to 100,000-hour life. If the curve isn’t smooth and well-behaved, I don’t trust the extrapolation.
-
The “Knee”: Some materials show a change in slope at long times. That’s where the microstructure degrades. E911’s tungsten delays that knee.
Personal Story: 2010, a utility in the UK was requalifying their superheater headers for a life extension. Original P91, 150,000 hours service. They pulled samples for creep testing. The data points fell below the original design curve. Microstructure showed the M23C6 carbides had coarsened badly—they were like pebbles instead of fine sand. They had to de-rate the unit. Had it been E911, with its tungsten-stabilized carbides, they’d probably have gotten another 50,000 hours. That’s the difference between a capital expense and an operating expense.
5.2.3 The Continuous Cooling Transformation (CCT) Curve
This isn’t something you test on the final product. This is something the mill should have used to design the quenching process. But when I’m troubleshooting a bad batch, I ask for it.
TEMPERATURE (°C) vs. TIME (seconds) - CCT Diagram for E911
===============================================================================
1100 |
| Austenite Region
1000 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
900 | Critical Cooling Curve
| /
850 +--------------------/--+-----------------------------------
| Ferrite Start / |
800 | / | Bainite Start
| / | /
750 | / | /
| / | /
700 | / | /
| / | /
650 | / | /
| / | /
600 | / | /
| / |/
550 | / +-----------------------------------
| / / Martensite Start (Ms ~ 400°C)
500 | / /
| / /
450 | / /
| / /
400 |/________________/________________________________________
| |
| Fast Cool Slow Cool
| (Water Quench) (Air Cool)
| 100% Martensite Mixed Microstructure
| (Bainite + Martensite)
|
0 +----+----+----+----+----+----+----+----+----+----+----+
1 2 5 10 20 50 100 200 500 1k 2k 5k 10k
Time (seconds, log scale)
===============================================================================
Critical Observations:
- To avoid ferrite formation, cooling from 850°C to 500°C must take < 120s
- For heavy wall pipe (>40mm), this requires internal + external water quench
- If you get bainite, you lose 15-20% creep strength
My rule of thumb: For every 10mm of wall thickness, you need approximately 3-4 seconds of cooling rate control through the nose of the curve. A 50mm wall pipe needs to cool from austenitizing to below 500°C in under 20 seconds. That’s aggressive. That’s why heavy-wall P91/E911 is a specialty product.
5.3 Acceptance Rules & Judgment
The standards tell you what to do when a test fails. Experience tells you what it means.
Table 5.3-1: Failure Mode Decision Tree
| Test Failure | Immediate Action | Possible Causes | My Field Decision |
|---|---|---|---|
| Tensile – Low UTS/YS | Retest 2 additional samples | Underaustenitized (low normalizing temp); Overtempered | If retest passes, accept. If fails, reject heat. |
| Tensile – High UTS/YS | Check hardness | Undertempered; Low tempering temp | If hardness > 280 HBW, reject. Can re-temper if caught early. |
| Impact – Low (<27J) | Retest 3 specimens | Coarse grain; δ-ferrite; High Al/N | If average of retest < 40J, reject. This is non-negotiable for me. |
| UT Indication | Rework area by grinding; Re-UT | Inclusion; Lamination; Crack | If depth < 5% WT, repair. If >5% WT or crack, cut out or reject. |
| Hardness – High | Check tempering record | Furnace deviation; Wrong temper | Can re-temper ONE TIME. If still high, reject. |
| Hardness – Low | Check microstructure | Overtempered; Mixed structure | Usually reject. Low hardness means low creep strength. |
| Flattening – Crack | Metallography on crack | Embrittlement; Inclusion stringers | Automatic reject. This pipe is brittle. |
The “One-Time Re-Temper” Rule:
I allow one re-temper. That’s it. Here’s why:
- First re-temper can correct an undertempered condition
- Second re-temper risks overtempering or even hitting the Ac1 temperature
- Multiple heat treatments coarsen the grain structure
I had a mill in Italy try to re-temper a batch three times. The hardness finally came down, but the grain size had grown from ASTM 8 to ASTM 4. The creep life was shot. We rejected 80 tons.
The 5% Grind-Out Rule:
For surface defects, we allow grinding, but:
- Must blend smoothly (no sharp notches)
- Wall thickness after grinding must still meet the minimum specified (not just nominal minus tolerance)
- Area must be re-inspected by MPI or UT
- No grinding in the last 150mm of the pipe end (weld area)
If they grind through the minimum wall, that pipe is scrap. I don’t care if it’s just a spot. A thin spot under creep conditions is a failure waiting to happen.
5.4 Statistical Process Control (SPC) in Acceptance
This is something the standards don’t tell you, but I do it on every major project. I don’t just accept individual values; I look at the distribution.
HARDNESS DISTRIBUTION - E911 PIPE (Target: 220 HBW)
===============================================================================
Frequency
^
|
20 | Normal Distribution
| (Good Process Control)
15 | ***********
| ***************
10 | *****************
| ******************
5 | ********************
| ********************
0 +---*---*---*---*---*---*---*---*---*---*---*---*--->
180 190 200 210 220 230 240 250 260 270 280 290 300
Hardness (HBW)
===============================================================================
Overlay:
Poor Process Control:
....................**....******....******....**.....
(Bimodal distribution - mixed microstructure!)
My Acceptance Criteria:
- Mean: 210-240 HBW
- Standard Deviation: < 15 HBW
- No individual readings > 260 HBW or < 180 HBW
- Distribution must be unimodal and symmetric
If I see a bimodal distribution (two peaks), it tells me the heat treatment wasn’t uniform. Maybe the furnace temperature varied, or the quenching was uneven. Even if all the individual values are “in spec,” I’ll reject the lot. Why? Because in service, the soft spots will creep faster, and the hard spots may be brittle. It’s a mismatch waiting to fail.
VI. Labeling, Packaging and Transportation
6.1 Product Labeling Standards
Every single pipe gets stenciled. It’s its passport.
-
Standard Marking:Manufacturer’s Name, ASTM/EN Specification (SA-335 P91 / EN 10216-2 1.4905), Size (NPS or OD x WT), Heat Number, Piece Number.
-
The “Ritchie” Rule:I always specify that the marking must be with a low-stress, non-hardening ink or paint.Never, ever, everuse a steel stamp to identify P91/T91. Those stamp marks are stress risers and potential crack initiation sites. I’ve seen it happen. A cold stamp on a high-strength martensitic steel is just asking for trouble. I’ve had more than one argument with a yard foreman about this.
6.2 Packaging and Protection
-
End Protection:Every pipe end gets a heavy-duty plastic or steel cap. The bevels are machined and must be protected from impact damage. A dented bevel is a bad weld start.
-
Bundling:Pipes are bundled with steel straps with protective corners. We use wooden dunnage between layers to prevent scratching and moisture trapping.
-
Storage:Keep them off the ground. On skids, under cover, in a dry environment. Water sitting on a pipe for months can lead to pitting corrosion, which is a crack starter.
VII. Special Instructions and Customization Requirements
This is where we separate the standard orders from the critical ones.
-
Hydrogen Bake-out:For heavy wall pipe intended for sour service or critical H2 environments, we may specify a post-manufacturing hydrogen bake-out at a low temperature (eg, 300°C) to ensure any residual hydrogen from the steelmaking process has diffused out. This prevents hydrogen-induced cracking (HIC).
-
Trace Element Control:For ultra-supercritical applications, we might impose additional limits on trace elements like Sn, As, Sb, and Cu (the “tramp” elements) to below 0.01% each. These can segregate to grain boundaries and embrittle the steel over decades of service.
-
Simulated PWHT:Often, the buyer will want the mechanical tests performed on material that has been given a simulated post-weld heat treatment (eg, 760°C for 4-8 hours). This confirms that the base metal’s properties won’t be degraded by the welding and heat treatment process in the field.

