Alternative Approach for Qualification of Temperbead ... Approach... · ALTERNATIVE APPROACH FOR QUALIFICATION OF ... to temper the base material, ... Alternative Approach for Qualification

1 Copyright © 2012 by ASME

ALTERNATIVE APPROACH FOR QUALIFICATION OF TEMPERBEAD WELDING IN THE NUCLEAR INDUSTRY

Steven L. McCracken, PE Electric Power Research Institute Charlotte, North Carolina, USA

Richard E. Smith, PhD, FAWS Structural Integrity Associates, Inc. Huntersville, North Carolina, USA

ABSTRACT Temperbead welding is common practice in the nuclear

power industry for in-situ repair of quenched and tempered low alloy steels where post weld heat treatment is impractical. The temperbead process controls the heat input such that the weld heat-affected-zone (HAZ) in the low alloy steel is tempered by the welding heat of subsequent layers. This tempering eliminates the need for post weld heat treatment (PWHT). Unfortunately, repair organizations in the nuclear power industry are experiencing difficulty when attempting to qualify temperbead welding procedures on new quenched and tempered low alloy steel base materials manufactured to modern melting and deoxidation practices. The current ASME Code methodology and protocol for verification of adequate fracture toughness in materials was developed in the early 1970s. This paper reviews typical temperbead qualification results for vintage heats of quenched and tempered low alloy steels and compares them to similar test results obtained with modern materials of the same specification exhibiting superior fracture toughness.

INTRODUCTION Carbon and low alloy steels are ferritic materials that

exhibit a toughness transition over a characteristic temperature range. The range over which the toughness transition occurs varies according to the metallurgical microstructure known as the ductile-to-brittle transition temperature (DBTT) range. It is important that pressure vessels operate at temperatures above the DBTT where fracture toughness is adequate to resist unstable propagation of defects. In contrast, the fracture toughness exhibited below the DBTT is low and can lead to unstable brittle fracture. The design of nuclear pressure vessels account for this fracture toughness transition of ferritic steels by specifying a minimum critical operating temperature before significant pressurization can take place. The ASME codes [1,2] require that a reference nil-ductility temperature (RTNDT) be determined for all heats of vessel materials (including weld metal) making up the pressure boundary. The highest RTNDT

for all the material heats within the pressure boundary becomes the limiting or lowest pressurization temperature. The RTNDT is correlated to a lower bounding trend curve that serves to characterize the fracture toughness properties for quenched and tempered low alloy pressure vessel steels. This reference curve has been established by evaluating an extensive database of static fracture toughness, dynamic fracture toughness, and crack arrest fracture toughness and has been codified by the ASME in Section III, Appendix G [3]. This toughness reference curve has been used successfully for many years to characterize material performance for nuclear pressure vessel low alloy ferritic steels.

Qualification of welding procedures requires that impact testing be performed to demonstrate that the welding process does not adversely affect the base metal fracture toughness properties in the weld heat affected zone (HAZ). The Code specifies that the toughness of the weld HAZ be measured according to a specific testing protocol that involves both the Charpy V-notch test (CV) and the drop-weight test. Both testing requirements have been codified by the American Society for Testing Materials (ASTM).

Modern melting practices, such as vacuum arc remelting, electroslag remelting, argon lancing, or argon-calcium lancing, can produce low alloy ferritic materials having impact properties that are superior to vintage materials produced in the past. The superior material impact properties are beneficial; however, inconsistencies can result when applying the ASME impact test protocol. The reason is that the temperature relationship between the nil-ductility transition temperature (TNDT) determined by the drop weight test and the CV transition temperature appear to be different for vintage materials compared to materials made with modern melting practice. In addition, the determination of TNDT by the drop weight test can be problematic for quenched and tempered low alloy steels (e.g. SA-508 forgings and SA-533 plates) [5,6]. This issue is believed to be related to the hardenability behavior of these steels and the microstructures produced by quenching and tempering following a full austenization heat treatment.

Proceedings of the ASME 2012 Pressure Vessels & Piping Conference PVP2012

July 15-19, 2012, Toronto, Ontario, CANADA

PVP2012-78571


ASME CODE IMPACT TEST REQUIREMENTS Measuring the fracture toughness of ferritic material

requires sophisticated test equipment and in-depth analysis procedures. Moreover, fracture toughness testing is expensive, time consuming, and requires large specimens. As a result, the ASME Code specifies simpler semi-quantitative methods, such as drop-weight and CV testing that can be used to validate adequate fracture toughness. The drop-weight and CV tests do not directly measure fracture toughness, but are useful to index materials to a reference curve that has been established to characterize the materials.

The ASME Code specifies ASTM Method E-208 for standardization of the drop-weight test and ASTM SA-370 for standardization of the CV test. The intended purpose of the drop weight test is to determine a temperature known as the nil-ductility transition temperature (TNDT). The key output material property values of the CV test are absorbed impact energy (ft-lb), mils of lateral expansion (MLE), and % shear.

A standardized relationship between the drop weight and CV test was adopted by the ASME Code in the early 1970s to validate adequate fracture toughness properties of low alloy pressure vessel steels. In concept, TNDT, as measured by the drop-weight test, signals the onset of a monotonic increase in fracture toughness properties. A temperature 60°F greater than TNDT is intended to define a position along the monotonically increasing CV transition curve where a minimum level of toughness can be expected. The minimum toughness for pressure vessel steels defined by the ASME Section III Code, as measured by the CV test, is 50 ft-lb absorbed impact energy and 35 MLE. When this result is achieved the RTNDT and the TNDT are identical. If the minimum impact properties are not achieved at TNDT + 60ºF additional CV testing is performed at higher temperatures until the minimum impact values meet the minimum requirements. For this latter case, the RTNDT is established 60°F below the temperature where the minimum required impact properties are achieved.

Rationale for ASME Code Impact Test Requirements The PVRC Ad Hoc Group on Toughness Requirements

provided the technical basis that defines the minimum CV impact test values and the relationship between the drop weight TNDT and the CV test temperature in the early 1970s. Appendix 2 of WRC Bulletin No. 175 [4] provides discussion and the rationale for this relationship. One basic premise was that impact energy absorption must increase with increasing yield strength in order to maintain constant fracture toughness. Described in another way; absorbed energy will increase with increasing yield strength in order to maintain constant mils lateral expansion. Energy absorption is controlled by two factors: 1) the yield strength of the steel which regulates the force required to plastically deform the CV specimen, and 2) the ductility of the steel which determines the distance over which the trip hammer force acts during the CV testing. In other words the loss of fracture toughness was attributed to a decrease in ductility rather than a decrease in strength. A test

parameter that evaluates the notch ductility, such as the CV lateral expansion rather than impact energy was therefore thought to be more significant and universal index. An index of ductility can readily be obtained from CV test specimen by measuring the lateral expansion of the specimen at the compression side directly opposite the notch. Measuring devices have been developed for this purpose and are commonly used.

The requirements for the CV impact energy and lateral expansion at a reference temperature TNDT + 60°F was rationalized based on an empirical relationship described by Corten and Sailors [10] and Shoemaker [11]. The data used to develop this empirical correlation was A-533-B steel plate properties that were tested in the mid to late 1960s. It is important to note that data for A-508 forging material, which typically has higher fracture toughness properties compared to A-533 plate, was not included in this correlation. It was assumed at the time that A-508 material would exhibit a similar pattern as A-533. The Corten-Sailors correlation is:

( ) 375.087.15 vId CK =

where CV is the absorbed impact energy (ft-lb), and KId is the dynamic stress intensity factor (ksi-in½). Corten and Sailors noted that this empirical relationship was valid for impact properties on the lower shelf and for impact properties in the temperature transition region. They recognized that the relationship was not valid for impact properties on the upper shelf. Furthermore, at TNDT + 60ºF a KId (dynamic fracture toughness) value of 70 ksi-in½ was reasonable and was a value that fell above the lower-bound reference curve developed from available test data on pressure vessel plate steels. Using the Corten-Sailors correlation, CV = 52 ft-lb for KId =70 ksi-in½.

Shoemaker [11] compared measured values of the CV lateral contraction with calculated values of the dynamic crack opening displacement. Shoemaker proposed the following empirical correlation:

yd

Id

EKσ

δ2

=

where δ is the calculated opening displacement, KId is the dynamic stress intensity factor in ksi-in½, and E is the elastic modulus in ksi, and σyd is the dynamic yield stress in ksi. Shoemaker then found empirically that:

δ20=LC

where LC is the CV lateral contraction. Assuming that the lateral expansion and lateral contraction are approximately equal, the PVRC Ad Hoc Group combined these equations as follows:

( )1000202

⎟⎟⎠

⎞⎜⎜⎝

⎛=

yd

Idyd KE

MLEσ

σ


where MLE is the CV lateral expansion expressed in mils. Assuming KId = 70 ksi-in½ and σyd = 80 ksi then MLE = 41mils.

Based on the Corten-Sailors and Shoemaker correlations the PVRC Ad Hoc Group determined that a specification of 35 MLE with not less than 50 ft-lb was reasonable. In addition to this evaluation, MLE versus absorbed impact energy plots were made with data obtained from qualification of a large number of A-533 B steel plates used in actual reactor vessel fabrication. The data were from plates having yield strengths between 65 to 70 ksi measured at the quarter-thickness plate location. The plot showed that 50 ft-lb corresponded to 40 MLE. A similar plot showing this relationship is shown in Figure 1.

Figure 1: Relationship at 40 Mils Lateral Expansion (MLE)

Between Impact Energy and Yield Strength [4]

TEMPERBEAD WELDING ON FERRITIC MATERIALS The nuclear construction code [1] requires that weldments

on low alloy heavy section vessel steels receive a post weld heat treatment (PWHT) to temper the base material, weld metal, and HAZ and to relieve weld residual stresses. Weld repairs on low alloy heavy section steels present a significant challenge since in many cases it is undesirable or impossible to perform the required PWHT to temper the weld repair area or relieve weld residual stresses. The temperbead welding process is an alternative that provides an acceptable level of toughness in the base material HAZ and weld metal without PWHT.

When weld metal is deposited on low alloy pressure vessel steels having high hardenability, such as SA-533 and SA-508 low alloy steels, the material adjacent to the weld deposit is elevated above the critical transformation temperature. This adjacent material, or heat-affected-zone (HAZ), cools rapidly as the welding arc moves along the surface. This rapid cooling forms a mixture of hardened transformation phases in the HAZ, including martensite and upper bainite. Martensite and upper

bainite are extremely hard and very strong microstructures but have limited ductility without tempering. The high hardness and strength can be attributed to carbon atoms that are unable to diffuse during cooling and become trapped in the martensite and upper bainite; effectively inhibiting material deformation. Tempering is a heating cycle below the critical transformation temperature that provides adequate thermal energy to enable carbon diffusion out of the trap locations and promote more thermodynamically stable metal carbides. This carbide precipitation also relieves some of the internal stress in the hardened phases which permits the atomic lattice to deform and thus increases the ductility while maintaining a significant level of strength associated with the hardened phases. The result is improved fracture toughness.

The optimum microstructure for high fracture toughness is tempered martensite. Consequently, the more martensite that is produced during the first weld layer the higher the toughness potential that can be obtained by a subsequent tempering heat cycle. A good temperbead welding procedure will develop a high volume of martensite in the HAZ during the first weld layer and effectively temper the martensitic during the second or third weld layer.

Temperbead Procedure Qualification The ASME Code governs the qualification of temperbead

welding procedures. The specific rules and test requirements are found in various parts of the ASME Code depending on the component and specific circumstances. ASME Section III, NB 2300 discusses fracture toughness requirements, Section III, NB 4330 discusses general requirements for welding procedure qualifications, Section IX provides specific welding procedure qualification requirements, and Section XI, IWA-4600, plus Nuclear Code Cases, provide additional guidance and rules.

The ideal temperbead qualification would use material archived from the same heat and heat treatment as the component to be repaired. Unfortunately, such archived material is seldom available. The ASME Code permits use of substitute material provided it is representative of the material to be welded. In many cases qualification material would have the same specification, type, and grade selected. It is noted that qualification on material of a different specification is permitted provided the material has the same P-Number and Group Number assigned by ASME Section IX.

Drop weight and CV test methods are used to evaluate the temperbead welding procedure qualification according to a test protocol defined by the ASME Code [1,2]. To qualify a temperbead procedure a weld test coupon is welded according to a fixed welding procedure qualification plan. The geometry of the test plate is controlled by Code and the test plate is heat treated to be representative of the component to be repaired. ASME Section IX or XI provides the test specimen geometry and testing protocol.

The requirement for acceptable qualification of the temperbead welding procedure is based on demonstration that the process does not degrade the fracture toughness of the base


material. A qualification groove weld test is prepared using the temperbead procedure and the completed temperbead weldment is tested according to the specified ASME Code test protocol. The test protocol specifies CV impact testing at TNDT + 60°F, where TNDT is the transition temperature for the base material determined by the drop weight test. This procedure is similar to the requirements described earlier for pressure vessel material qualification except that three additional CV tests are placed in the weld HAZ.

Both the unaffected base metal and HAZ CV tests are performed at the TNDT + 60°F temperature. The average unaffected base metal CV results at this temperature must be equal to or exceed 35 MLE and 50 ft-lb impact energy. The average MLE measured in the HAZ samples must be equivalent or greater than the average MLE of the unaffected base material for successful qualification. If these requirements are not satisfied the ASME Code permits alternative rules; however, these alternatives are not preferred since the minimum specified RTNDT (or lowest permissible service temperature) imposed on the system after temperbead repair may increase the lowest permissible service temperature.

Evaluation of Industry Temperbead Qualifications Documented data from twenty-one Procedure

Qualification Records (PQRs) for temperbead procedures developed in the nuclear industry were reviewed to investigate the typical results when following the current ASME Code temperbead qualification rules. Temperbead PQRs were obtained from Framatome ANP (now AREVA NP), Welding Services Inc., Entergy, and EPRI. One goal of the investigation was to determine the scatter associated with typical temperbead qualification impact test results. Only PQRs with the necessary and essential data were included in the review.

The temperbead welding process is acceptable when the ratio of impact properties in the HAZ to unaffected base material is equal to or greater than one. Since the CV tests for the temperbead qualifications were performed at varying temperatures based on the drop-weight test results of the base material, it is of little value to plot the ratios of impact test results versus temperature. Instead, the impact property ratios are plotted versus the unaffected base metal average % shear at the CV test temperature set by the drop-weight test (TNDT + 60°F). The % shear was selected for comparing the impact property ratios since it generally indicates the relative position on the toughness transition curve; with 0% shear indicating lower shelf behavior, 50% shear about mid way in the transition temperature range, and 100% shear indicating upper shelf behavior. Average test results were used to compute the impact property ratios for each of the 21 different PQRs evaluated. Figure 2 shows the MLE ratios (red square markers) and impact energy ratios (blue diamond markers) versus % shear measured at the applicable TNDT + 60°F temperature for the 21 temperbead PQRs reviewed. One striking result is that the TNDT + 60°F test temperature causes CV tests to be performed over a wide range of ductility conditions from 20%

to 100% shear. This PQR data shows that test temperatures set by the drop weight test are inconsistent in terms of attempting to measure HAZ impact properties at a fixed position along the toughness transition relationship. Figure 2 also illustrates cases where the procedure qualification failed (HAZ to base metal ratio is less that 1.0) where the CV specimen exhibited 100% shear. CV specimens exhibiting 100% shear are measuring material behavior on the upper shelf of the transition curve where scatter is high and there is little chance that improved or degraded HAZ impact properties can be demonstrated.

Another interesting observation is that absorbed impact energy produced higher ratios at all levels of shear compared to MLE ratios. This behavior is not unexpected since absorbed impact energy is a function of both strength and ductility, and both are improved by temperbead welding; whereas the magnitude of MLE measured by the CV test is governed primarily by ductility without consideration for strength.

Figure 2: Charpy V-notch Property Ratios (HAZ to Base Metal)

Versus % Shear for Temperbead Weld Procedures

EVALUATION OF TEMPERBEAD WELDING PROCESS ON LOW ALLOY STEELS

The impact properties of the HAZ and unaffected base metal that occurs with temperbead welding procedures were investigated on heavy section SA-508 Class 2 (Grade 2 Class 1)1, SA-508 Class 3A (Grade 3 Class 2)1, and SA-533 Type B Class 2 pressure vessel low alloy steels. These vessel materials are quenched and tempered carbon-manganese low alloy steels with nominal ½ chromium, ½ molybdenum, and ¾ nickel. ASME Section IX designates these low alloy steels as P No. 3 Group 3 materials.

Temperbead layers were installed using a machine GTAW process with ambient temperature preheat. The consistent layer temperbead technique (CLTT) was used to control the heat of welding and to provide proper tempering of the weld HAZ. The CLTT technique was developed in the early 1990s by EPRI

1 Current SA-508 grade and class designation is noted in parenthesis.


for temperbead weld repair of nuclear grade reactor pressure vessel steels [8].

The objective was to define the transition temperature behavior for the unaffected base material so that it could be compared directly to the temperbead HAZ transition curve. The impact properties of the unaffected base metal and temperbead HAZ were evaluated using single CV tests over a wide range of temperatures. The weld metal was not tested. It is important to note that current ASME Code protocol requires testing three CV samples at the drop-weight TNDT + 60°F temperature for temperbead weld process qualification. The approach for this study was to examine impact toughness behavior over the complete brittle-to-ductile transition temperature range rather than at a single temperature. In this way any bias or confusion resulting from testing at a single test temperature could be assessed.

Vintage SA-508 Class 2 Forging Test Results The first temperbead welding experiment was done on a

heavy section SA-508 Class 2 forging. The drop weight TNDT for this material was 0°F and, as per the ASME test protocol, would be impact tested at 60°F. The ratio of HAZ to unaffected base metal impact energy at 60°F and the ratio of HAZ to base metal MLE at 60°F are plotted in Figure 2 (open symbols at 70% shear) for reference to the industry temperbead welding procedures. Note that these specific ratios are based on single CV specimens at the prescribed TNDT + 60°F test temperature instead of the normal average of three. These results suggest that this temperbead welding process would be qualified if additional samples had been tested to meet Code testing requirements.

The full CV transition curves for the SA-508 Class 2 base material and temperbead HAZ are plotted in Figures 2a and 2b. Trend lines are added in Figure 2a to help display the dramatic shift of the toughness transition curve to lower temperatures. A comparison of the HAZ and unaffected base metal impact properties at roughly 50% shear indicate a shift of approximately 130°F to 150°F for impact energy and about 100°F for MLE. This shift to lower temperatures for both impact energy and MLE suggest significantly improved HAZ impact properties were achieved by the temperbead welding procedure.

Close examination of the base material transition curves in Figures 2a and 2b indicate that the minimum required 50 ft-lb and 35 MLE criteria are achieved at about -50°F, which is significantly lower than the 0°F TNDT temperature measured by the drop weight test. This indicates that the RTNDT determined from the CV transition curves is near -110°F; a value well below an RTNDT of 0°F suggested by the drop-weight test. It is notable that the HAZ exhibits both superior absorbed impact energy and MLE compared to the unaffected base material at the -50°F test temperature established from the CV transition curves. In contrast, the HAZ exhibits less improvement in impact energy and only a marginal improvement in MLE at the 60°F test temperature determined from the drop weight test.

This clearly shows the difficulty created by performing CV testing at the temperature (TNDT + 60°F) established by the drop weight test; especially considering that MLE is the property used for temperbead welding procedure acceptance.

As previously noted, Figures 2a and 2b show that the 50 ft-lb and 35 MLE minimum required properties in both the base material and HAZ are achieved at a temperature much lower than predicted by the drop weight TNDT + 60°F. This implies that the temperature established to index the reference fracture toughness curve found in Appendix G [1] is too high, and an alternate approach is needed.

Figure 2a: Impact Energy Versus Temperature for SA-508

Class 2 Forging Material and Temperbead HAZ

Figure 2b: Charpy V-notch MLE Versus Temperature for SA-

508 Class 2 Forging and Temperbead HAZ


New SA-508 Class 3A and SA-533 Type B Class 2 A second set of temperbead welding experiments were

done on a SA-508 Type 2 Class 3A forging and a SA-533 Type B Class 2 plate manufactured by modern melting practices. The consistent layer temperbead technique was again used for these experiments with the exact same temperbead qualification parameters with the exception that neither material received a 40 hour simulated PWHT. It should be noted that the SA-508 Class 2 material used for the first experiment did receive a 40 hour simulated heat treatment prior to temperbead welding. As was done previously, only one CV sample was tested at each temperature in order to develop a full transition curve with a minimum number of impact tests.

The SA-508 Type 2 Class 3A forging exhibited a drop weight TNDT of -30°F and would be impact tested at +30°F per the ASME protocol. The full CV transition curves for this SA-508 base metal and temperbead HAZ are plotted in Figures 4a and 4b. The SA-533 Type B Class 2 plate exhibited a drop weight TNDT of -18°F and so the prescribed impact testing would be performed at 42°F. The full CV curves for the SA-533 plate and HAZ are plotted in Figures 5a and 5b.

Inspection of Figures 4a, 4b, 5a, and 5b suggest that both the plate and forging materials melted using modern methods have such high toughness that impact testing at the TNDT + 60°F (indicated by the vertical dotted line on each plot) would be on the upper shelf; which means that full shear or nearly full shear behavior would be exhibited for both the base material and HAZ. The scatter and overlap of impact properties in the TNDT + 60°F range show significant overlap and scatter such that it is not possible to easily discern the difference in properties between the unaffected base metal and HAZ. Further, the scatter observed could easily result in failure of the temperbead procedure qualification even though the impact properties are significantly higher than the 35 MLE and 50 ft-lb ASME Code minimum.

DISCUSSION The CV test temperature indexed by the drop-weight test

(TNDT + 60°F) results in a wide range of results as displayed in Figure 2. These data represent 21 commercial PQR tests contributed from industry to assess the results. There was no screening performed to differentiate vintage materials from current melting practices; however there have been many reports of great difficulties finding vintage material and the use of the modern materials has resulted in problems qualifying procedures. The data clearly show that for SA-533 and SA-508 low alloy steels the drop-weight test is not a precise method for setting the CV impact test temperature. Scatter in nil-ductility transition temperatures determined for quenched and tempered low alloy steels by the ASTM E 208 drop weight method is not unexpected. Onodera, et al. [5] investigated the validity of the drop weight test method for A-508 and A-533 nuclear vessel steels. They reported a scatter from -4°F to -67°F for A-508 and from -4°F to -40°F for A-533. Satoh et al. [6] evaluated

the drop weight test for A-508 and observed similar scatter. Both Onodera and Satoh attributed the scatter to variations in how the brittle weld starter bead was applied and to the hardenability behavior of A-508 and A-533. Steps have been taken to minimize these variations, but they still exist to some degree.

Comparison of HAZ and base metal CV impact data acquired at temperatures on the upper shelf region of the transition curve show that the inherent scatter can lead to erroneous negative results. This scatter is readily apparent on MLE versus temperature curves where the MLE reaches a peak threshold. Once this maximum MLE is reached along the temperature axis, higher MLE values are not observed due to 1) the geometry limitation of the small CV specimen and 2) the fixed placement of the anvils in the test fixture that allow the CV sample to be pushed through the testing device when the peak threshold is reached.

The work presented in this paper shows that the drop weight test tends to index the CV impact test temperature to the upper shelf of the transition curve for modern materials with superior fracture toughness. Though not as prominent as compared to new materials, this trend is evident when comparing vintage SA-533 and SA-508 Cv impact data. Figure 7b shows that the impact properties of the vintage SA-508 forge material at temperatures equal to or above the drop weight TNDT + 60°F are much higher compared to the vintage SA-533 plate. This difference indicates that the SA-508 was tested on the upper shelf and the SA-533 was tested in the transition temperature region as desired. As previously discussed, the Corten-Sailors correlation that set the drop weight TNDT + 60°F and Cv test temperature relationship was based on empirical data for A-533 plate material.

The purpose of CV impact testing for base materials, in addition to verification of acceptable impact properties, is to index the material to the ASME Code design reference fracture toughness curve. The objective of temperbead welding qualification is not to index the HAZ fracture toughness, but rather to demonstrate that the temperbead welding process does not degrade the material that is to be repaired. Evidence of degraded HAZ fracture toughness would be seen by a shift of impact properties to higher temperatures when compared to the base material. Testing over a range of temperatures to show a HAZ toughness transition shift is clearly beneficial in demonstrating the effectiveness of the temperbead welding process. For example, the transition temperature shift of 100°F to 150°F for the temperbead HAZ shown in Figures 2a and 2b for SA-508 represents a significant improvement in fracture toughness that is not seen by merely examining the ratios of CV properties at a single temperature. This temperature shift in toughness behavior is best described by a full CV transition curve. Multiple CV tests from the lower to upper shelf temperatures would allow full development of the curve and permit examination of impact properties in the transition temperature region. Data trend curves could then be developed with which to accurately quantify transition temperature shifts.


Figure 4a: Impact Energy Versus Temperature for SA-508

Class 3A Forging Material and Temperbead HAZ

Figure 4b: MLE Versus Temperature for SA-508 Class 3A

Forging Material and Temperbead HAZ

Figure 5a: Impact Energy Versus Temperature for SA-533 Type

B Class 2 Plate Material and Temperbead HAZ

Figure 5b: MLE Versus Temperature for SA-533 Type B Class 2

Plate Material and Temperbead HAZ


The disadvantage of this approach is the large number of CV tests required to develop the full transition curve for both the HAZ and base material.

The current Code qualification procedures for temperbead welding measure the impact properties at a single temperature defined as the TNDT + 60°F temperature and a full toughness transition is not required to be measured for either the base material or HAZ. If the base material used for qualification produced a TNDT + 60°F sufficiently high for the CV testing to be measuring upper shelf properties for the base material, then no improvement in the HAZ transition toughness temperature behavior can be shown. In fact the current Code methodology may even fail a procedure when the properties of the HAZ are much better than the base material, but cannot be shown. An examination of the full CV transition curves for both base material and HAZ eliminates the potential for this dilemma and would be a much better approach; however this approach requires a large number of HAZ samples and base metal samples for evaluation. A better testing protocol is needed.

MLE Versus Absorbed Impact Energy Relationship Figures 6a and 6b display the MLE versus absorbed impact

energy for a large number of heats of pressure vessel low alloy steels (P-No. 3 Group 3 materials). Figure 6a plots impact data for steels produced with modern melting practices and Figure 6b plots vintage materials. The TNDT, as measured by the drop weight test, is listed in parenthesis in the marker legend block. The open round circle marker at 50 ft-lb and 35 MLE is the ASME Code minimum required impact properties. Similar linear relationships between MLE and impact energy for SA-106 Grade B piping and Type 4340 bolt steel were provided in WRC Bulletin No. 175 to help define the Code test requirements for reactor pressure vessel steels [4].

Figures 6a and 6b show that, with the exception of the temperbead HAZ in the SA-508 Class 2 forging (EPRI SA-508 HAZ blue open rectangles), both modern melting practice and vintage materials have a linear MLE-impact energy relationship up to about 80 MLE. The behavior of the EPRI SA-508 HAZ appears to be the exception since this same non-linear relationship is not exhibited by the temperbead HAZ of the other steels in this study. One notable difference for the EPRI SA-508 material is the higher TNDT of 0°F. The other pressure vessel steels that were evaluated with temperbead welding in this study have a much lower TNDT. When considering the MLE-impact energy relationship one would normally expect an increase in ductility to the left of the linear relationship and higher strength to the right. The higher impact energy is a function of the strength but also reflects the ductility that is present. This suggests that if the ductility remains the same or slightly better, higher energy absorption will be developed if the strength increases and still remains ductile. When testing is on the upper shelf with higher hardenability steels (such as is the case for vessel steels made with modern melting practice) this is exactly what happens. The impact strength continues to increase while the MLE plateaus at a maximum. Notice this

MLE plateau or peak ceiling relationship in Figure 6a. Figure 6b shows a similar plot with the same MLE-impact energy linear relationship for vintage low alloy vessel steels. Notice the increased ductility at lower absorbed energy for the vintage materials. This is likely due to lower test temperatures where this result can be seen.

Figure 7a plots the same data as shown in Figure 6a except the data measured below TNDT + 60°F has been removed. It is readily apparent that all the tests at TNDT + 60°F and greater display this same mixed upper shelf behavior where the MLE is relatively constant due to Charpy specimen geometry limitations, but the impact energy continues to increase. One could take the position that this is a good result because the testing requirements ensure tough material for pressure vessels. Unfortunately the purpose of impact testing is to index materials to the design reference toughness curve and upper shelf properties do not accomplish this. The 50 ft-lb and 35 MLE property requirements are the index values. Since the upper shelf values represent higher temperatures the justification has been that it is conservative and that is acceptable. The problem with this approach for qualifying temperbead welding procedures is that the purpose is not to index to the toughness reference curve, but rather to demonstrate that the toughness achieved in the temperbead weld HAZ is equivalent or better than the base material. This requires the ability to discriminate toughness behavior between the base material and the HAZ. The current test temperature indexing protocol to the TNDT + 60°F does not accomplish this goal. An alternate approach is needed.

Finally, one additional plot is provided in Figure 7b showing the behavior of vintage material. This plot, like Figure 6b, removes impact data taken at temperatures below the TNDT + 60°F. Notice that the vintage SA-533 plate (NP-119 heats) data is concentrated around the 50 ft-lb and 35 MLE point. As previously discussed, the material properties of A-533 plate is the basis for the Corten-Sailors relationship that was subsequently used in the WRC Bulletin 175 to specify the minimum 50 ft-lb and 35 MLE minimum requirement for pressure vessel steels. It is also important to note that the vintage SA-508 forging materials (AV heats) plotted in Figure 7b do not follow the same behavior as the SA-533 plate materials (NP-119 heats).

Figure 6b clearly shows that the Cv impact properties of new SA-533 and SA-508 vessel steels, tested at or above the drop weight TNDT + 60°F, are in a region where the MLE plateaus at a maximum value and the impact energy continues to increase. Though not plotted, it is obvious that this region is on the upper shelf of the transition curve. Figure 7b shows a similar pattern for the vintage SA-508 forging material. This indicates that vessel steels with superior fracture toughness, such as new SA-533 plate and new and vintage SA-508 forgings, and their temperbead HAZ, do not follow the drop weight test protocol and pattern established for vintage SA-533 plate material in the 1970s.


Figure 6a: Lateral Expansion Versus Impact Energy for

New Melting Practice Materials (SA-508 and SA-533) Base Material and Temperbead HAZ

Figure 6b: Lateral Expansion Versus Impact Energy for Vintage Plate (SA-533, NP-119 Heats) [9] and Vintage Forging (SA-508,

AV Heats) Materials

Figure 7a: Impact Properties for New Melting Practice Materials

Tested at Temperatures ≥ TNDT + 60°F (Compare to Figure 6a)

Figure 7b: Impact Properties for Vintage Plate (SA-533, NP-119

Heats) [9] and Vintage Forging (SA-508, AV Heats) Materials Tested at Temperatures ≥ TNDT + 60°F

(Compare to Figure 6b)


ALTERNATIVE APPROACH FOR QUALIFICATION OF TEMPERBEAD WELDING

This study provides the technical basis showing the need for an alternative approach for qualifying temperbead welding procedures. The current ASME Code testing protocol that indexes the CV test temperature to the drop weight TNDT + 60°F is not appropriate for materials produced using modern melting practices. The reason is the significantly improved cleanliness, the finer prior austenite grain sizes, and increased hardenability of these new low alloy steels. These properties provide significantly higher toughness and shift the transition to lower temperatures; temperatures and toughness properties that are not properly accounted for in the current temperbead test protocol. The result is that in most cases the CV impact testing is performed on the upper shelf (near 100% shear behavior) where the MLE of the HAZ cannot be effectively distinguished from the base material MLE.

An alternative approach is proposed that eliminates the use of the drop weight test. This alternative also minimizes the number of CV tests required and properly sets the CV test temperature in the transition region. The proposed alternative is a follows: • Multiple CV tests at a range of temperatures are performed

as necessary on the unaffected base material to identify an acceptable mid transition temperature. An acceptable temperature would be in the 50% shear range or between 35 and 50 MLE.

• Three CV tests in the unaffected base material are performed at the selected mid transition temperature. The average of the three CV specimens must be 50 ft-lb or greater and 35 MLE or greater.

• Three CV tests in the temperbead HAZ are performed at the selected mid transition temperature. The temperbead welding procedure is acceptable if the average MLE of the three HAZ CV specimens is equal to or greater than the average MLE of the unaffected base metal CV specimens.

ACKNOWLEDGMENTS This work was funded under EPRI’s Welding and Repair

Technology Center program. Mike Newman of the Welding and Repair Technology Center, Charlotte, North Carolina, USA, fabricated the temperbead specimens used for this project. Charpy testing was performed by Howard McGehee of AMC-Vulcan, Inc., Birmingham, Alabama, USA.

REFERENCES [1] ASME Boiler and Pressure Vessel Code, ASME Section

III, Rules for Construction of Nuclear Facility Components

[2] ASME Boiler and Pressure Vessel Code, ASME Section XI, Rules for Inservice Inspection of Nuclear Power Plant Components

[3] “Fracture Toughness Criteria for Protection Against Failure,” ASME Boiler and Pressure Vessel Code, Section XI, Nonmandatory Appendix G, 2004 Edition.

[4] “PVRC Recommendations on Toughness Requirement for Ferritic Materials,” Welding Research Council, Bulletin No. 175, August 1972

[5] Shinsaku Onodera, Keizo Ohnishi, Hisashi Tsukada, Komei Suzuki, Tadao Iwadate, and Yasuhiko Tanaka, “Effect of Crack-Starter Bead Application on Drop-Weight NDT Temperature,” ASTM STP 919, American Society of Testing and Materials, Philadelphia, 1986, pp. 34-55.

[6] Masanobu Satoh, Tatsuo Funada, and Minoru Tomimatsu, “Evaluation of Valid Nil-Ductility Transition Temperatures for Nuclear Vessel Steels,” ASTM STP 919, American Society of Testing and Materials, Philadelphia, 1986, pp. 16-33.

[7] R.E. Smith and D. J. Ayres, “Statistical Analysis of Charpy-V Impact Properties SA 533 Grade B Class 1 and SA 516 Grade 70 Plate Material,” Transactions of the ASME Journal of Engineering for Industry, February 1973.

[8] David W. Gandy and Shane J. Findlan, “Temperbead Welding Repair of Low Alloy Pressure Vessel Steel: Guidelines,” EPRI Repair Replacement Applications Center, TR-103354, December 1993

[9] Fracture Toughness Data for Ferritic Nuclear Pressure Vessel Material, Final Report. EPRI, Palo Alto, CA: April 1976. NP-119

[10] H. T. Corten and R. H. Sailors, “Relationship Between Material Fracture Toughness Using Fracture Mechanics and Transition Temperature Test,” T. & A. M. Report No. 346, University of Illinois, Urbana, Ill., Aug. 1, 1971

[11] A. K. Shoemaker, “Notch-Ductility Transition of Structural Steels of Various Yield Strengths,” ASME Page No. 71-PVP-19, presented at the First National Congress on Pressure Vessels and Piping, San Francisco, Calif., May 10-12, 1971

Alternative Approach for Qualification of Temperbead ... Approach... · ALTERNATIVE APPROACH FOR QUALIFICATION OF ... to temper the base material, ... Alternative Approach for Qualification

Documents