P91 Piping Fabrication Guidelines

David Buzza American Electric Power

Combined Cycle User’s Conference October 31, 2011

History

• Grade 91 was developed for high temperature service. Allowable stress of P91 is 2.5 times greater than P22 at 1080F.

• Grade 91 approved by ASME B&PV Code Section I and VIII in 1983-1984.

• Grade 91 has been in service since the 1990’s for retrofits of traditional fossil plants, new supercritical plants and new combined cycle plants.

Most Common Fabrication Issues with P91 Identified by Inspection

• Target Hardness – Components = 190-250 Hb – Welds = 190-280Hb

• Small Bore (</= 4 inch) – Hard Components – Hard Socket Welds – Soft Components

• Large Bore (>4 inch) – Hard Attachment Welds – Soft Components (mostly elbows)

P91 Fabrication Issues – Hard Components

• ASME A335 specifies max allowable hardness of 250Hb.

• Risks - Overload or creep failure due to improper metallurgy

• HP stop valve before seat drain line failure had hardness of 318-351Hb.

• Plant X - identified 4 small bore pipe sections (330-390Hb).

• Plant Y– identified 13 large and small bore components (316-366Hb) HP Stop Valve Before Seat

Drain Line Failure

Failure

Drain Line

P91 Fabrication Issues – Hard Socket Welds & Attachment Welds

• AEP P91 Fabrication specifies less than 280Hb weld hardness

• Risks - Cracking due to SCC or lack of toughness

• Plant X – 30 of 300 socket welds (300-417Hb)

• Plant Y – 76 of 773 socket welds (300-437Hb)

• Plant Y – 34 of 94 attachment welds (315-439Hb)

Socket Weld

Pipe

Elbow

P91 Fabrication Issues – Soft Components

• AEP P91 Fabrication Standard specifies hardness of >190Hb.

• Risks – shortened creep life due to improper metallurgy.

• Plant X and Y – identified numerous components with hardness of <165Hb

Metallurgy

P22 P91 Base Metal Ferrite / Pearlite Martensite

Welding Bainite microstructure. Hardness = 280Hb

Martensite microstructure. Hardness = 400Hb

PWHT Stress relieving (weld shrinkage) and tempering hardness. Bainite transforms to ferrite Hardness = 200Hb

Stress relieving, tempering hardness and formation of precipitates Martensite must be maintained Hardness = 250Hb

If PWHT is above LCT, then microstructure remains the same.

If PWHT is above the LCT, then microstructure changes drastically and component can become soft or hard.

LCT = Lower Critical Temperature

Improper PWHT is the Root Cause for P91 Fabrication Issues

Fabrication Issue Root Cause Hard Components PWHT temperature exceeded LCT

followed by rapid cooling.

Hard Welds PWHT temperature at the weld never reached hold temperature.

Soft Components PWHT temperature exceeded LCT followed by slow cooling. Or PWHT hold time is excessive.

PWHT Demonstration I Hold Temperature > LCT

2” Sch 160 P91 Pipe, Hold Time = 3 hours

Code Hold Temperature = 1350-1425F

Lower Critical Temperature = 1475-1575F

After PWHT target = 190-250Hb

PWHT Demonstration II Hold Temperature > LCT

Monitoring TC

Control TC

1. Thermocouple (TC)

2. Heating Beads

3. Insulation & Cook

PWHT Demonstration II Chart

1400F

1200F

1600 F Both socket welds hold at 1430 F

One Pipe T peaks at 1728 F

Other Pipe & Elbow remain under 1430 F

1800 F

2 Hr

PWHT Demonstration II Final Hardness, Hb

Low hardness - pipe temperature exceeded lower critical temperature

Good hardness - pipe & elbow temperature remained below lower critical temperature

Hardness Target = 190-250Hb

Waterford Drain Pipe Fabrication

• Material – 2 inch socket welded, P91 pipe spools.

• Pipe spools were fabricated in shop then shipped for install.

• Shop PWHT – Furnace – Induction

Spool installation

Drain Pipe Fabrication Furnace PWHT

• Temperature chart showed that target temperature and hold time were met.

• Final hardness check showed 100 of 170 welds to be hard.

• Conclusion – furnace temperature did not accurately reflect the actual weld temperature during heat treating.

Furnace layout of spools

Drain Pipe Fabrication Induction PWHT

• Additional thermocouples, TC’s were added.

• Temperature chart of pipe, elbow and weld showed that target temperature and hold time were met.

• Final hardness check showed that all 28 welds were properly heat treated.

• Conclusion – Proper heat treatment results in good hardness readings.

Keys to Proper Fabrication of P91

1. Check incoming material A. N:Al on the Material Test Report (MTR)

should be greater than 4:1 and under no circumstances less than 2.5:1

B. Hardness on the MTR and own testing should be 190-250Hb. A hardness of 220Hb is preferable because hardness will decrease with each PWHT

Keys to Proper Fabrication of P91 (con’t)

2. Improve heat treating procedures A. Follow AWS D10.10

i. Number of control zones based on pipe diameter ii. Heating blanket length based on pipe diameter and thickness iii. Monitoring thermocouples for soak zone and heat band zone

B. Use heat treating set-up sheets i. Procedure – ramp rates, hold times / temperatures ii. Wrap sheet

a. Heating pad size / placement for each heating circuit b. Control thermocouple placement for each heating circuit c. Monitoring thermocouple placements

C. Review heat treating set up sheets i. Check that hold time has been reduced for smaller welds. ii. Check that heating pads will contact the pipe as much as possible. iii. Check that thick sections and thin sections are on separate heating circuits. iv. Check that control thermocouples are towards center of pad and at the location that

you would expect to be the hottest. v. Check that monitoring thermocouples are where you want the heat (weld) and

where you do not want to overheat (base metal). vi. Check that someone will be monitoring the heat treating and can abort if needed.

Heat Treating Wrap Sheet that Shows Proper Control & Monitoring

• Heating blankets • Control

thermocouples – valve body – pipe soak band

• Monitoring thermocouples – weld – pipe heating band

Heat Treating Wrap Installed

TC’s

weld

Heating blankets

If in doubt, then add TC’s

Keys to Proper Fabrication of P91 (con’t)

3. Confirm proper heat treating by hardness testing

A. Weld hardness reduced. B. Base material hardness maintained within 0 to 20

Brinell hardness points of initial readings.

Note: Hardness testing should be performed on a grid by a qualified tester. Testers will think that they are qualified, but they are not. We use GE MIC10 and Pin brinell or telebrineller

Conclusions

• If your fabricator continues to treat grade 91 materials the same as they treat other traditional high temperature materials (i.e. grade 22), then you will continue to have quality issues.

• You must insist that your fabricator upgrades their heat treating (improved control and monitoring) and adds hardness testing to their procedures to successfully and consistently fabricate grade 91 materials.