58 Weld Repairs- Creep Performance

Published on December 2016 | Categories: Documents | Downloads: 86 | Comments: 0 | Views: 540
of 21
Download PDF   Embed   Report

Comments

Content

Creep Performance of Weld Repairs

OMMI (Vol. 1, Issue 3) December 2002

Factors, Defined from Analysis, Contributing to the Creep Performance of Weld Repairs
T H Hyde 1, J A Williams 2 and W Sun 1 School of Mechanical, Materials and Manufacturing Engineering and Management, University of Nottingham, University Park, Nottingham NG7 2RD 2 Consultant, Pine Bank, West Leake Road, East Leake, Loughborough, Leics. LE12 6LJ
1

Professor Tom Hyde is the Head of School of Mechanical, Materials, Manufacturing Engineering and Management and Director of the Rolls Royce UTC in gas turbine transmission systems at the University of Nottingham. He has around 30 years experience in the application of analysis to the high temperature deformation and creep of materials and welds/weld repair and his work has close links with the UK power industry [email protected] Eur Ing Dr Adrian Williams has been an independent consultant since 1992 and, in addition, an Industrial Fellow at the School of Mechanical, Materials, Manufacturing Engineering and Management at the University of Nottingham. He has 27 years experience in the UK power industry, up to 1992, in the fields of high temperature materials, including similar and dissimilar weld performance component testing, life assessment and weld repair. [email protected] Dr Wei Sun is a research fellow in the Structural Integrity and Dynamics Research Group at the University of Nottingham. He have been working in the field of high temperature creep since 1993, with emphasis on the creep testing, material creep properties at elevated temperature and failure assessment of welded components using numerical creep and continuum damage modelling. [email protected]

ABSTRACT Welded components operating within the creep regime at elevated temperatures can accumulate time dependent damage/cracking that can influence the future safe and economic operation of the plant. Following the assessment of this damage, one option could be to weld repair the component. Thus, weld repair can be an important factor necessary to ensure the continued safe operation within the design life. In addition, following a general total life assessment of a plant to examine the feasibility of life extension, repair welding and component replacement can be alternatives to the construction of new plant necessary to meet national and local power requirements.

2
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

This paper is illustrative only, and is intended to summarise the current understanding of weld repair studies and, in particular, the corroboration of actual repair data and numerical studies. This paper considers the general background to weld repair, provides some examples of repair needs and considers the predictions that are made by numerical finite element analysis. It finally identifies areas where additional effort is required. Key words: weld repairs, creep performance, appraisal methods, FE analysis 1. INTRODUCTION

It is necessary to inspect components operating at elevated temperatures in the creep regime to ensure that any defects/cracking that occur will not influence the future safe operation of the plant. Any damage that is discovered will require assessment and will result in a series of possible options, namely:• • • • Ignore or do nothing Re-inspect following a specific strategy, Repair, Replace the component with an identical or an improved design incorporating geometry and/or material changes.

Such decisions are necessary to comply with essential safety requirements and also to optimise the economics of plant operation. Furthermore, although damage may be time dependent and can occur within the original design life of the plant, it can be more important if life extension, in excess of the original design life, is considered for economic reasons. This can be a feasible alternative to the construction of new plant necessary to meet national and local power requirements. This paper is for illustrative purposes only and is intended to summarise current understanding of weld repair studies, in particular, the corroboration of actual repair data and numerical studies. This paper considers the general background to weld repair, provides some examples of repair needs and considers the predictions made by numerical finite element analysis. It finally identifies some areas where additional effort is required. 2. HISTORY AND THE NEED FOR WELD REPAIR

The need for weld repair in high temperature plant components arises from a number of causes, as summarised in Table 1.
Type of defect Fabrication defects Fabrication defects not found on initial inspection Damage occurring during service Operating time of material 0 h. The time to the first major inspection, typically 20-30,000 h. Generally in excess of 50,000 h.

Table 1 Typical classes of defects requiring weld repair.

3
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

2.1

Examples of extensive studies involving actual weld repair data

There are very few extensive studies of weld repair needs, based on actual repair data, and for completeness, a number of examples are given below. 2.1.1 Performance of dissimilar metal welds The failure of tube and pipe size dissimilar metal welds between ferritic and typically Type 300 austenitic material, using Type 300 or high Ni filler metals, typified by Inconel 82, 182, and Incoweld A filler metals, has been extensively studied by many workers, [1,2]. Price, [1], and Bagnall, Rowberry and Williams, [2], examined the failure numbers in UK power plant up to 1981 on both a national and, for tube weld data, a specific station basis. The improved performance of joints using a high Ni filler rather than a Type 300 filler was illustrated, Figure 1. Application of the Weibull statistical approaches allowed estimates of the failure incidence to be made for specific plant on an unit by unit basis, [1].

Figure 1 Failure statistics for tube size austenitic:ferritic dissimilar welds using Type 300 or high Ni filler metals, [1].

2.1.2 Failure and weld repair in CrMoV welds fabricated with a 2.25Cr1Mo filler metal The main cracking forms expected in ferritic high temperature weldments are summarised in Table 2.

4
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

Crack type HAZ stress relief cracking

Position of damage Circumferential cracks in the coarse grained low ductility region in the HAZ Transverse weld metal Transverse cracks on the weld cracking in 2.25Cr1Mo weld metal side adjacent to the weld metal, [3]. metal interface. Type IIIa cracking in Circumferential cracking in CrMoV:2.25Cr1Mo:CrMoV the carbon depleted region at welds [4]. the weld metal:HAZ interface. Type IV cracking, [5]. Circumferential cracking in the weaker fine grained or intercritical region of the HAZ. Final failure mode Short axial cracking in the low ductility HAZ and weld metal regions.

Driving force for cracking Residual stress during PWHT and operational stresses during service. Residual stresses, service stresses and low ductility structures. System stresses in axial direction and weak zones due to Carbon transfer. System stresses in axial direction and weak zones.

Pressure stresses.

Table 2 Typical cracking modes found in ferritic, undermatched filler, welds exposed to creep loading. 2.1.2.1 Transverse weld metal cracking and stress relief cracking. Around 1978, a number of instances of transverse weld metal cracking, TWMC, were found in 0.5CrMoV material welded with 2.25Cr1Mo weld metal in UK power plant. The cause of the cracking was a reduced ductility in the weld metal, partly influenced by composition, and the lack of an adequate post weld heat treatment, PWHT, during fabrication. In general, the PWHT was carried out at 600-650o C rather than the expected 700o C, [3]. The influence of the residual stress and the reduced weld metal ductility led to damage occurring during PWHT and, more usually, during early service under pressure, at times typically up to 20,000 h. The influence of weld thickness is shown in Figure 2, [6], which illustrates two main features for two forms of stress relief cracking, that in the HAZ Weld Repair % and that in the weld metal.
60 50 40 30 20 10 0 Weld repair, %, (HAZ cracks) Weld repair, %. (TWMC). Thickness mm 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100

Figure 2 The weld repair % for HAZ, stress relief, and Transverse weld metal cracking in CrMoV:2.25Cr1Mo:CrMoV welds, as a function of weld thickness in mm,[6].

5
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

The incidence of weld repair due to HAZ stress relief cracking increased with section thickness, Figure 2. This was a direct consequence of two effects, residual stress increasing with thickness and the slightly higher Vanadium levels, (~ 0.3%), used in thicker section or cast CrMoV, but still within the specification at that period. The higher V levels can lead to a reduced HAZ ductility. Figure 2 also shows the incidence of transverse weld metal cracking in the 2.25Cr1Mo weld metal, characterised by the percentage of weld repairs. The repair rate is relatively independent of thickness up to around 70 mm. Thereafter, it is strongly dependent on thickness with ~ 60 % repaired at thicknesses > 70 mm. Again, this is due to residual stress effects and the damage caused by relaxation of these stresses in a lower ductility material. 2.1.2.2 Type IV cracking in CrMoV welds fabricated with a 2.25Cr1Mo filler metal More recently, the incidence of Type IIIa and Type IV cracking in CrMoV:2.25Cr1Mo:CrMoV welds has been reported, [7], Figure 3. These cracks, outlined in Table 2, occur at the interface or within the HAZ respectively and are driven by any cross-weld stresses, or for a pipe butt weld, axial or bending system stresses.

Figure 3 Crack incidence data for CrMoV:2.25Cr1Mo :CrMoV welds, [7]. These cracking forms are a current concern in these alloys and also in the advanced 9-12 % Cr steels. 2.2 Weld repair excavation shape

The weld repair excavation is generally tailored to the type and position of the damage being repaired. The general excavation forms, [8,9], are illustrated in Figure 4 and show a

6
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

full repair and two different partial repair forms, focussed on HAZ damage, as well as a local partial excavation which is also used for weld metal damage. It should be noted that:• • • Excavations should remove ALL of the damaged regions. For a full repair, a section of the root is generally left to minimise any fit up and support problems. Certain excavation types will form new HAZ regions in the service exposed weld metal and can lead to the formation of weaker zones in these regions. This can be due to overaging, of the already service aged weld metal, from the thermal fields generated by the repair.

Virgin weld

Full repair, frw

Partial repair, prw2

Excavation.

Partial repair, prw1

Partial repair, weld metal

Figure 4 Typical excavations used in weld repair, (PowerGen plc, Innogy plc, British Energy plc.), [8]. 2.3 Welding procedure options

The details of the welding procedures can be important. Conventional weld repair procedures will generally be identical to those used for primary construction, using preheat, hydrogen control if necessary, interpass temperature control, a chemically matching filler metal and a full PWHT. More advanced procedures have the additional objective of local control of the microstructure, to minimise the amounts of any undesirable phases, and heat input control to allow some degree of controlled re-heating of deposited beads. They rely on the thermal fields generated by each new weld bead to control any transformation and tempering in existing beads, [10,11]. Hence, the key controls are bead overlap and the heat input of each layer. With such techniques, there is the possibility of omitting the PWHT stage for some alloys as a degree of heat treatment can be obtained by careful control during welding. The current advanced welding procedures are summarised below in Table 3.

7
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

Technique Half bead.

Details Butter and grind half away. Build up three layers with different size rods. 6 layer with controlled heat input. 230 – 290 C PWHT. 6 layer with controlled heat input.

Comments Difficult to control, cannot see bead boundary once 1st. layer is ground away. Debris from grinding. Used for A508 and A533B. No grinding and debris. Steels not in creep range. Automated GTAW. GTAW and SMAW. Tempering with no transformation. Steels in creep range. Use changes in rod size to control heat input. 50% overlap and > 50% refinement. Steels in creep range. Layer 1. 80% refinement. Layer 2, 20% refinement and some tempering. Layer 3. 100 % temper. Steels in creep range.

Referenc e [12]

Temper bead.

[11,13]

Consistent layer temper bead.

[11,14,15 ]

Controlled deposition. Ontario Hydro, CEGB, University if Tennessee. EWI/TWI.

3 layer with heat input control.

[16,17,18 ]

3 layer with heat input control.

[19]

Table 3 A summary of the advanced welding procedures used for weld repair

All of these methods are acceptable under certain circumstances and, indeed, are based on the same basic principles. The main differences are the control of the power levels, welding speed, the overlap of adjacent beads and the number of bead layers. The procedures can be complex and welder training is essential. Furthermore, the procedure must be monitored during use. 3. METHODS OF APPRAISAL

A number of different methods have been suggested for the general appraisal of primary and repair welds. The methods in some cases are generally applied to laboratory studies to identify the main controlling parameters for creep life. Based on these methods and conclusions, assessment methods are generated which can be applied to the components in question. A typical example of an assessment method is the R5 methodology. 3.1 Uniaxial cross-weld specimen tests

Cross weld creep and rupture testing is used to define the possibility of weld failure during long term service, [20]. The weld is placed at the centre of a conventional uniaxial creep specimen, although there is no explicit specimen design within the codes, and is loaded in conventional creep test frames. The test is relatively easy to perform, but the results may be difficult to interpret and are usually compared with the parent material properties, [21].

8
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

This can show a weakening, no difference or a strengthening. Such a procedure, with the stress applied in a direction approximately normal to the weld interfaces, will generally provide data for a form of failure where the weld is subjected to a significant axial stress component. 3.2 Full size component testing

It is possible to incorporate all of the features of welds within full size component testing. Such test vessels are essential for the validation of analysis and assessment methods and provide the final underwriting of our understanding for application to plant. The work is expensive and only provides a limited number of failure points. However, the analysis of strain during the test, and the detection of damage, allow complete monitoring of the vessel and weld behaviour. Typical examples of this type of work are given in the published literature, for example, [22,23,24]. 3.3 Plant statistics

Statistical methods rely on the presence of a number of failures of a specific type from a known series of components. They incorporate the weld features by default, primarily by working with the actual welds or families of welds. More specifically, the method has been applied to the failure of heat exchanger tubes in power plant boiler systems although, less advanced systems have been applied to pipe weld failures. The principles have been illustrated by Price, [1], for dissimilar metal welds and by Davison, [25], for general failures in heat exchanger tubing where there are known families of failures. The method used is to fit the failure pattern, for failures from the same family, to probability functions, such as that due to Weibull. Having generated a good fit to the existing data by calculating a series of characterising parameters, the distribution function can be used to generate a series of future expected failures for that particular power plant. A number of examples are shown by Davison, [25], for a range of different boiler tube failure modes and the agreement is excellent. Similar analyses have been used for studies of repair weld failures but are not in the published literature. Such a procedure, originally developed in the UK in the 1980s, has been in continuous successful use by UK power companies. Such methods rely on the detailed accounting of failures and the identification of the cause of failure which requires a supported strategy of failure recording and examination. The approach has been excellent for units such as heat exchangers which contain large numbers of specific families of potential failures and which may need effort, over a period of time, to correctly identify failure modes. This method of assessment has been primarily used in the UK for weld failure analysis with success. It obviously cannot give predictions for materials where there is no prior experience but, for existing plant, it can have a major input into repair and maintenance strategy.

9
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

3.4

Analysis methods, steady-state and damage

Assessment and analysis methods are usually based on the use of numerical modelling techniques, such as the finite element (FE) analysis methods, [26,27], or through simplified rules methods but based on stress analysis approaches, see R5, [28]. Two main types of material behaviour models are generally used in analyses of creeping structures or, specifically in this case, for repair welds. These are power law creep laws and continuum damage constitutive equations. Continuum damage equations of the following general type [29] 3  σ  S ij m & = A  eq  ε t 2 1 − ω  σ eq
n c ij

(1) (2) (3)

and where

& = ω

Mσ r χ tm (1 + φ )(1 − ω )φ

σr = ασ1 + (1 - α)σeq

can be used, in conjunction with the FE methods, to study welds. σeq and σ1 are the equivalent and maximum principal stresses, respectively, Sij is the deviatoric stress. ω is the damage variable which varies from 0 (no initial damage) to 1 (failure) and A, m, n, M, φ, χ and α are material constants. Generally, such detailed approaches are used to define the influence of specific parameters on creep life. Once understood, sensitivity analysis can generalise the approach. FE codes for such approaches and precise material properties, especially for the HAZ material and weld metal, are not widely available. Thus, in many cases, the results of steady-state analyses are used as a simplified life prediction approach. Steady-state creep solutions, using Norton’s creep law, [22],
c & ij ε =

3 ' −1 A' σ n eq S ij 2

(4)

can be obtained using commercial FE codes. Such simplified stress-based approaches using steady-state stresses have been used for life estimations of welds and repaired welds [30]. Generally, the a values are not available and must be estimated, bounded between 0 and 1 or measured. There are other forms of constitutive laws which can also be used to characterise the full creep curves, including tertiary creep. Typical examples of these are forms employing two or more damage parameters, [e.g. 31] and the THETA approaches [e.g. 32]. Although these have been used for the analysis of specific homogeneous parent metal and weld situations, the published data have not generally been applied to repair welds. For this reason, this paper is limited to the use of steady state conditions with a single variable damage law.

10
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

In general, the early modelling ignored the presence of the HAZ but later analysis, incorporated one or two HAZ zones, [22,23,33], or three, [34], in a linear form as zones parallel to the weld interface, across the component thickness. Very few, if any, analyses have modelled the specific HAZ structure distribution found in practice. 4. GENERAL RESULTS FROM ANALYSIS

There are only limited FE results on the creep analysis of repaired welds. The work described here has focused on ferritic steel circumferential pipe welds, [e.g. 27, 30, 34]. These studies have mainly involved stress and damage analyses and failure prediction, in which some typical factors, such as material mismatch, repair dimensions/profiles, system load etc., and their influence on the predicted stresses and failure times, are discussed. For the case of a full weld repair, which has a similar geometry to that of a virgin weld, it is possible to use analysis results which have been obtained for virgin welds. In all cases, the repair welds are assumed to be post weld heat treated and hence should contain low residual stresses. It is accepted that advances in welding technology and procedure control have allowed the potential use of repair welds in the as-welded state to be considered. However, the magnitudes and distributions of residual stress are not known in detail for these geometries and are not considered here although some data are available, [35]. 4.1 Damage analysis and simplified steady-state analysis

The continuum damage laws define failure as occurring when the damage parameter, w, approaches unity. In practical FE damage modelling of pipe welds, creep failure was assumed to have occurred when the failure damage, ω → 1, was achieved through a significant part of the wall thickness. Detailed descriptions on the FE modelling can be found in [29,36]. The accuracy of life estimates from damage analyses of pipe welds, relies heavily on the accuracy of the characterisation of the material properties, within the various zones, at the nominal operating conditions. However, accurate determination of the material properties for each constituent part of a weld can be difficult, expensive and, in some cases, is not possible, [e.g. 37]. Another drawback of the damage analysis method is that the calculations are time consuming and cannot generally be performed by standard FE packages. In addition, the multi-axial damage criteria are not available for many materials within the operating stress and temperature ranges. Therefore, simpler life prediction methods, such as that based on steady-state stresses, have been widely used. Simple life prediction techniques, based on steady-state analyses, have been shown to be conservative, when compared to full damage analysis, in the general case of welded components [e.g. 36,38]. One of the simplest methods of life prediction is based on the peak rupture stress, [38], σr (=ασ1+(1-α)σeq), which takes into account the influence of stress state on failure in the failure dominant zone, and the appropriate creep rupture material properties. Integrating equation (2) gives:

11
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

 1 + m  (1+ m ) tf =  χ  σr )  M (

1

(5)

One of the main concerns, in the use of such an approach, is the need to determine the a value, which can only be obtained experimentally, although lives can be bracketed using a values of zero and unity, required to define s r, as two extremes. Other multi-axial failure criteria have similar problems. Examples of failure prediction obtained from two approaches, for a number of new, service-aged and repaired CrMoV pipe welds under pressure load [39], are shown in Table 4. These results were obtained for input data at 640o C but the principles should be equally applicable at operational temperatures, and showed that conservative life predictions were obtained from steady-state analyses, compared to the corresponding damage predictions, underestimating the failure life by about 30-40%. However, the failure positions predicted by continuum damage and steady-state analyses were consistent, confirming the validity of the application of steady-state analysis as a conservative, alternative failure prediction approach.
Damage Model tf (hrs) Position Welds New weld Aged weld Full repair Partial repair-i Partial repair-ii 21,018 15,640 10,803 9,892 ---HAZ, near OD HAZ, near OD HAZ, near OD HAZ, near OD ---Steady-State Model tf (hrs) Position 14,795 8,942 7,077 6,467 6,298 HAZ, near OD HAZ, near OD HAZ, near OD HAZ, near OD HAZ, near OD

Table 4 Comparison of failure prediction for a number of CrMoV pipe welds under pressure loading.

4.2

Property balance across weld

Although the geometries of repaired welds in pipes can be significantly different, the general form of the stress distributions within a weldment shows the characteristics of off loading, [40, 41], namely, the redistribution of hoop stress from the weaker to the stronger materials, which, in general, results in a lower stress in the weaker material zones and a higher stress in the stronger material zones. For steady-state stress analyses, the degree by which a material is “stronger” or “weaker” than another structure is defined by the ratio of the secondary creep rates of the relevant materials at a constant stress. A schematic illustration of the hoop stress distributions across a pipe weld for a weaker weld metal and a stronger weld metal is shown in Figures 5(a) and (b), respectively. For the former, the & haz < ε & pm < ε & wm relative secondary creep rates for the parent, weld and HAZ materials are ε
& haz > ε & pm > ε & wm . There is little effect on the axial stress distribution. and for the latter, ε

12
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

As with a normal circumferential weld, the hoop stress distributions within a repaired weld, in a pressurised pipe, usually show significant off-loading, [41]. Therefore, the representative stresses, such as the peak rupture stresses in each of the material zones, which are used for life estimations, are functions of relative material creep properties. Generally, in practice, the repair weld metal will be similar to that which was used for the primary weld fabrication. Hence, in the weld repair state, where both the original parent and weld metal have been exposed to the service conditions of pressure and temperature, the repair weld metal will generally be stronger, in creep, than the exposed parent material and weld metal. However, repair welding will generate additional HAZs, which, in many cases, could be the positions controlling failure. There is a marked effect of the relative creep properties, across the repair weld, on the stresses and damage accumulation in the weld. However, the general effect of the material mismatch on life is difficult to establish, as the life will be influenced by the strain and rupture properties as well as the multi-axiality factor, a . A typical example of throughthickness damage variations for a partial repair CrMoV weld is shown in Figure 6, where the weld was repaired by a stronger weld metal, and the failure was due to the high damage which occurred in the weaker HAZ generated in the un-repaired parent material [e.g. 36]. Ideally, a perfect match between the creep properties in all of the zones would be the objective but this would be impractical to achieve. The analysis does however suggest that, in an internally pressurised pipe weld, ductility is also very important as there will be constraint applied to the weld. Thus, there could be an advantage in using weld metals that are slightly weaker than the parent.

13
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

Hoop stress Stress

a)

& haz < ε & pm < ε & wm ε
HAZ

WM PM PM

Hoop stress Stress

b)

& haz > ε & pm > ε & wm ε

Figure 5 Schematic representation of off-loading in a weld: (a) weak weld metal; (b) strong weld metal.

1.0
t (hour)

0.8 0.6 ω 0.4 0.2 0.0 0

3000 6000 7500 8500 9000 9200 9400 9600 9880

0.2 0.4 0.6 0.8 1 Normalised Distance along the HAZ

Figure 6 Damage variation in and along the HAZ, starting from the outer surface, near the Type IV region at different times for a partially repaired weld, under pressure load, pi = 16.55 MPa.

14
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

4.3

Repair profiles and weld width

The final geometry of a repaired weld depends on the repair profile and dimensions, where the excavation width, for example, for a full repair, Fig. 4, is an important parameter to both ensure the full removal of damaged material and to minimise preparation and welding costs. Results [39, 42] for a number of CrMoV repaired welds have shown that the peak stresses are slightly reduced, (< 10%), by increasing the excavation width within the practical range. Also, there is a limited effect of the repair profiles, i.e. full repair or partial repairs, Fig. 4, on the peak stress and failure life. The results given in Table 4, for example, clearly show the weak dependency on repair profiles, with the order in failure life of t f (full) > tf (partial-i) > tf (partial-ii), suggesting that there may not be any significant benefit in choosing a particular repair profile, with regard to the peak stress. As noted above, similar small effects were also found for a specific set of material properties in the weld [39, 42]. Generally, the “J” type weld preparation, with an interface angle of ~15o, will be used for Class 1 welds intended for internally pressurised pipework. For tube size welds, a “V” preparation is often used and the interface angle for these may be higher. However, for weld repair excavation profiles, Figure 4, it is more usual to have preparation angles of the same order as those of the “J” preparation. In practice therefore, the included angle of the preparation will not exceed 45o unless there is a specific access problem for the welder. Again, although such an effect has not explicitly been studied for the prw type repair geometries, the effect has been studied for conventional welds which will give guidance for the full weld repair form, [43, 44]. Again the effect is relatively small for fixed property differences across the weld. Any effects have been found to be larger, albeit still small, when additional loading is higher and for included angles in excess of 45o. However, in general, it would appear that the effect of weld preparation angle can be ignored unless the additional loading and weld angles are large. 4.4 End (system) load

System loading is another important factor which could significantly affect the failure behaviour of repaired welds [30,36,45] In FE modelling, the system loads were usually simply considered by applying additional constant axial loads or bending moments to the pipe ends [45,46]. The results obtained from a series of CrMoV repaired welds [45] have shown a non-linear increase of the peak rupture stresses with the axial load. In this case, for which a weaker HAZ was used, stresses were found to vary insignificantly across the HAZ with low or moderate end loads. However, the variations are more significant when the axial load is high, and the peak stresses, within the HAZ, occur near the HAZ/PM boundary. An example of the variations of the rupture life with axial loading, for these CrMoV welds obtained from damage analyses [39], is shown in Fig. 7, where the results obtained for the plain pipe, (aged PM), are shown for comparison. This clearly illustrates the main features of life reduction due to the presence of welds, an additional life reduction due to increased axial stress and the more limited life reduction due to geometry factors. An additional feature is that when the axial load is high, where the failure is expected to occur by type IV cracking, the life is relatively independent of geometry and weld metal.

15
Creep Performance of Weld Repairs
1E+5

OMMI (Vol. 1, Issue 3) December 2002

tf (h)

1E+4
New weld Aged weld Full repair Partial-i Pipe (Aged PM)

1E+3 0.2 0.4 0.6 σax/ σmh 0.8 1

Figure 7 Failure life versus σax/σmh, predicted by damage analyses, for the new, aged and repaired CrMoV welds, with pi = 16.55 MPa.

4.5

Position of damage and subsequent failure

Steady-state stresses can only be used to give a prediction of the failure location in the repaired welds, where, in fact, the failure location should be referred to as the position of crack initiation in welds. However, the results of damage analyses can be used to suggest crack growth patterns within repaired welds [47]. High damage usually occurs in the weak material zone near the material boundaries and the position and the time at which the first failure damage ( ω → 1) occurs can be referred to as the position and time of the crack initiation. Examples of the variations of the length of the failure damage zone, a, in and along the HAZ, near the parent material, with time, under pressure loading, for a number of CrMoV welds, are shown in Figure 8. For these welds, it has been shown, that in all cases, the failure damage starts at or near the OD, in the HAZ, and grows inwards along the HAZ. The analysis suggests, for repaired welds in thick wall CrMoV pipes, that typical failure location is at or close to the OD. In the case of relatively low additional axial load, failure is more likely to start at a position near the HAZ:WM interface, while, when system load is sufficiently high, the failure is more likely to start at the type IV zone, adjacent to the HAZ/parent material interface. There are also other cases, where the failure can occur in the HAZ generated within the
1.0 0.8 a/T 0.6 0.4 0.2 0.0 0 5000 10000 15000 20000 25000 t (h) New Aged Fully repaired Partially repaired

Figure 8 Variations of a/T with creep time, obtained for the new, aged, fully

16
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

weld metal by a partial repair [39], which is consistent with practical service evidence [48]. 5. DISCUSSION

Service experience of weld cracking has been published for various material combinations. Examples of these are tube and pipe dissimilar metal weld cracking or failures, [1,2], TWMC, [6], stress relief cracking, [6], Type IIIa and Type IV cracking, [4], and for general weld cracking in pressure vessel steels, [49]. Although specifically dealing with failure or the detection of cracking, these data can be, or are, related to weld repairs. However, these publications are few and generally deal with a specific failure form in a limited number of plant. Furthermore, the mechanism by which the cracking occurs is generally understood but information is lacking on specific operating data; in particular, system load, its form and magnitude, etc. Other surveys, generally linked to large research contracts, have attempted to survey experience on a wider level. Examples are reported by EPRI, [50], Storesund, [51], Concari et al, [52], and, whereas they are useful accumulated experience, there is generally insufficient information to provide a comprehensive view of the problem. The reasons for this are that such data are expensive to collect, are often classified as confidential by the donor company and some operational data may not be available. Thus, there are reasons why a complete comparison between laboratory data, analysis and experience is difficult to carry out. Hence, current understanding of weld repair, and the factors influencing it, must, at this stage, be based on qualitative comparison with what is known from modelling, analysis and practice. The behaviour of virgin and repair welds will be directly influenced by the balance of creep properties across the weld, the specific values of these properties and the operating conditions. There are additional differences between virgin and repaired welds. The main difference is that the repair will generally be made in a parent material that has been aged in service and is weaker than in the virgin state. Thus, the repair weld metal, which is virgin, will generally be stronger in creep than the exposed parent material. Secondly, the weld may not have been subjected to a specific PWHT. To summarise, the numerical results suggest that:•
q q

Factors based primarily on geometry There is no major effect of weld width, within practical limits, on the creep performance of repair welds of the “full repair, frw” category. There is no major effect of weld interface angle on the creep performance of repair welds of the “full repair, frw” category as long as the included angle is limited to < 40o. By inference, there is also no effect of these parameters on the creep performance of repair welds of the “partial repair, prw1, prw2” categories.

q

17
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

However, for partial repairs of the prw2 category and where the excavation is solely within the weld metal, the HAZ formed in the service exposed weld metal can show weakness and can be a weak link, initiating damage and cracking in both full size pressure vessel tests and under service conditions, [48]. The analysis of such a model is in general agreement with this result. These predictions are not at variance with the existing operational data. For example, there has been no identified effect of weld width and interface angle from plant. Indeed, most interface angles, for the J type preparation generally used for pressure pipe welds, are generally 15o or less and most welds have a width which is about equal to the weld thickness. In addition, the use of the narrow gap weld design has not led to an increased number of failures. Whilst true for the similar welds considered here, it has been suggested that there are interface angle effects for austenitic:ferritic dissimilar welds, where there are differences in the coefficients of thermal expansion. • Factors based on operational conditions such as temperature and pressure and applied system stress

For any creep-controlled process, there should be a direct correlation between failure rate and temperature, generally controlled by an exponential relationship of the activation energy. Although, this can generally be shown from studies of boiler tube failure statistics, data from plant are not available for heavy section repairs. Numerical analysis has suggested that system stress has a direct influence on the life of virgin and repair welds, influencing both the failure life and the position of failure. It is known that Type IV cracking is directly influenced by axial components of system stress. However, it is generally difficult to explicitly identify the magnitude of the stress, although it will generally be less than the value of the hoop stress. Numerical analysis further suggests that the effect of increasing axial stress is non linear and that the position of initial damage changes from the HAZ adjacent to the weld metal, at lower axial stresses, to being adjacent to the parent material, namely the Type IV position, at high axial stresses. Indeed, the dominant stress becomes axial at higher system loads whereas it can be the hoop stress at low axial system loads. This again is not at variance with the practical data but unfortunately cannot be verified explicitly. A similar case must exist with Type IIIa cracking although, in this case, there is an additional direct metallurgical effect due to the carbon depletion at the weld metal/HAZ interface, leading to a weaker region and a plane of weakness which can be acted on by the axial system stresses. Thus, it is difficult, with the existing weld repair database from operating plant, to fully confirm and validate the details of these predictions. In addition, there will be a range of parent material and weld metal strengths, dependent on the position within any creep data scatter band and also on the exposure time and temperature, which can complicate any comparison.

18
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

Benchmark experiments, of various forms, can be used to validate the results of such analyses for specific test cases. However, for the case of the examination of service failures, there are often variables, which cannot be directly quantified in detail, although it is known that all the creep properties will fall within the allowed scatter band. For performance evaluation of service failure cases, it will be essential to increase the number of recorded failures within the literature, or which are made accessible for study, together with the operational data which led to the repair. There is a genuine need for evaluation of the performance of repair welds on a wide scale where the details of failure/cracking and operational data and their interpretation can be made available to all parties. Such an approach would have a direct input into the ongoing assessment of the performance of repair welds for the normal high temperature creep resistant materials, the ongoing evaluation of repair welds through analysis and experimental techniques such as cross weld testing and would allow statistical approaches to be used on a wide data base. Such an evaluation, when completed, would have a direct input into design rules for repair welds and a major input into the safety assessment of high temperature welded plant. It is suggested that efforts should be made to initiate such studies on a National and International basis. Indeed, the internet could be an ideal vehicle for contact between operators, transfer of data and the general evaluation of the weld repair performance. It is accepted that such data collection can have difficulties, but there would be definite economic benefits to companies in ensuring that weld repairs were carried out in the most realistic and cost effective manner. In addition, it could act as a realistic vehicle for the validation of experimental and numerical laboratory based approaches. This would allow the improved validation of all assessment methods. The possibility of this should be examined. ACKNOWLEDGEMENTS The authors wish to acknowledge EPSRC, PowerGen Plc, Innogy Plc and British Energy Plc for their financial and technical support of the work, through a collaborative EPSRC/ESR21 grant.
REFERENCES [1] Price A T, 1982, CEGB experience with small diameter dissimilar metal welds in coal fired boilers, Conf. on Joining Dissimilar Metals, Pittsburgh, Pa, USA, EPRI-ASME, USA. [2] Rowberry T R, Bagnall B I and Willliams J A, 1979, Analysis of service experience with large austenitic:ferritic steampipe welds in CEGB Midland region plant, Conf. on Welding and Fabrication in the Nuclear Industry, London, 1979, BNES, London, UK. [3] Wolstenholme A A, 1978, Transverse cracking and creep ductility of 2CrMo weld metals, Conf. on Trends in Steels and Consumables for Welding , London, The Welding Institute, Abingdon, UK. [4] Brett, S J and Smith P A, 1998, Type 3A cracking at 2CrMo welds in CrMoV pipework, Conf. Baltica IV, Helsinki, VTT, Finland. [5] Schuller H J, Hagn L and Woitscheck A, 1974, Cracking in the weld region of shaped components in hot steam lines – Materials investigations, Der Machinenschaden, 47, 1-13.

19
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

[6] Cheetham D, Fidler R, Jagger M and Williams J A, 1977, Relationships between laboratory data and service experience in cracking of CrMoV welds, Conf. on Residual stresses and their effect on welded constructions, London, TWI, Cambridge, UK. [7] Brett S J, 1994, Cracking experience in steam pipework welds in National Power, Conf. on Materials and Welding Technology in Power Plants, Vereinigung der Grosskraftwerksbetreiber (VGB), Essen, Germany, March 1994. [8] Sun W, 1996, Creep of serviced aged welds, PhD Thesis, University of Nottingham, England, UK. [9] Klenk A, Williams J A, Shibli A and Issler S, 2000, Integrity of repair welds in high temperature plant operating under creep and creep:fatigue conditions, 2nd Int. HIDA Conf. on Advances in Defect Assessment in High Temperature Plant, Stuttgart, Germany, Oct. 2000, EEC. [10] Neary C M, 1996, Utility guidelines for controlled deposition repair welding, Conf. on Challenges and solutions in repair welding for power and process plant , 1996, Proc. of a workshop, San Diego, USA, 1996, Welding Research Council Bulletin No. 412, Pressure Vessel Research Council, (PVRC), Welding Research Council, (WRC), New York, USA. [11] Doty WD, 1996, History and need behind the new NBIC rules on weld repair without PWHT, Conf. on Challenges and solutions in repair welding for power and process plant , San Diego, WRC Bulletin 412, 1996. [12] American Society of Mechanical Engineers Boiler and Pressure Vessel Code, Section IX, Division 1, Rules for in-service inspection of nuclear power plant components, ASME, New York, USA, 1978. [13] Gandy D W, Findlan S J and Viswanathan R, 1998, Temperbead welding of P numbers 4 and 5 materials, Conf. on Integrity of high temperature welds , Nottingham, UK, Professional Engineering Publishing, UK. [14] Lundin C D, 1996, Overview of results from PVRC programmes on half bead/temper bead/controlled deposition techniques for the improvement of fabrication and service performance of CrMo steels, Conf. on Challenges and solutions in repair welding for power and process plant , 1996, Proc. of a workshop, San Diego, USA, 1996, Welding Research Council Bulletin No. 412, Pressure Vessel Research Council, (PVRC), Welding Research Council, (WRC), New York, USA. [15] Lundin C, 2001, Controlled deposition welding, Conf. on Repair Welding and Serviceability , San Diego, Ca., USA., Jan. 2001, PVRC, NY, USA. [16] Lau T W, Lau M L and Poon G C, 1996, Development of controlled deposition repair welding procedures at Ontario Hydro, Conf. on Challenges and solutions in repair welding for power and process plant, San Diego, WRC Bulletin 412, 1996. [17] Alberry P J and Jones W K C, 1976, An improved welding technique for HAZ refinement, Welding and Metal Fabrication, 45, 11, 549. [18] Alberry P J and Jones W K C, 1982, A computer model for the prediction of HAZ microstructures in multipass welds, Metals Technol., 9, (10), 419. [19] Friedman L M, 1996, EWI/TWI controlled deposition repair welding procedure for 1CrMo and 2.25Cr1Mo steels, 1996, Conf. on Challenges and solutions in repair welding for power and process plant, San Diego, WRC Bulletin 412. [20] Hyde T H and Tang A, 1998, Creep analysis and life assessment using cross weld specimens, Int. Mats. Rev., 43, 6, 221. [21] Etienne C F and Heerings J H, 1993, Evaluation of the influence of welding on creep resistance, International Institute of Welding, (IIW), Document IX-1725-93, Steel Research, 1994, 65, 5, pp. 187-196. [22] Coleman M C, Parker J D and Walters D, 1985, The behaviour of ferritic weldments in thick section CrMoV pipe at elevated temperatures, Int. J. Press. Vess. & Piping, 18, pp. 277-310. [23] Williams J A, 1992, The validation of procedures for assessing the life of components, Conf. BALTICA 2, Life and Performance of High Temperature Materials and Structures , Helsinki, Finland, Technical Research Centre of Finland, (VTT), Espoo, Finland, Session 2c, Paper 2.

20
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

[24] Van Wortel H, 1999, Effectiveness of repair of in service exposed components by grinding or by welding without PWHT, Conf. on Langzeitverhalten warmfester Stahle und Hochtemperaturwerkstoffe, Dusseldorf, Nov. 1999, VdeH. [25] Davison J K, 1991, Experience in the use of boiler tube failure prediction to aid maintenance planning, Conf. on Boiler tube failures in fossil power plants , San Diego, USA, EPRI, Palo Alto, Ca., USA. [26] Hyde T H, Sun W and Williams J A, Creep analysis of pressurised circumferential pipe weldments − A Review, J. Strain Analysis, 37, 6 (in press). [27] Hyde T H, Becker A A, Sun W and Williams J A, 2002, Review of finite element analysis of repaired welds under creep conditions, 3rd Int. HIDA and INTEGRITY Conf. on Integrity of high Temperature Repair Welds, Oeiras-Lisbon, Portugal, Sept. 2002. European Commission/ISQ/MT, Portugal. [28] R5 (Issue 2): Assessment Procedure for the High Temperature Response of Structures, 2001, Eds. R A Ainsworth, R Hales and P Budden, British Energy Ltd, Gloucester, UK, [29] Hall F R and Hayhurst D R, Continuum damage mechanics modelling of high temperature deformation and failure in a pipe weldment, Proc. R. Soc. London, A443, 1991, pp. 383-403. [30] Samuelson L A, Segle P and Storesund J, Life assessment of repaired welds in high temperature applications, Welding & Repair Technology for Fossil Power Plants , EPRI Int. Conf., Virginia, 1994. [31] Perrin I J and Hayhurst D R, A method for transformation of creep constitutive equations, Int. J. Pres. Ves. & Piping, 68, 1996, pp. 299-309. [32] Evans R W and Wilshire B, 1985, Creep of Metals and Alloys, London, Institute of Metals, London. [33] Hayhurst D R and Perrin I J, CDM analysis of creep rupture in weldments, Proc. of Engineering Mechanics, 1995, 1, pp. 393-396. [34] Klenk A, Schemmel J and Maile K, 2002, Numerical modelling of ferritic welds and repair welds in P22 and P91 steel, 3rd Int. HIDA and INTEGRITY Conf. on Integrity of High Temperature Repair Welds, Oeiras-Lisbon, Portugal, Sept. 2002. European Commission/ISQ/MT, Portugal. [35] Dong P, 2001, Residual stresses in repair welds and local PWHT effects, Conf. on Repair Welding and Serviceability, San Diego, USA, Jan. 2001, Welding Research Council, New York, USA, pp. 145 – 169. [36] Sun W, Hyde T H, Becker A A and J A Williams, Comparison of the creep and damage failure prediction of the new, service-aged and repaired thick-walled circumferential CrMoV pipe welds using material properties at 640o C, Int. J. Pres. Ves. & Piping, Vol. 77, 2000, pp. 389-398. [37] Hyde T H and Sun W, Determining high temperature properties of weld materials, JSME Int. J. Solid Mechanics & Material Engineering, Series A, Vol. 43, No. 4, 2000, pp. 408-414. [38] Hyde T H, Sun W, Becker A A, Assessment of the use of finite element creep steady-state stresses for predicting the creep life of welded pipes, Proc. 4th Int. Conf. on Computational Structures Technology, Civil-Comp Press, Edinburgh, Ed. B. H. V. Topping, pp. 247-251, 1998. [39] Sun W, Hyde T H, Becker A A and Williams J A, A study of weld repairs in a CrMoV serviceexposed pipe weld, Proc. 9th Int. Conf. on Creep & Fracture of Engineering Materials & Structures, Swansea, April 2001, pp. 613-622. [40] Walters D J and Cockcroft R D M, A stress analysis and failure criteria for high temperature butt welds, Colloquium on Creep Behaviour of Welds in Boilers. Pressure Vessels and Piping, International Institute of Welding, (IIW), Toronto, 1972. [41] Walters D J, The stress analysis cylindrical butt welds under creep conditions. 1976, CEGB, Note RD/B/N3716, CEGB, London, UK. [42] Hyde T H, Sun W and Becker A A, Life assessment of weld repairs in 1/2Cr1/2Mo1/4V:2 1/4Cr1Mo main steam pipes using the finite element method, J. Strain Analysis, Vol. 35, No. 5, 2000, pp. 359-372.

21
Creep Performance of Weld Repairs OMMI (Vol. 1, Issue 3) December 2002

[43] Fidler R, The effect of weld width on the performance of narrow gap welds in creep conditions, CEGB Report TPRD/M/1583/R86, 1986. [44] Hyde T H, Sun W, Becker A A and Williams J A, Effect of weld angle and axial load on the creep failure behaviour of an internally pressurised thick walled CrMoV pipe weld, Int. J. Press. Vess. & Piping, 78, 2001, pp. 365-372. [45] Hyde T H, Sun W, Becker A A, Effects of end loading on the creep failure behaviour of CrMoV welds in main steam pipelines, Proc. 6th Int. Conf. on Damage and Fracture Mechanics, Ed. A P S Selvadurai and C A Brebbia, pp. 415-424, WIT Press. May 2000, Montreal, Canada. [46] Hyde T H and Sun W, Effect of bending load on the failure behaviour of pressurised thick-walled CrMoV pipe weldment, Int. J. Press. Vess. & Piping, 2002, 79, pp. 331-339. [47] Hyde T H, Sun W and Becker A A, A damage mechanics approach to predicting initiation and growth of type IV cracks in CrMoV weldments, 2nd Int. ‘HIDA’ Conf. on Advances in Defect Assessment in High Temperature Plant, MPA Stuttgart, Germany, October 2000. Also in Int. J. Pres. Ves. & Piping, 78, No. 11-12, 2001, pp. 765-771. [48] Hickey J J, Bernard P J, Bissell A M and Jirinec M J, Investigation and repair of a failed seam welded reheater outlet header, 2nd Int. EPRI Conf. on Welding and Repair Technology of Power Plants, Florida, May 1996. [49] Darlaston J, 1999, Seminar on Boiler shell weld repair Sizewell A NPS , London, Prof. Eng. Publications, UK [50] Gandy D W, Findlan S J and Viswanathan R, 1999, Weld repair of steam turbine casings and piping-an industrial survey, PVP Vol. 388, Fracture, design, analysis of pressure vessels, piping components and fitness for service. [51] Storesund J, Samuelson L E and Klasen B, 2002, Creep life assessment of pipe girth weld repairs with recommendations, 3rd Int. HIDA and INTEGRITY Conf. on Integrity of High Temperature Repair Welds, Oeiras-Lisbon, Portugal, Sept. 2002. European Commission/ISQ/MT, Portugal. [52] Concari S, Fedelli G, Pinto M and Willliams J A, 2002, The collation of in-house data on weld repair of high temperature components, 3rd Int. HIDA and INTEGRITY Conf. on Integrity of High Temperature Repair Welds, Oeiras-Lisbon, Portugal, Sept. 2002. European Commission/ISQ/MT, Portugal.

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close