THE EFFECT OF A LONG POST WELD HEAT TREATMENT ON THE INTEGRITY OF A WELDED JOINT IN A PRESSURE VESSEL STEEL

C. Smith: Department of Materials Science and Metallurgical Engineering, University of Pretoria, Pretoria 0002, South Africa

P. G. H. Pistorius: Department of Materials Science and Metallurgical Engineering, University of Pretoria, Pretoria 0002, South Africa

J. Wannenburg: Department of Mechanical and Aeronautical Engineering, University of Pretoria, Pretoria 0002, South Africa

Keywords: Post Weld Heat Treatment, Microstructure, Mechanical properties, C-Mn-Mo Steel, Pressure vessel, Toughness, Impact properties

ABSTRACT

The effect of a long post weld heat treatment on the microstructure and mechanical properties of a welded joint in a 0.2%C – 1.4%Mn – 0.5%Mo pressure vessel steel was studied. Multipass submerged-arc welds were made at a heat input of 1.2 and 4.3 kJ mm^-1. Individual microstructural regions observed in the heat-affected zone of the actual weld were simulated. These regions were brittle in the as-simulated condition. Post weld heat treatment for periods of up to 40 h at 620 ⁰C resulted in a significant improvement in the Charpy impact toughness. At the same time, a loss of the heat-affected zone and weld metal hardness and transverse weld strength occurred. A fracture toughness (J_IC) of 134 kJm ^-2 was measured in the heat-affected zone of the 4.3 kJ mm^-1 welds after prolonged post weld heat treatment. The improvement in weldment toughness with post weld heat treatment was primarily attributed to softening of the structure.

1 INTRODUCTION

Various codes and specifications require that welded structures such as pressure vessels and offshore platforms be post weld heat treated (PWHT), depending on the type and thickness of the welded joint. Post weld heat treatment reduces the effect of any stresses induced by the welding process and tempers the heat-affected zone. The PWHT may be performed several times on a structure during fabrication and after weld repairs, resulting in an accumulation of the total time at soaking temperature. In some instances, the original welds of the structure or vessel may be subjected to cumulative PWRT cycles which exceed the amount of time qualified for by the original welding procedure qualification tests. Because PWHT, in some instances, may result in the loss of both the heat-affected zone and weld metal strength and toughness,¹ the mechanical properties of the weld-joint may deteriorate, as the vessel is repaired repeatedly. This may be undesirable, since it is not known whether the mechanical properties of a weldment are still acceptable.

Studies^1,2 have been conducted to study the effect of a long PWHT on the properties of constructional and pressure vessel steel weldments. Post weld heat treatment may have a beneficial, detrimental or negligible effect on the properties (especially toughness) of the weidments, depending on the chemical composition of the steel, welding procedure used and PWHT time and temperature. The purpose of the current study is to provide detailed information on the effect of a long PWHT on the microstructure and mechanical properties of a welded joint in ASTM A302 Gr B pressure vessel steel. Weld-joints of the steel in the as-welded and post weld heat treated conditions were studied. The mechanical properties of the weldments were determined by heat-affected zone (HAZ) and weld metal Charpy impact and hardness tests, HAZ fracture toughness tests and transverse weld tensile tests. Tests on the HAZ were supplemented by Charpy impact and hardness tests on HAZs generated from thermal simulation techniques. Metallographical examination included optical and transmission electron microscopy.

2 EXPERIMENTAL PROCEDURE

2.1 Materials and welding procedure

Normalized ASTM A302 Gr B plate material from manufacturers A and B, with dimensions of approximately 500 x 500 x 30 mm was used. The test steels had a ferrite-pearlite/bainite microstructure. Welds were deposited using the submerged-arc welding (SAW) process; the materials used were a commercial SD3-Mo wire (3.2 mm dia.) in combination with an OERLIKON OPl2lTT basic welding flux. The chemical compositions of the base metal and weld metal are listed in Table 1.

Element	Plate A (HAZ Tests)	Plate B (Weld Metal Tests)	Weld Metal
C	0.19	0.18	0.084
Mn	1.49	1.35	1.35
Si	0.28	0.30	0.291
Mo	0.54	0.58	0.513
S	0.0018	0.0027	0.003
P	0.0017	0.014	0.001
Ni	0.165	0.251	0.079
Cr	0.167	0.126	0.070
Cu	0.010	0.028	0.076
Al	0.016	0.028	0.016

Table 1: Chemical compositions (major elements) of base materials and weld metal obtained from welding with SD3-Mo wire, wt%.

A K-groove preparation (Fig. 1) was used to ensure that the notch in a Charpy specimen would be entirely in the straight-sided HAZ. The weldments were made with the weld axis parallel to the final plate-rolling direction. Impact testing was done transverse to the rolling
direction. The plates were welded without any external restraint. The plates were welded at nominal heat inputs of 1.2 and 4.3 kJ mm^-1. For the lower heat input welds, 12 to 14 runs were needed to complete the joint. In the case of the higher heat input welds, only one 4.3 kJ mm^-1 weld run was deposited on each side. These beads were followed by one capping pass at a heat input of 2 kJ mm^-1. The preheat temperature, interpass temperature and welding parameters are listed in Table 2.

Parameter	Low Heat Input	High Heat Input
Preheat temperature	150°C min; 200°C max	150°C min; 200°C max
Interpass temperature	250°C max	250°C max
Arc voltage	23 V DCEP	32 V DCEP
Current	475 A	600 A
Travel Speed	9 mm s^-1	4.5 mm s^-1
Nominal heat input	1.2 kJ mm^-1	4.3 kJ mm^-1

Table 2: Welding Parameters

2.2 Thermal simulation

The measured impact properties of a weld-joint are critically dependent on the exact location of the tip of the notch of a Charpy specimen. The presence of tougher regions (e.g. the fine grained HAZ) amid expected local brittle regions (e.g. the coarse grained HAZ), often makes it difficult to assess the true contribution of local brittle zones³ to the toughness of the welded joint. With weld simulation, it is possible to replicate certain HAZ regions of the actual weld HAZ to test the toughness of these regions in the as-welded and/or PWHT conditions. In this study, individual microstructures observed in the HAZ of the actual weld were simulated and the toughness tested. Specimen blanks with dimensions of 11 x 11 x 70 mm were machined from steel A transverse to the final rolling direction. After simulation and PWHT the specimens were reduced to standard (10 X 10 X 55 mm) Charpy specimens. The microstructures of four different regions were generated with a simulator which functions on the same principles as that of a Gleeble simulator. The regions simulated were the coarse grained HAZ (CGHAZ), intereritically reheated CGHAZ (ICCGHAZ), subcritically reheated CGHAZ (SCCGHAZ) and the intercritically heated HAZ (ICHAZ).³

Through dilatometry measurements, it was established that the Ac₁-temperature for this steel is about 725⁰C, with the Ac₃-temperature about 835⁰C. Therefore, to simulate the structure of the ICHAZ, a peak temperature of 780⁰C was used. A peak temperature of 1200⁰C was used to simulate the CGHAZ of the actual weld. Specimens were also subjected to a thermal cycle reaching 1200⁰C followed by second thermal cycles with peak temperatures of 780 and 740⁰C to simulate the intercritically reheated CGHAZ, and 450⁰C to simulate the subcritically reheated CGHAZ. The peak temperatures of the above regions were determined through comparison of the microstructure and hardness of the simulated HAZ with that of the actual weld. The cooling time from 800 to 500⁰C during weld simulation corresponded to that of a weld with a heat input of 2.4 kJ mm^-1. The effective heat input during weld simulation was therefore between that of the low (1.2 kJ mm^-1) and the high (4.3 kJ mm^-1) heat input welds.

2.3 Post weld heat treatment

In addition to testing in the as-welded condition, the low heat input weld was treated at 620⁰C for soaking times of 3, 20 or 40 h. A particular section was placed into the furnace and heated from room temperature to the soaking temperature at a rate of 120⁰C h^-1. The high heat input welds were treated at 620⁰C for 3 and 40 h, respectively. The maximum soaking time of 40 h was used to investigate the effect of prolonged PWHT on the microstructure and properties. Simulated specimens were tested in the as-simulated condition as well as PWHT condition, the soaking periods being the same as for the low heat input welds. After completion of the soaking periods the coupon cooled down in the furnace to below 300⁰C at a rate of 60°C h^-1, whereafter it cooled in air. This procedure met the PWHT requirements of the ASME pressure vessel code.⁴

2.4 Mechanical testing

2.4.1 Charpy impact, hardness and tensile tests

For the low heat input welds, Charpy specimens were machined from 2 mm below the surface of the plate. In the case of the high heat input welds, Charpy specimens were machined in such a way that only the weld metal or HAZ of the 4.3 kJ mm^-1 run (and not the 2 kJ mm^-1 cap weld) would be sampled. The impact test specimens were etched and marked for location of the notch in either the weld metal or straight-sided HAZ. For the HAZ, notches were placed as close as possible to the weld fusion line without sampling the weld metal. The notch was cut normal to the plate surface. CVN specimens were tested over a range of temperatures to obtain ductile-to-brittle transition temperature curves. From the fractured surfaces of the specimens, the fracture-appearance transition temperature at 50% cleavage (FATT₅₀) was determined.

Specimens were cut from weldment sections for metallographic inspection and hardness testing. The Vickers hardness (HV₃₀) and microhardness (HV_0.3) were established in the as-welded condition, and after PWHT. Tensile tests were conducted at room temperature, using round tensile specimens machined transverse to the weld direction. Specimen dimensions complied with ASME codes⁵ and with ASTM A370.⁶ For the low heat input welds, the reduced section diameter of the specimens was 12~5 mm. The diameter of the high heat input weld specimens was 8.75 mm, and only the weld metal and HAZ of the 4.3 kJ mm^-1 run were tested.

2.4.2 J-integral fracture mechanics tests

Transverse weld three-point bend specimens (B X B geometry) were machined from the high heat input welds subjected to 40 h of PWHT. The thickness (B) of each specimen was 25 mm, which is almost the plate thickness (30 mm). A through-thickness notch within dimensions conforming to ASTM E813⁷ was machined into the vertical HAZ of each specimen as close as possible to the fusion line. After pre-fatigue cracking, specimens were tested at -10⁰C using standard bend test fixtures. Crack extension was measured using the single specimen unloading compliance method. J_IC - values were determined by using a program developed by Duvenhage,⁸ which complied with ASTM E813.⁷

2.5 Metallography

Optical and transmission electron microscopy were used to identify and characterise the HAZ and weld metal microstructures. Specimens for optical microscopy were mechanically polished and etched with 2% Nital before examination. For the HAZ, thin foils for the transmission electron microscope (TEM) were prepared from specimens of the five simulated HAZ microstructures in the as-simulated condition and after 40 h of PWHT. For the weld metal, thin foils were prepared from the last weld pass in the as-welded condition and after 3 and 40 h of treatment. The preparation procedure for both cases involved wet grinding of specimens of 3 mm in diameter down to about 40 mm thick, whereafter it was electrolytically polished in a 5% percholoric acid-95 % acetic acid solution with an added 0.1% chromate (CrO₃).

The fractured test pieces of the three-point bend specimens were sectioned and the microstructure examined optically to determine the actual microstructure at the point of unstable fracture initiation. Sectioning was also done perpendicular to the plate surface and the fractured surface.

3 RESULTS

3.1 Mechanical properties

3.1.1 Impact toughness

The Charpy impact toughness of the HAZ is shown in Fig. 2 as ductile-to-brittle transition temperature curves.
Table 3 lists the FATT₅₀ for each heat treatment condition. It is evident from Fig. 2 and Table 3 that PWHT did not alter the impact properties of the HAZ of the low heat weld greatly. A significant decrease in impact toughness occurred as a result of welding at a heat input of 4.3 kJ mm^-1, such that the FATT₅₀ of specimens post weld heat treated for 3 h increased by 89⁰C. The impact energy of the weld simulation specimens are shown in Fig. 3. The change occurring in the FATT₅₀ of each simulated region due to PWHT is illustrated in Fig. 4. Low impact toughness of the HAZ regions in the as-simulated condition can be observed, with the FATT₅₀ higher than 50⁰C. PWHT at 620⁰C reduced the FATT₅₀ significantly.

Heat Treatment Period at 620°C (h)	FATT50 (°C) 1.2 kJ mm^-1	FATT50 (°C) 4.3 kJ mm^-1
Heat-affected zone
No PWHT	-74	-
3	-78	11
20	-66	-
40	-68	-12
Weld metal
No PWHT	-65	-
3	-60	-29
20	-71	-
40	-69	-31

Table 3: Fracture-appearance transition temperatures (at 50% cleavage) after various periods of PWHT (low as well as high heat input) for both the HAZ and weld metal.

The Charpy impact energy of the weld metal is shown in Fig. 5 and Table 3. As with the HAZ toughness, no appreciable changes occurred with PWHT. Welding at a higher heat input decreased the weld metal toughness at intermediate testing temperatures, but did not influence the upper shelf energy values.

Comparison of the FATT₅₀ of the low heat input weld metal with that of the HAZ in the PWHT conditions (Table 3) indicated the toughness to be fairly similar. Welding at a higher heat input seemed to be less detrimental to the weld metal than to the HAZ. Therefore, the integrity of a joint produced by high heat input welding seems to be dependent on the toughness of the HAZ and not on the weld metal.

3.1.2 HAZ fracture toughness

For a high heat input weld PWHT for 40 h at 620⁰C an average HAZ fracture toughness (J_IC) of 134 kJ m^-2 was measured. Fracture was unstable without any stable crack growth prior to fracture, and was the result of 'pop-ins', or short-arrested brittle cracks. Fracture originated from the coarse grained heat-affected zone and arrested in either the tougher weld metal or base metal. From the J_1C-value the corresponding K_1C-value were estimated using eqn (1):⁹

K_IC² = EJ_IC (1)

By using an elastic-modulus of E = 210 GPa, the J_IC-value of 134 kJ v ^-2 gives an equivalent K_IC-value of 168 MPaÖm, a relatively high value.

3.1.3 Hardness

Micro-hardness profiles of welds in the as-welded and PWHT conditions are shown in Figs. 6 and 7. Higher hardness levels were obtained with welding at a low heat input than with a high heat input. The width of the HAZ increased with increase in the heat input. PWHT for 3 h at 620°C had little effect on the hardness of the CGHAZ close to the fusion boundary, but softening in this zone occurred after 40 h of treatment. In the weld metal, only a small amount of softening occurred with PWHT. Post weld treatment also resulted in a decrease in the amount of scatter in the weld metal hardness. The weld metal hardness did not decrease significantly with an increase in the heat input.

3.1.4 Weld strength and ductility

Results of the tensile tests on the 1.2 and 4.3 kJ mm^-1 welds are shown in Table 4 for different periods of PWHT. The yield and tensile strength generally decreased with PWHT. In all of the specimens, fracture occurred in the base metal some distance away from the fusion boundary. This is to be expected since the base metal is the softest part of the welded joint (Figs. 6 and 7). Both the yield and tensile strength complied with the specified limits for ASTM A302 Gr B steel.¹⁰ The reduction in area of the specimens essentially remained constant with PWHT. The elongation generally increased with longer periods of PWHT, as the hardness difference between the heat-affected zone and the base metal decreased (Figs. 6 and 7). The elongation complied with specified limits stipulated by ASTM A302 Gr B.¹⁰

3.2 Metallography

3.2.1 Heat-affected zone

Typical as-welded microstructures for the five HAZ regions observed adjacent to the weld interface of the low heat input weldments are presented in Fig. 8(a-e). Similar structures were also obtained with the simulation experiments (see Fig. 9(a, b)). The hardness values of the weld HAZ microstructural regions agreed fairly well with that of the simulated microstructures. For instance, the hardness of the as-simulated intercritically reheated CGHAZ (1200 + 740^°C) differed from that of the as-welded intercritically reheated CGHAZ by only 19 hardness points.

Figure 8: Optical Micrographs of the Various HAZ Regions in the 1.2 kJ mm-1 weld. As-Welded Condition.

The CGHAZ of the last weld pass consisted of a lath martensitic/lower bainitic microstructure, as shown in Fig. 8(a). Transmission electron microscopy of the CGHAZ and subcritically reheated CGHAZ material in the as-simulated (or as-welded) condition also showed lath martensite and lower bainite (Figs 10 and 11). Subcritical reheating did not alter the microstructure of the CGHAZ of the previous pass significantly, and a similar structure to the CGHAZ was observed. Intercritical reheating of the COHAZ of the previous pass (i.e. heating to a peak temperature of 780⁰C after exposure to a peak temperature of 1200⁰C) changed the microstructure significantly. Islands of martensite primarily formed on prior austenite grain boundaries, with a tempered type of martensitic structure being observed intragranularly (see Fig. 8(c)). Reheating to higher temperatures in the intercritical region (Fig. 8(d)) only resulted in larger martensite islands delineating prior austenite grain boundaries, and more martensite present inside the prior austenite grains. In the intercritically heated region of the last weld pass, a microstructure consisting of ferrite and martensite was observed (Fig. 8(e)). Welding at a higher heat input resulted in a change in the CGHAZ microstructure from martensite/lower bainite to a structure dominated by upper bainite.

Figure 9: Optical Micrographs of two of the As-Simulated HAZ Regions

The only microstructural changes produced by PWHT of the HAZ for 40 h seemed to be the formation and/or spheroidisation of carbides on prior austenite grain boundaries, ferrite lath boundaries and prior island boundaries (Fig. 12).

3.2.2 Weld metal

Examination of the as-deposited weld metal for both heat inputs showed the columnar structure¹¹ of the top bead to consist of acicular ferrite (AF) and primary grain boundary ferrite (PF(G)), the acicular ferrite being the largest microstructural component in the structure (Fig. 13(a)). A small percentage of ferrite with aligned second phase (FS(A)) was identified in certain regions together with some polygonal ferrite. The coarse grained regions of the high heat input welds were characterised by ferrite envelopes delineating the prior austenite grain boundaries.

Figure 13: Photomicrographs of High Heat Input Weld Metal in the As-Welded Condition.

In the normalised regions of the low heat input weld metal the ferrite envelopes of the as-deposited structure were eliminated completely and replaced with a more non-uniform fine grained structure of ferrite with second phase. The normalized region of the 4.3 kJ mm^-1 welds consisted of an equiaxed ferrite-pearlite/bainite structure (Fig. 13(b)). For both heat inputs, optical examination of the intercritically reheated regions revealed the presence of a dark etching phase on the prior austenite grain boundaries (Fig. 13(c)). This structure survived the post weld heat treatment.

An increase in the heat input resulted in a general coarsening of the microstructure. The width of the columnar grains and the acicular ferrite lath size increased. Post weld-heat treatment of the weld metal resulted in the precipitation and spheroidisation of carbides on ferrite grain boundaries, as illustrated by the thin foil micrograph in Fig. 14.

4 DISCUSSION

4.1 Impact toughness of the heat-affected zone

The hard, brittle martensitic structure of the as-simulated CGHAZ and subcritically reheated CGHAZ (Fig. 11(a)) is probably responsible for the low toughness of these regions. Tempering of these regions resulted in softening of the ferrite matrix by relaxation of transformation stresses and the precipitation and spheroidisation of M₃C carbides¹² on ferrite lath and prior austenite grain boundaries. This, together with the fine ferritic lath structure, resulted in a high toughness in the PWHT condition (Fig. 4). The presence of lower bainite in these regions probably increased the toughness by obstruction of propagating cracks along the intra-lath Fe₃C carbides.¹²

The microstructure of the intercritically reheated CGHAZ developed by nucleation and growth of austenite at the prior austenite grain boundaries and martensite lath boundaries during heating into the two-phase region. During the heating, carbon partitioning possibly occurred,¹³ resulting in carbon enrichment in the austenite and carbon depletion in the surrounding ferrite matrix. The austenite islands formed on prior austenite grain boundaries and on ferrite lath boundaries (Fig. 8(c)). This new austenite transformed to a fine lath martensite on cooling. The network of connected martensite islands resulted in low impact toughness of the as-simulated intercritically rehated CGHAZ (Fig. 4), even lower than that of the CGH AZ. Akselsen et al.¹⁴ suggested that low impact toughness might be associated with the stress concentrations developed within the surrounding ferrite matrix, resulting in the initiation of brittle fracture in the ferrite. Stress relieving the intercritically reheated CGHAZ resulted in a significant increase in toughness (Fig. 4), more so than that of the CGHAZ and subcritically reheated CGHAZ. Tempering of the martensite probably resulted in the reduction of stress concentrations at the martensite/ferrite interface and the relaxation of transformation strains in the ferrite matrix.¹⁴

Heating of the base metal to a temperature between the A₁ and A₃ temperatures resulted in the formation of the intercritical HAZ (ICHAZ). This region has a martensite-ferrite structure (Fig. 8(e)). This implies that austenite formed during heating into the two phase region. The austenite then transformed to martensite during cooling. The brittle behaviour of the as-simulated ICHAZ may be attributed to the same type of mechanism responsible for the brittleness of the intercritically reheated CGHAZ (i.e. the development of stress concentrations within the ferrite surrounding the martensite, resulting in the initiation and propagation of the microcracks in the ferrite). PWHT again resulted in an increase in toughness (Fig. 4). The same principles which account for the increase in toughness of the intercritically reheated CGHAZ may be applied here. The fine substructure of the islands may also play an important role in obtaining a high impact resistance. No retained austenite was found in any of the simulated HAZ regions.

The high toughness of the as-welded HAZ (Fig. 2) obtained from welding at a low heat input, indicates two important aspects:

(a) Determining the effect of possible local brittle zones in the HAZ on the toughness of the welded joint is difficult, since crack propagation will be influenced by the fine grained HAZ or weld metal of the multi-pass weld, which usually has a high impact resistance.

(b) If a crack is initiated at a local brittle zone, it may be arrested in tougher regions of the weldment. The fact that double-bevel groove preparations generally are used on large structures and vessels, makes arresting of a crack in either the relatively tough base metal or weld metal highly possible. This ensures an inherent degree of safety of the weld joint.¹⁵

Post weld heat treatment had little effect on the HAZ toughness of the low heat input welds. Low temperature impact toughness remained high after stress relief heat treatment for 40 h. It is therefore concluded that prolonged PWHT would not be detrimental to the impact toughness of the HAZ of a low heat input weld. Welding at a higher heat input resulted in a significant decrease in HAZ toughness, even after 3 h of PWHT (Fig. 2). Due to the fact that fewer weld runs were made as a result of the higher deposition rate, the influence of the fine grained region was less. Ductile tearing at higher temperatures occurred mostly in a wider HAZ region, with less weld metal being sampled. Welding with a high heat input also resulted in the formation of an upper bainite structure in the CGHAZ. The heat-affected zone also had a coarser grain size and less fine grained region. The combined effect of these factors possibly resulted in the decrease in the HAZ toughness. Prolonged PWHT (40 h) did not decrease the toughness (Fig. 2), and adequate toughness was still obtained. The reason for the apparent increase in toughness from 3 h to 40 h of PWHT (Fig. 2) is probably attributed to the softening of the CGHAZ with PWHT. The industry¹² usually requires an impact toughness of 36 J at –10°C for this steel with thickness less than 50 mm. In this respect, the average impact toughness of 60 J at –10°C barely exceeded the requirement. The lowest value was 49 J. However, the HAZ toughness of this steel was still adequate up to 4.3 kJ mm^-1, in the stress relieved condition.

4.2 HAZ hardness and strength

The lower hardness of the last pass CGHAZ of the higher heat input welds (Fig. 6) compared to that of the unaltered CGHAZ of the low heat input welds (Fig. 7) is primarily due to the formation of the softer upper bainite structure at slower cooling rates. An increase in toughness was not obtained with the decrease in hardness, indicating the detrimental effect of the bainite. For both heat inputs, PWHT resulted in a decrease in the HAZ hardness to values in the region of the base metal hardness after about 20 h of treatment. The yield and tensile strength still conformed to the minimum requirements of ASTM A302 Gr B (Table 4).

4.3 Impact toughness of weld metal

The high as-deposited weld metal toughness observed in this investigation is attributed to the presence of a high percentage of acicular ferrite with fine interlocking laths in the structure. The absence of high volume fractions of grain boundary ferrite and ferrite with aligned second phase which could result in low toughness, further contributes to the high impact toughness.^16,17

Post weld heat treatment did not alter the impact toughness greatly (Table 3). The detrimental effect of carbides on the grain and lath boundaries, as described by Dunne and Pollard¹⁸ for a Mo-bearing steel weld metal, was not observed. The beneficial effect of ferrite softening probably compensated for the deleterious effect of carbide precipitation on the grain and lath boundaries.¹⁹ The net effect was that the weld metal toughness had been unaffected by PWHT.

In the normalised regions of the low heat input weld metal, the ferrite envelopes delineating prior austenite grain boundaries were replaced by a non-uniform structure of ferrite with second phase. Evans²⁰ studied the effect of Mo on the microstructure and properties of C-Mn weld deposits and found that with the addition of increasing amounts of Mo (from 0.25 - 1.1% Mo) to the weld metal, the well defined equi-axed nature of this region would gradually be replaced by a more non-uniform structure of colonies of ferrite with an aligned second phase. This type of structure was also observed in the present study. The effect of PWHT was to spheroidise the carbides, similar to that of the columnar structure. The ferrite grains also became more equi-axed. For the high heat input welds, it is expected that the coarser equi-axed structure of the normalised region (Fig. 13(b)) would be a factor contributing to the lower toughness experienced by the weld metal.

Optical examination of the intercritically reheated zones revealed a dark etching grain boundary phase (Fig. 13(c)), believed to be Fe₃C.²⁰ During reheating into the austenite-ferrite region, nucleation of austenite possibly occurred at original d-ferrite solidification boundaries because of segregation.²⁰ During subsequent cooling these austenite regions transformed to ferrite/carbide aggregates. This grain boundary phase was also observed in the PWHT condition. The possible detrimental effect of these phases on the impact toughness would not be ascertained because of the presence of the tough acicular ferrite. It is expected, though, that these phases would be detrimental to toughness since they could serve as crack propagation paths.

Welding at a higher heat input resulted in a decrease in the impact toughness of the weld metal. The decrease in toughness is associated with the coarsening of the prior austenite grains, the increase in ferrite lath size and an increase in grain boundary ferrite and ferrite sideplates.²¹ Despite the fall in toughness with high heat input welding, the toughness was still adequate, even after PWHT.

4.4 Weld metal hardness

The weld metal hardness and strength decreased only slightly with prolonged PWHT. The hardness was in the region of 200 Vickers after 40 h of treatment. Welding at a higher heat input did not alter the hardness significantly, despite the coarsening of the structure.

4.5 Fracture toughness of the welded joint

Although the Charpy impact test provides useful information on the impact toughness of a welded joint, it cannot be used in the design of structures against brittle fracture. On the other hand, the fracture toughness may be used to calculate the defect size required to cause fracture, and is thus a measure of the integrity of the joint. In this study, fracture toughness tests were used to:

(1) Assess the applicability of CVN-K_1C correlations to the HAZ of this particular steel.

(2) Determine the integrity of the joint by using a simple fracture assessment.

Efforts to correlate Charpy impact data with K_IC have received considerable notice among structural engineers.^22,23 A typical equation for pressure vessel steel relating Charpy impact energy with K_IC is:²³

K_IC = 14.6ÖCv [MPa Öm] (2)

with Cv = Charpy impact energy in joules. From the Charpy impact test results in Fig. 2, it was established that the average impact toughness of the high heat input weld heat-affected zone after 40 h of PWHT is approximately 103 J at –10°C. Using eqn (2), the K_IC-value is estimated to be 148 MPaÖm. This value agrees well with the measured K_IC value of 168 MPaÖm. PWHT for 40 h is therefore not detrimental to the toughness of joints produced by welding at 4.3 kJ mm^-1. Evidently, the CVN-K_IC correlation in eqn (2) may be used to predict the fracture toughness in the weld HAZ.

A simple assessment of the integrity of the joint, similar to that described in PD-6493-Level One,²⁴ may be performed. Consider the case of a section of a vessel being subjected to an internal pressure. Assume that some of the fabrication welds were subjected to a cumulative post weld heat treatment time of 40 h at 620⁰C. The possibly deleterious effect of the pressure of an external semi-elliptical surface flaw in the wall of the vessel will now be evaluated. Assume that the flaw is situated along a longitudinal weld in the heat-affected zone (Fig. 15). The purpose of the fracture assessment is to determine the critical flaw size which would result in brittle/catastrophic failure of the vessel.

The internal pressure produces a circumferential uniform tensile hoop stress²⁵ in the wall of the pressure vessel (Fig. 15). For a thin-walled pressure vessel²⁵ the hoop stress can be calculated as:

s_hoop = pr/B (3)

where

p = internal pressure

r = inner radius of the transfer line

B = wall thickness.

For an internal pressure of 2.3 MPa, an inner diameter of 1.65 m, and a plate thickness of 30 mm, the hoop stress is 63.3 MPa. Furthermore, for the sake of this argument, assume that uniform tensile residual stresses amounting to about 60% of the yield strength of the base metal are still present in the weld after stress relief heat treatment. This level of residual stress is rather conservative. These levels are usually much lower after PWHT.²⁴ If the room temperature yield strength is 458 MPa after 40 h of PWHT,²⁶ then summing the hoop stress and 60% of yield strength gives a total stress, s_t, of 338 MPa.

For this type of loading and crack geometry the stress intensity factor is given by:²²

for a << B (4)

Where:

a = crack depth.

Q = ø², where ø is an elliptic integral of the second kind.

B = plate thickness.

Assume a severe case for the crack depth to crack length ratio (a/2c) of 0.01. The value for Q is 1.0²² For K_IC = 168 MPaÖm the critical crack depth for failure of the vessel wall is then 65 mm. This value is more than twice the thickness of the wall. The crack will therefore break through the vessel wall causing the vessel to leak. This condition is commonly known as ‘leak-before-break’.²² Provided that no fires or other dangers result from leakage, this represents a safe condition. The above fracture analysis showed that welding ASTM A302 Gr B steel with a heat input of up to 4.3 kJ mm^-1 would result in weld-joints of high integrity, even after a cumulative stress relief heat treatment of 40 h at 620^°C.

5 CONCLUSIONS

For normalised ASTM A302 Gr B steel welded with commercial SD3-Mo wire at a nominal heat input of 1.2 and 4.3 kJ mm^-1, and then stress relieved for up to 40 h, the following conclusions have been drawn:

(1) The coarse grained heat-affected zone, subcritically and intereritically reheated coarse grained heat-affected zone and intercritical heat-affected zone regions in the as-welded (as-simulated) condition were brittle. Post weld heat treatment at 620^°C resulted in a significant increase in toughness of these regions.

(2) Welding at a higher heat input results in a decrease in the heat-affected zone toughness compared to that of the low heat input welds. This is attributed to the presence of upper bainite in the coarse grained heat-affected zone.

(3) Post weld heat treatment for 40 h results in a general decrease in the strength of the weldments. The strength still meets the minimum specified limits. No deterioration in toughness did occur.

(4) High weld metal toughness was obtained for both heat inputs in the as-welded condition. This high toughness is attributed to the presence of fine acicular ferrite. Very little grain boundary ferrite and ferrite with aligned second phase was observed. Welding at a higher input did, however, decrease the toughness due to coarsening of the prior austenite grains and the acicular ferrite laths, and the formation of a coarser equi-axed structure in the normalized regions.

(5) No significant change in the weld metal impact toughness in the as-welded condition occurred with PWHT for 40 h. PWHT resulted in the precipitation and spheroidisation of carbides along grain and ferrite lath boundaries. Contrary to expectation, these carbides did not lower the toughness. Softening during stress relieving possibly compensated for the detrimental effect of the formation of grain boundary and lath carbides.

(6) A relatively high heat-affected zone fracture toughness (J_IC) was measured after 40 h of post weld heat treatment, despite the occurrence of pop-ins. The fracture toughness correlated with the corresponding Charpy impact toughness of the heat-affected zone.

(7) A heat input of up to 4 kJ mm^-1 may be used while adequate HAZ and weld metal toughness for this steel in the PWHT condition are still maintained.

(8) A cumulative PWHT time of up to 40 h does not pose any danger to the integrity of the welded joint.

ACKNOWLEDGEMENTS

The authors wish to thank Sasol Synthetic Fuels for financial and technical assistance. Dr N. van der Berg and A. Botha provided assistance and helpful discussions during the TEM studies. The authors thank M. Meyering at Iscor Pilot Plant for the preparation of the three-point bend specimen notches.

REFERENCES

1. Konkol, P. J., Effects of long-time post weld heat treatment on the properties of constructional-steel weldments. Welding Research Council Bulletin, 1988, 330, 11-26.

2. Provost, W., Effects of a stress relief heat treatment on the toughness of pressure vessel quality steels welded with high heat input processes. International Journal of Pressure Vessels & Piping, 1981, 9, 125-154.

3. Denys, R. M., Local brittle zones and the implications for HAZ toughness testing. Revue de la Soudurel Lastijdschrift, 1988, 44(2), 24-38.

4. ASME Boiler and Pressure Vessel Code, Section VIII-Division 1. Part UCS Carbon and low alloy steel vessels, American Society of Mechanical Engineers, 1989.

5. ASME Boiler and Pressure Vessel Code, Section IX, Part QW Welding. American Society of Mechanical Engineers, 1989.

6. Annual Book of ASTM Standards. Standard specification for general requirements for steel plates for pressure vessels, Designation A 370/A 370M, 1988.

7 Annual Book ofASTM Standards. Standard test method for J_IC, a mreasure of fracture toughness, Designation E, 813, 1987.

8. Duvenhage, G., Remaining life of a pressure vessel subjected to cyclic pressure loading based on probabilistic fracture mechanics. M. Eng. thesis, University of Pretoria, 1994.

9. Annual Book of ASTM Standards. Standard test method for J_IC, a measure of fracture toughness, Designation E, 813, 1981.

10. Annual Book of ASTM Standards. Standard specification for pressure vessel plates, alloy steel, manganese-molybdenum and manganese-molybdenum-nickel, Designation A 302/A 302M, 1987.

11. Evans, G. M., The effect of manganese on the microstructure and properties of C-Mn all-weld metal deposits. Welding Journal 1980, 59(3), 67s-75s.

12. De Villiers, W. M., De Klerk, H. J. and Schwartzer, G. A. G., Microstructure of ASTM A302 Gr B pressure vessel steel intended for use in ashlocks. ISCOR Internal Report by Research and Development, 1991.

13. Koo, J. K. and Ozekcin, A., Local brittle zone microstructure and toughness in structural steel weldments. In Welding Metallurgy of Structural Steels, ed. J. Y. Koo. The Metallurgical Society, 1987, pp. 119-135.

14. Akselsen, 0. M., Grong, 0. and Solberg, J. K., Structure-property relationships in intercritical heat-affected zone of low-carbon microalloyed steels. Materials Science Technology, 1987, 3, 649-655.

15. Webster, S. E. and Bateson, P. H., Significance of local brittle zones to structural behaviour. Materials Science Technology, 1993, 9, 83-90.

16. Grong, O. and Kluken, A. O., Microstructure and properties of steel weld metals. In Ferrous Alloy Weldments, ed. D. L. Olsen et al. Trans Tech Publications, Clausthal, Germany, 1992, pp. 21-46.

17. Dolby, R. E., Factors controlling HAZ and weld metal toughness in C-Mn steels, Proceedings of First National Conference on Fracture, Johannesburg, South Africa. In Engineering Applications of Fracture Analysis, eds G. G.Garret and D. L. Marriott. Pergamon Press, New York, 1979, pp.117-134.

18. Dunne, D. J. and Pollard, G., The effect of stress relieving on the microstructure of C-Mn-0.4%Mo weld metal. In Recent Trends in Welding Science and Technology, ed. S. A. David et al. ASM International, Materials Park, OH, 1989, pp. 269-272.

19. Farrar, R. A., Taylor, L. G. and Harrison, E. M., Effect of stress relieving on fracture properties of submerged-arc welds of C-Mn steels. Metallurgical Technology, 1979, 6(10), 380-389.

20. Evans, G. M., Effect of molybdenum on the microstructure and properties of C-Mn all-weld metal deposits. Joining and Material, 1988, 1(5), 239-245.

21. Evans, G. M., The effect of heat input on the microstructure and properties of C-Mn all-weld metal deposits. Welding Journal, 1982, 61(4), 125s-132s.

22. Broek, D., The Practical Use of Fracture Mechanics. Kluwer Academic Publishers, Amsterdam, 1989, pp. 73-121.

23. Hertzberg, R. W., Deformation and Fracture Mechanks of Engineering Materials, 2nd edn. Wiley, New York, 1983, pp. 339-348.

24. British Standards Institution. Guidance on methods for assessing the acceptability of flaws in fusion welded structures, PD 6493, 1991.

25. Benham, P. P. and Crawford, R. J., Mechanics of Engineering Materials. Longman Scientific & Technical, Essex, UK, 1987, pp.45-46.

26. Hattingh, C., The effect of repeated heat treatment on the mechanical properties of ASTM A302 Gr B (Mn-Mo) steel and API 5L Gr B (C-Mn) steel. M.D. thesis, Pretoria Technicon, 1987.