Nuclear Power Control, Reliability and Human Factors Part 8 docx

The illustration diagram for estimation of crack instability point for J/T method 3.2 Dynamic J-R curve testing for long crack extension To obtain the effective J-R curve under the cond

Evaluation of Dynamic J-R Curve

for Leak Before Break Design of Nuclear Reactor Coolant Piping System 199 conventional fitting method for tearing modulus curve However, analytical approach has uncertainty basically by fitting In this paper, to evaluate reliable Tmat curve at long crack extension region experimentally, we have researched the method for measurement of dynamic J-R curve with crack extension as long as possible

Fig 9 Graphical illustration of J/T method

Fig 10 The illustration diagram for estimation of crack instability point for J/T method

3.2 Dynamic J-R curve testing for long crack extension

To obtain the effective J-R curve under the condition of long crack extension, two specimens were used where one is for short crack extension and the other is for long crack extension

By using two test data, the dynamic J-R curve was evaluated over the crack extension length range according to ASTM code Table 1 shows test matrix for reactor coolant piping base metal for Shin-Wolsung

Item Material (Inner Dia.) Pipe size

Number of test Short crack

extension

Long crack extension Main Loop

Piping

Table 4 Dynamic J-R test conditions for short and long crack extension conditions

The load - displacement curve for each piping material is shown in Fig 11 In the dynamic

J-R curves obtained by normalization method, for hot leg pipe and elbow materials, dynamic J-R curves were similar regardless of crack extension length; whereas for cold leg piping material, J-R curve for short crack extension length was lower than that for long crack extension length as shown in Fig.12 To analyze the reason for the difference between short and long crack extension for cold leg pipe, normalized load-displacement curve is described

in Fig 13 Normalized load-displacement curve, PN - ν’pl curve shows different shape between two tests with different crack extension length In general, normalized load – displacement curve should maintain a constant shape regardless of crack extension size Therefore, optimal normalized PN - ν’pl curve should be calculated by considering both PNi - ν’pli data pair for short and long crack extension

Load Line Displacement (mm)

Short Crack Extension Long Crack Extension Hot Leg Pipe

0 10 20 30 40 50 60

Load Line Displacement (mm)

Cold Leg Pipe

Short Crack Extension Long Crack Extension

0 10 20 30 40 50 60

Evaluation of Dynamic J-R Curve

for Leak Before Break Design of Nuclear Reactor Coolant Piping System 201

2 )

Crack Extension Length (mm)

Cold Leg Pipe Short Crack Extension Long Crack Extension

0 500 1000 1500

Fig 12 The comparison of dynamic J-R curve by normalization method between the tests for short and long crack extension

0.00 0.05 0.10 0.15 0.20 100

150 200 250 300 350

Long Crack Extension

Short Crack Extension

Fig 13 Normalized load, displacement data pair and its each fitting curve for short and

long crack extension of cold leg piping material

3.3 Combined analysis

Based on this concept, combined analysis is proposed as the evaluation method of J-R curve

to long crack extension using the test results with two different crack extensions The procedure is as follows; At first, the PNi - ν’pli data pair is obtained by using load – load line displacement curve for long crack extension length in accordance with Eqs.(9) and (10), and final PNi - ν’pli data pair is obtained for two specimens respectively, where final PNi - ν’pli

values are

pl

f Ni

f

PFinal P

W aWBW

<0.001 are excluded from effective PNi - ν’pli data pair The coefficients of the fitting function

of Eq.(11) instead of Eq.(6) are calculated for two final PNi - ν’pli values and the effective PNi - ν’pli data pair

by checking with slightly increasing crack lengths from initial crack length a0, where load - displacement curve for long crack extension length is used However, J-R curve obtained using combined analysis was deviated from individual J-R curve for short and long crack extension respectively in the case of hot leg pipe material as shown in Fig 14 This reason is

Evaluation of Dynamic J-R Curve

for Leak Before Break Design of Nuclear Reactor Coolant Piping System 203 that load - displacement curve between short and long crack extension have slightly different shape as shown in Fig 11 Therefore, it is needed to adjust the position of middle point by reflecting the characteristics of J-R curves for short and long crack extension To do

so, the coincidence level is evaluated by comparing the J-R curves between normalization analysis by only short crack extension and combined analysis As a method of evaluation for coincidence, best fit curve of Eq.(14) for the J-R curve of short crack extension is used

 m

0.00 0.05 0.10 0.15 0.20 100

150 200 250 300

Hot Leg Pipe Short Crack Extension Long Crack Extension Initial Combined Analysis

Crack Extension Length (mm)

Fig 15 Dynamic J-R curve for hot leg pipe material prior to adjustment of middle point on normalized load versus displacement curve in combined analysis

Next, the standard deviation σ of Eq.(15) is calculated from J value by combined analysis and J value obtained by J-R curve of Eq.(14) Such that, the data of combined analysis to short crack extension are used in calculating σ

where Jfit is J value obtained by fitting function of Eq.(14) Jcombined is J value obtained by combined analysis and n is the number of effective J-R data to short crack extension Optimal middle point on the normalized load-displacement relationship is determined as

a point when standard deviation σ value of Eq.(15) is reached to minimize by adjusting PN

value at ν’pl value at final point of short crack extension Using the optimal middle point, final PNi - ν’pli data pair of long crack extension and effective PNi - ν’pli data pairs, J-R curve can be estimated Figure 9 shows the comparison of dynamic J-R curve among the combined method and normalization method of short and long crack extension For all three kinds of piping, dynamic J-R curve by combined analysis is well described with the behavior of that for two different crack extensions From this combined analysis, we could obtain reasonable dynamic J-R curve until long crack extension for nuclear piping materials In combined analysis, one J-R curve is obtained using two specimens Therefore, the scatter of material properties with the position of taking specimen is required not to be large In LBB analysis, the lowest material property is used among three test results for material property scatter In this approach, the J-R curve tends to be estimated as an average J-R data for two test results Further investigation is therefore needed for low bound curve of J-R curve with long crack extension effectively based on the statistical concept

Crack Extension Length (mm)

Hot Leg Pipe Short Crack Extension Long Crack Extension Combined Analysis

Crack Extension Length (mm)

0 500 1000 1500

Fig 16 The dynamic J-R curve by combined analysis for each material

Evaluation of Dynamic J-R Curve

for Leak Before Break Design of Nuclear Reactor Coolant Piping System 205

4 Conclusion

From the comparison test results between DCPD and normalization method as a dynamic

J-R curve testing method, short crack extension, dynamic J-J-R curves were similar but, for long crack extension, J-R curve estimated by normalization was higher by 10~30% at the initial loading stage than that by DCPD For reliable J/T analysis for LBB design of nuclear piping, material J-R curve for long crack extension is needed However, normalization method is applicable for only short crack extension To overcome this problem, combined analysis based on normalized method was proposed In combined analysis, dynamic J-R curve with long crack extension is estimated by two dynamic J-R curve tests with different crack extension length The dynamic J-R curve beyond the crack extension length range designated by ASTM code could be estimated using the combined analysis

5 References

ASTM (2009) ASTM E1820-09e1 Standard Test Method for Measurement of Fracture

Toughness, In: Annual Book of ASTM Standard, Vol 03.01, ASTM International, West

Conshohocken, Pennsylvania, USA

Ernst, H.A., Paris, P.C., Rowssow, M & Hutchinson, J.W (1979) Analysis of Load

Displacement Relationship to Determine J-R Curve and Tearing Instability Material

Properties In: ASTM STP 677 Fracture Mechanics, Smith, C.W (Ed.), pp 581-599,

ASTM International, ISBN EB 978-0-8031-4746-1, West Conshohocken, Pennsylvania, USA

Ernst, H.A., Paris, P.C & Landes, J.D (1981) Estimations on J-integral and Tearing Modulus

T from a Single Specimen Test Record In: ASTM STP 743 Fracture Mechanics,

Roberts, R (Ed.), pp 476-502, ASTM International, ISBN EB 978-0-8031-4809-3, West Conshohocken, Pennsylvania, USA

Hackett, E.M., Kirk, M.T & Hays, R.A (1986) NUREG/CR-4550 : An Evaluation of J-R Curve

Testing of Nuclear Piping Materials Using the Direct Current Potential Drop Technique, U.S Nuclear Regulatory Commission

Johnson, H.H (1965) Calibrating the Electric Potential Method for Studying Slow Crack

Growth Materials Research and Standards, (September 1965), Vol.5, No.9, pp

442-445, ISSN 0025-5394

Joyce, J.A (1996) Manual on Elastic-Plastic Fracture Laboratory Test Procedures, ASTM

International, ISBN 0-8031-2069-9, West Conshohocken, Pennsylvania, USA

Kim, J.W & Kim, I.S (1997) Investigation of Dynamic Strain Aging on SA106-Gr.C Piping

Steel Nuclear Engineering and Design, Vol 172, No 1-2, (July 1997), pp 49-59, ISSN

0029-5493

Scott, P.M., Olson, R.J & Wilkowski, G.M (2002) NUREG/CR-6765: Development of Technical

Basis for Leak-Before-Break Evaluation Procedures, U.S Nuclear Regulatory Commission

Landow, M.P & Marschall, C.W (1991) Experience in Using Direct Current Electric

Potential to Monitor Crack Growth in Ductile Metals, In: ASTM STP 1114

Elastic-Plastic Fracture Test Methods, Joyce, J.A (Ed.), pp 163-177, ASTM International, ISBN-EB 978-0-8031-5172-7, West Conshohocken, Pennsylvania, USA

Landes, J.D., Zhou, Z., Lee, K & Herrera.,R (1991) Normalization Method for Developing

J-R Curve with the LMN Function Journal of Testing and Evaluation, Vol 19, No 4,

(July 1991), pp 305-311, ISSN 0090-3973

Lee, B.S., Yoon, J.H., Oh, Y.J., Kuk, I.H & Hong, J.H (1999) Static and Dynamic J-R Fracture

Characteristics of Ferritic Steels for RCS Piping, 15th International Conference on

Structural Mechanics in Reactor Technology, Vol V, pp 297-302, ISBN 89-88819-05-5

94500, Seoul, Korea, August 1999

Lee, J.B & Choi, Y.H (1999) Application of LBB to High Energy Pipings of a Pressurized

Water Reactor in Korea, Nuclear Engineering and Design, Vol.190, No.1-2, (June

1999), pp.191~195, ISSN 0029-5493

Nakamura, T., Shih, C.F & Freund, L.B (1986) Analysis of a Dynamically Loaded

Three-Point-Bend Ductile Fracture Specimen, Engineering Fracture Mechanics, Vol 25, No

3, pp 323-339, ISSN 0013-7944

Oh, Y.J, Kim, J.H & Hwang, I.S (2002) Dynamic Loading Fracture Tests of Ferritic Steel

Using Direct Current Potential Drop Method Journal of Testing and Evaluation, Vol

30, No 3, (May 2002), pp 221-227, ISSN 0090-3973

Sharobeam, M.H & Landes, J.D (1991) The Separation Criterion and Methodology in

Ductile Fracture Mechanics International Journal of Fracture, Vol 47, No.2, (January

1991), pp 81-104, ISSN 0376-9429

Wallen, K (2009) Extrapolation of Tearing Resistance Curves 2009 Proceeding of the ASME

Pressure Vessel and Piping Conference, Vol.3, pp 281-286, ISBN 978-0-7918-4366-6, Prague, Czech Republic, July 2009

Thermal stratification phenomenon results from a temperature differential across the pipe cross section with the top fluid stream hot and bottom stream relatively cold During normal plant operations at low flow conditions, when the feed water nozzle is not completely full, hot water from the steam generator remains in the nozzle to fill up the rest of the volume The difference in buoyancy between the hot and cold fluids inhibits their mixing so that the feed water becomes and remains thermally stratified Separation of these two flow regions is due to the density difference in the hot and cold streams The stratified temperature conditions can produce very high stresses, and can occur may times during normal low power operations; therefore this has the potential to initiate cracks in a relatively short period of time Thermal striping is a local phenomenon that occurs at the interface between hot and cold flowing fluids The interface level oscillates with periods ranging from 0.1 to 10 seconds The oscillating fluid temperature gives rise to fluctuating stresses The magnitudes

of the striping stresses are not as high as those due to stratification itself, but the number of cycles is so large that they contribute significantly to fatigue crack initiation

During normal plant operation, a series of temperature measurements has been taken around the pipe circumference at the vicinity of the of the feed water nozzle/pipe weld Analysis of the data indicates that the stratified temperature distributions may be grouped into a handful of basic profiles corresponding to different levels of the interface between the hot and cold fluids For analysis purposes these profiles could be assumed to be at steady state conditions because of their long duration observed during the tests Nuclear piping systems (Class 1) are designed according to the rules of NB 3600 of the ASME Boiler and

Pressure Vessel Code, Section III The loads producing the stresses originate from the internal pressure, mechanical loads due to deadweight, seismic and thermal expansion and the operating thermal transients Normally piping systems are not designed for circumferential temperature variation The effect of the thermal stratification on the state of stress in the pipe is manifested in two ways: (a) the difference in temperature between the top and bottom of the pipe causes greater thermal expansion at the top tending to bow the pipe When such bowing is restrained global bending stresses result; (b) the interface between the two fluid layers causes a local stress in the pipe due to thermal discontinuity across the pipe section The fatigue damage produced by thermal stratification and the associated thermal striping are a good indication of the contribution of these phenomena to the observed feed water line cracking

A detailed finite element stress analysis has been carried out using a three dimensional model that includes the steam generator shell, the feed water nozzle, and the elbow/pipe The shell nozzle/elbow model contains three distinct regions with different heat transfer characteristics between the metal and the adjacent fluid Each of the stratification profiles produces a complex state of stress throughout the nozzle and the elbow (pipe) Different levels of interface produce peak stresses at different locations around the circumference Since the interface level varies during low flow operating conditions, each point in the counter bore area is subjected to a state of varying stresses of large magnitudes A maximum range of stress intensity analysis was carried out prior to fatigue evaluation to determine whether the simplified elastic plastic analysis procedure would be required, and if so, to calculate the plastic intensification factor Ke by which the peak alternating stresses would be multiplied The analysis predicted crack locations that that correlated well with the observed cracking

The major cause of growth of the cracks is due to the thermal stratification cycles, which occur during low flows, primarily at hot standby The thermal striping phenomenon or the oscillations occurring at the interface between hot and cold fluids has some influence on the crack growth, but it certainly impacts the crack initiation predictions Thermal stratification causes a stress distribution in a pipe that is similar to what happens in a bimetallic strip In the hot upper region compressive stresses develop as a result of constrained expansion, with the tensile stresses occurring in the lower region This has been demonstrated using a simplified 2-dimensional finite element model These are essentially the membrane stresses

in the axial direction Since the piping is flexible, the thermal moment gives rise to a bending stress that is added to the membrane stresses to obtain the total stresses

It is suggested that the equations for obtaining stresses in piping systems as outlined in the ASME Code contain a term addressing circumferential temperature gradients in the pipe A number of remedial measures have been implemented or suggested in operating power plants to minimize the stress amplitudes and frequency of load cycling during the stratification events

In recent years, thermal stratification phenomenon has been observed to exist on several piping systems in pressurized water reactors Damages have been observed in the main feed water lines, pressurizer spray lines, unisolable branch piping connected to reactor coolant piping, and pressurizer surge lines, with evidence linked to thermal stratification The stratification phenomenon results from a temperature differential across the pipe cross-section with the top fluid stream hot and the bottom stream relatively cold This condition occurs under relatively low flow conditions by cold feed injection into a stagnant hot pipe region or vice versa Separation of two fluid flow areas is due to density differences in the

Feed Water Line Cracking in Pressurized Water Reactor Plants 209 hot and cold streams This gives rise to gross thermal bending moments across the pipe section resulting in bowing deformation of the pipe

In May 1979, a pressurized water reactor plant in operation approximately a year developed

a through-wall crack in one of its feed water lines at the entrance to the steam generator Subsequent investigation of the remaining lines revealed cracking in the same vicinity but limited to partial wall penetration As a result of this incident, the United States Nuclear Regulatory Commission submitted a directive to all PWR operating plants to perform inspection of their feed water lines A number of plants produced same degree of cracking

in the same general area with wide variety of size, orientation and length of plant operation Because of the involvement of many variables, it was impossible to immediately identify the specific mechanisms of crack initiation and growth A number of activities were initiated to investigate the structural, thermal, hydraulic, operational and environmental conditions which individually or collectively contributed to the observed cracking

2 Observed crack locations

Figure 1 illustrates the feed water pipe to steam generator nozzle junction where majority of cracking occurred Cracks were found to be oriented circumferentially and located in the base metal outside the heat affected zone There were intermittent pitting throughout the inside surface The deepest cracks were found at the base of the counter-bore transition

Fig 1 Location of Cracks in PWR Feed water Pipe to Nozzle Attachment Region [1]

Typically a majority of PWR plants produced the circumferential cracking, the pattern of depth orientation varied considerably for different plants Generally the deepest cracking was observed at the top, although in a number of plants this was found to occur at the sides,

as well as the bottom With the exception of one through-wall condition, most plants produced relatively small shallow cracks

3 Metallurgical studies

The metallurgical investigations revealed that although corrosion may have been a major factor in initiating the cracks, the primary driving force for crack growth was mainly mechanical in nature The corrosion fatigue may have resulted the cracking; both high and low cycle fatigue were involved, with high cycle initiating the fatigue and the low cycle propagating it The fracture appearances were studied at high magnification by electron microscopy Striations were found (Figure 2) substantiating the evidence that crack growth

was taking place by fatigue, although the striation spacing was unreliable as a measure of the growth rate, since a large range of temperatures were involved (200 - 450°F)

Fig 2 Fractographs of the tip of a Deep Crack [1]

Instrumentations were installed at various plants to measure vibration and displacements of the feed water piping as well as temperatures in the vicinity of pipe to nozzle junction The plants (both with and without observed cracking) were surveyed to determine their transient operation history and chemistry control Particular attention was paid to the feed water oxygen content because of the presence of pitting Thermocouple data of the on-site testing demonstrated the existence of persistent pipe thermal stratification during low feed water flow operations such as feed water makeup cycling during hot standby

4 Flow model studies

Based on flow model tests it was shown that the temperature profile in a stratified cross section is mainly correlated with two thermal hydraulic parameters: (a) the flow rate in the line, and (b) the temperature difference between the top and the bottom of the pipe cross-section under consideration The flow model test was a full scale feed line and nozzle assembly made of Plexiglas for visual observation and fluid temperature measurement

Feed Water Line Cracking in Pressurized Water Reactor Plants 211 (Figure 3) The test was designed to establish the temperature profile of the stratified water more accurately than the field measurements and to determine that thermal striping exists

at the stratified interface, and if so determine the magnitudes and frequencies

Fig 3 Flow Model Test showing Stratification: upper clear layer hot water, lower gray layer cold saline solution [2]

The fluid temperature oscillations were recorded and it was subsequently confirmed that thermal striping mechanism led to feed line thermal fatigue

5 Structural analysis

`During normal plant operation at low power conditions water is supplied to the steam generators at very low flow rates When the flow rate is not high enough to completely fill the nozzle, hot water from the steam generator remains in the nozzle to fill up the rest of the volume The difference in buoyancy between the hot and cold fluids inhibits their mixing so that the feed water becomes and remains thermally stratified as long as the flow rate is less that that required to completely fill the nozzle During normal plant operation a series of temperature measurements was taken around the pipe circumference at the vicinity of the pipe weld Analysis of the test data indicated that the stratified temperature distributions may be grouped into six basic profiles corresponding to different levels of the interface between the hot and cold fluids and are shown in Figure 4

Fig 4 Stratified Temperature Profiles [3]

Feed Water Line Cracking in Pressurized Water Reactor Plants 213

A finite element model has been prepared that includes a part of the steam generator shell, the feed water nozzle and the connecting elbow The model uses 20-node isoparametric solid elements, two elements through the thickness and twelve around the circumference of the model

The shell/nozzle/elbow model contains three distinct regions with different heat transfer characteristics between the metal and the adjacent fluid The first region is that of the inside of the steam generator shell exposed to slowly moving hot water The other regions are the section of the nozzle under the thermal sleeve, and the rest of the nozzle and the elbow

Each of the stratification profiles produces a complex stress state throughout the nozzle and the elbow The highest stresses occur in the weld counter bore region at the root of the elbow transition For each profile there is a zone of compressive stress above the hot/cold interface and a region of tensile stress below it Different interface levels produce peak stresses at different locations around the circumference Since the interface level varies during low flow operating conditions, each point in the counter bore area is subjected to varying stress state

Fatigue evaluations have been performed around the circumference for the counter bore transition root and along the top and side of the counter bore region The load conditions and the number of cycles were combined with a pressure of 7.6 MPa A maximum range of stress intensity analysis was performed prior to each fatigue evaluation to determine whether the simplified elastic plastic analysis procedure would be required and if so, to calculate the plasticity intensification factors, Ke factors by which the peak alternating stresses are to be multiplied

The results for a typical plant fatigue evaluation [3] indicate that the peak usage factors are well above 1.0 and occur at the top and sides These correlate with the observed locations of the deepest cracks for that plant The high usage factors conclusively implicate thermal stratification and thermal striping during low flow conditions as prime contributors to the observed feed line cracking

The stress distributions for the Profiles 1 through 6 has been computed using the approximate numerical model and are shown in Figure 6

The maximum range of stresses occurs at the top of the pipe and equals 72-(-124) = 196 MPa (based on profiles 2 and 1) Although the peak stress due to through the thickness temperature has not been explicitly considered, a conservative value of 2.0 is used This makes the alternating stress amplitude as 196 MPa, which gives the allowable number of cycles about 30,000 using the design curve of [5] The plant data in [4] indicates a comparable number of stratification temperature excursions This leads to a significant fatigue usage factor at the top of the pipe that correlates with the fatigue cracks observed at this location