Dominique Carrouge Houston February 2002 Phase Transformation Group H. K. D. H. Bhadeshia MCAS Technology Group Dr. P. Woollin.

Dominique Carrouge

Houston

February 2002

• Phase Transformation Group • H. K. D. H. Bhadeshia

• MCAS Technology Group • Dr. P. Woollin

Microstructural Change in HighTemperature Heat-Affected

Zones of Welded

Why this project ?

Super Martensitic SS are :

• Increasingly employed in the Oil & Gas industries

• Cost effective alternative to Duplex SS

Engineered to be :

• High strength, Corrosion resistant, Weldable

But :

• Work on understanding microstructure development in HAZ could improve service properties

Phases in HAZ

• Low carbon martensite (ok) • Retained austenite (ok)• Delta-ferrite (?)

• Can reduce toughness and corrosion resistance !

Aim

• Understand HAZ microstructure• Locate delta-ferrite in HAZ of single-pass and

multi-pass welds• Investigate effect of some welding parameters• Predict ferrite retention

HAZ & Phase diagram

Ae5

Ae4

Ae3

Single-pass weld

• Plate : 5 mm thick• Mechanised TIG process• No filler material• Heat-input : 0.7 kJ/mm

Steel A (UNS S41426) : 0.01 C, 12 Cr, 6.5 Ni, 2.5Mo

Microstructure

• Electrolytic sulphuric acid• Villela’s reagent

Coarse-Grained HAZ

Widmanstätten Austenite

Ferrite

(now martensite)

Amount of ferrite negligible

Dual Phase HAZ

• ~ 20 % ferrite • DPHAZ maximum length : 1 mm

X-ray Mapping of Ferrite

MoMo

CrCr

NiNi

Partitioning during

welding !

Less ferrite with high heat-input

0.7 kJ/mm 0.3 kJ/mm

Slow cooling rate Fast cooling rate

Steel A : 0.01 C, 12 Cr, 6.5 Ni, 2.5Mo

Situation in multi-pass weld

Steel A : 0.01 C, 12 Cr, 6.5 Ni, 2.5Mo

• 12 mm thick pipe

• HI : 0.4 - 1.2 kJ/mm

• 25 Cr Duplex filler wire

• 7 passes TIG + MMA

Ferrite distribution

Root

Cap

Fast cooling rate

Slow cooling rate

Modelling ferrite content

• Physical model need reliable thermodynamic data• MT DATA software has been used• Prediction of equilibrium phase transformation

temperatures and volume fractions

A : 0.01 C, 12 Cr, 6.5 Ni, 2.5Mo

B : 0.01 C, 11 Cr, 1.5 Ni, 0.5 Cu

C : 0.01 C, 12 Cr, 3 Ni

1223oC1255oC1235oC

Ae4

High temperature dilatometry

• Heating rate : 10 K/min

Ae4

Comparison

• Poor agreement• Lack of data in the

database• Physical Modelling

cannot be used

Neural network model

• Non linear regression tool that can be used to make extrapolation and predictions

• Not enough data to create a neural network model due to difficulty to measure ferrite content in Martensitic SS

• But a model has been developed for predicting ferrite retention in Austenitic stainless steels

• For a given composition the Ferrite Number can be estimated

http://www.msm.cam.ac.uk/phase-trans/2001/ferrite.number.html/index.html

To avoid ferrite

Minimum level of Nickel required to minimise ferrite :

To avoid ferrite

Minimum level of Nickel required to minimise ferrite :

To avoid ferrite

Minimum level of Nickel required to minimise ferrite :

Summary &Conclusion• HT-HAZ microstructure has been studied• Volume fraction of ferrite trapped in HAZ are small• Mo containing alloy retain more ferrite

Effects of this ferrite on properties are not well established

But if ferrite content is to be minimised :• Slower cooling rate would reduce ferrite content• Alloy chemical composition could be adjusted

Thank you for your attention