1、STP-PT-064EVALUATION OF FRACTURE PROPERTIES TEST METHODS FOR HYDROGEN SERVICESTP-PT-064 EVALUATION OF FRACTURE PROPERTIES TEST METHODS FOR HYDROGEN SERVICE Prepared by: M. T. Miglin, PhD. Stress Engineering Services, Inc. Date of Issuance: June 24, 2013 This report was prepared as an account of work
2、 sponsored by ASME Pressure Technology Codes other forms of hydrogen damage are not included. The feasibility of a safety factor approach for the adjustment of the da/dN data from experiments to correspond to actual design parameters could not be specified because of the lack of full-scale test data
3、. Recommendations for an appropriate KTH value that represents expected vessel behavior are made. 2.1 Measurement of KTH Current testing procedures (paragraph KD-1045 ASME BPVC VIII, Div. 3) allow the determination of KTH by either constant displacement or constant load testing of fatigue precracked
4、 specimens. A review of these procedures and of the rising load test procedure is presented below. The available literature has been searched for the following information: (a) Static constant load or displacement vs. rising load test data (b) Effect of exposure time on static tests (c) Effect of st
5、rain or loading rate on rising load test (d) Effect of specimen size on both test methods (e) Application of static and rising load test data to actual pressure components in the presence of a crack. A critical evaluation has been made of the existing test procedures specified in ASTM and ISO standa
6、rds for measurement of KTH in hydrogen. The existing KTH test data for pressure vessel and piping steels in hydrogen has been compiled and compared. The literature was searched for full-scale test data to compare with properties obtained using the above-defined test procedures. No full-scale test da
7、ta were found. 2.2 Measurement of da/dN The available literature has been searched for information on the effects of test frequency, shape of the load-time cycle, and R-ratio on threshold fatigue crack growth rate, as follows: (a) The appropriate test frequency and R-ratio to be used in da/dN measur
8、ements are suggested. The test frequency and the R-ratio should adequately reflect the actual loading conditions expected in vessels during service, but are also be realistic for laboratory test conditions. No data was found for shape of the load-time cycle. (b) The effect of R-ratio on da/dN has be
9、en evaluated. Evaluation of Fracture Properties Test Methods for Hydrogen Service STP-PT-064 3 (c) A value of KTH for Q however, the presence of hydrogen increases the crack growth rate. Hydrogen reduces the KTH, and raising the R-ratio decreases KTH further. Figure 11 - Crack Growth Rate as a Funct
10、ion of K Showing Effects of Hydrogen and R-Ratio on the Shape of the Fatigue Crack Growth Rate Curve Note: IG = intergranular, T = transgranular, KTmax = threshold value, below which hydrogen has no effect on crack growth rates until the lower threshold region is reached 12. Evaluation of Fracture P
11、roperties Test Methods for Hydrogen Service STP-PT-064 13 The effect of hydrogen on the Ni-Cr-Mo steels HY80 and HY130 is shown in Figure 12. Hydrogen accelerates the fatigue crack growth rates for both steels over the air levels. Hydrogen has a greater effect on the lower strength steel. Figure 12
12、- Fatigue Crack Growth Rate of HY-80 and HY-130 in Air and Hydrogen 11 Figure 13 - Fatigue Crack Growth Rate of HY100 Steel in Hydrogen and Helium 11 Above the threshold region, the effects of hydrogen on fatigue crack growth are not simply due to the absence of oxygen. Hydrogen also increases fatig
13、ue crack growth rates relative to inert environments, K levels above threshold. Figure 13 shows higher fatigue crack growth rates (at K levels of 20 MPam and above) for HY100 in hydrogen than in helium. STP-PT-064 Evaluation of Fracture Properties Test Methods for Hydrogen Service 14 STB-X-2005Style
14、 Guide4.1 Effects of Frequency Cyclic frequency affects fatigue crack growth rates in hydrogen, but the magnitude of the effect depends upon the R-ratio and the steel. Figure 14 shows fatigue crack growth rates for DOT 4130X cylinder steel in hydrogen at two frequencies and R-ratios. Increasing the
15、frequency from 0.1 to 1.0 Hz affects the crack growth rates at R=0.5 but not at R=0.1 13. Figure 14 - Fatigue Crack Growth Rates for 4130X in Hydrogen at Two Frequencies and R-Ratios As shown in Figure 15, the effects of frequency on fatigue crack growth rate are prominent at high K values. Decreasi
16、ng frequency from 25 Hz to 0.01 Hz increases the fatigue crack growth rate by more than an order of magnitude. Figure 15 - Fatigue Crack Growth Rates for A3750 in Hydrogen at Frequencies Ranging from 0.01 to 25 Hz 14 Evaluation of Fracture Properties Test Methods for Hydrogen Service STP-PT-064 15 T
17、he effects of low frequencies are shown in Figure 16 for SA-105 C-Mn steel in 100 MPa H2. Crack growth rate is observed to increase with decreasing frequency from 1 Hz down to 0.00083 Hz, and the effect does not appear to saturate or reverse over that frequency range. Figure 16 - Fatigue Crack Growt
18、h Rate of SA105 C-Mn Steel in Hydrogen and Helium over a Range of Frequencies from 0.00083 to 1 Hz 10 In contrast, Figure 17 shows fatigue crack growth rates for X52 pipeline steel that remain stable as frequency is decreased from 0.1 Hz to 0.001 Hz for K=17.5 MPam. However, these tests were perform
19、ed at lower K levels than those in Figure 16. The frequency effect on crack growth rate is a function of the magnitude of da/dN. Figure 17 - Fatigue Crack Growth Rates as a Function of Frequency for X52 Pipeline Steel Tested at R=0.5 in Hydrogen Gas 13 STP-PT-064 Evaluation of Fracture Properties Te
20、st Methods for Hydrogen Service 16 STB-X-2005Style GuideIn Figure 18, fatigue crack growth data in hydrogen are plotted with respect to inverse frequency, such that the abscissa is in seconds. All testing was done at K=23 MPam 10. At this K level, in gaseous hydrogen, taking five seconds or more to
21、reach Kmax produces the fastest crack growth rates. This corresponds to a frequency of 0.1 Hz. Figure 18 - Effect of Frequency on Crack Growth Rates for C-Mn Steels Charged with Hydrogen in Aqueous Solution and in Gaseous Hydrogen The frequency selected for fatigue crack growth rate testing in hydro
22、gen must balance the conflicting issues of test duration and data reliability. The effect of frequency on crack growth rates diminishes at low da/dN levels, so near the threshold, higher frequencies can be used. From the data above, a frequency in the vicinity of 0.1 Hz appears to be a reasonable va
23、lue. 4.2 Effects of R-Ratio The applied load ratio R=Kmin/Kmax affects fatigue crack growth rates in the threshold region. Increasing the R-ratio increases the crack growth rates and thereby pushes the threshold to lower K levels. The data in Figure 19 illustrate this effect for 4340 low alloy steel
24、 in dry hydrogen. Figure 19 - Near-Threshold Fatigue Crack Growth Rate Curves for 4340 Steel in Dry Hydrogen at Various R-Ratios 15 Evaluation of Fracture Properties Test Methods for Hydrogen Service STP-PT-064 17 Figure 20 shows near threshold results in dry hydrogen plotted for several NiCrMoV ste
25、els. The data are plotted as KTH (termed K0) as a function of R-ratio. From this figure it is apparent that increasing R-ratio reduces KTH, and the effect appears to be linear and larger for lower strength steels. Based on these results, testing at an R-ratio of 0.8 appears to be conservative. It is
26、 apparent from Figure 19 and Figure 20 that KTH for high-strength steels (yield strength above 1000 MPa) in hydrogen is quite low at 2-3 MPam. Figure 20 - Effect of R-Ratio on KTH for NiCrMoV Steels Tested in Dry Hydrogen 16 Figure 21 presents fatigue crack growth data for 4130X specimens with yield
27、 strength equal to 607 MPa tested in hydrogen at two frequencies and R-ratios. The air data represents the crack growth relationship provided in ASME VIII-3 Article KD-4 17. At high K, fatigue crack growth rates in hydrogen are more than an order of magnitude greater than rates in air. Figure 21 - F
28、atigue Crack Growth Rates for Four 4130X Specimens (Open Symbols) in 45 MPa Hydrogen and Expected Fatigue Crack Growth Data in Air (Solid Line) STP-PT-064 Evaluation of Fracture Properties Test Methods for Hydrogen Service 18 STB-X-2005Style GuideThe most aggressive combination of test conditions is
29、 the lower frequency (0.1 Hz) combined with the higher R-ratio (0.5). Individual sections of each curve can be fitted to power-law relationship such as equation (1) da/dN=CKm (1) in which the coefficient, C, and the exponent, m, are constants. The values of the constants for several K ranges are giv
30、en in Table 2. Table 2 - Fatigue coefficients derived from fitting the data in Figure 20 to the power-law relationship given as Equation 1 17 Figure 22 presents fatigue crack growth rates for X80 pipeline steel in hydrogen at two R-ratios. It is apparent from Figure 22 that increasing the R-ratio fr
31、om 0.1 to 0.5 increases the fatigue crack growth rates. At low K levels, the hydrogen data approach the air data. Figure 22 - Crack Growth Rate as a Function of Applied K for X80 Pipeline Steel Tested in Dry Hydrogen 1 Evaluation of Fracture Properties Test Methods for Hydrogen Service STP-PT-064 19
32、 Figure 23 shows the data from Figure 22 replotted as crack growth rate per cycle versus Kmax. It is apparent from Figure 23 that, if the Kmax can be maintained at a low enough level in a real structure, that crack growth rates in hydrogen will be close to those in air. When plotted versus Kmax, cra
33、ck growth rate decreases with increasing R-ratio. Figure 23 - Data from Figure 21 Replotted as Crack Growth Rate versus Kmax 1 The same effect of Kmax on crack growth rates is observed for X60 (see Figure 24 and Figure 25). In Figure 25, the data in 21 MPa H2 tested at R=0.5 approach and meet the ai
34、r curve at Kmax18 MPam. Suresh and Ritchie 12 proposed a stress intensity factor threshold (KTmax as shown in Figure 11) below which hydrogen does not affect fatigue crack growth rates. This is entirely conceivable, given that crack growth rates in hydrogen approach those in air as Kmax decreases. H
35、owever, as shown in section 4.3 below, this generalization does not apply in the threshold region. Figure 24 - Crack Growth Rate as a Function of Applied K for X60 Pipeline Steel Tested in Dry Hydrogen 1 STP-PT-064 Evaluation of Fracture Properties Test Methods for Hydrogen Service 20 STB-X-2005Styl
36、e GuideFigure 25 - Data from Figure 21 Replotted as Crack Growth Rate versus Kmax 1 In summary, increasing R-ratio results in increased fatigue crack growth rates when plotted versus K. According to Figure 20, testing at high R-ratio produces a lower KTH, suggesting that threshold tests should be co
37、nducted at an R-ratio of 0.8 or higher. The data in Figure 19, 21, 22 and 24 all show that testing at higher R-ratios is more aggressive than testing at low R-ratios. From Figure 19 and Figure 20, it appears that an R-ratio of 0.8 or larger is needed to produce conservative crack growth data. 4.3 Th
38、reshold Effects Testing at low K levels is complicated by the effects of crack closure. Specimens tested in moist air have oxide films on the crack faces, which may be thickened by fretting oxidation. These films wedge the crack open, reducing effective K levels at the crack tip. This does not happe
39、n in dry hydrogen, so near-threshold crack growth rates in dry hydrogen are higher than those in air. The same is true for dry argon, although it is inert. Near-threshold fatigue crack growth rates in dry argon are comparable to those measured in hydrogen, and higher than those measured in air. Corr
40、espondingly, KTH values measured in dry hydrogen and dry argon are similar, and both are lower than KTH measured in air 18. Threshold behavior is influenced by R-ratio. Figure 26 shows that, as R-ratio is increased, KTH approaches 4 MPam (4.4 Ksiin). Evaluation of Fracture Properties Test Methods fo
41、r Hydrogen Service STP-PT-064 21 Figure 26 - Variation of Alternating and Maximum Stress Intensities at Threshold, K0 and K0,max, Respectively, with Load Ratio R for SA387-2-22 Tested in Moist Air and Dry Hydrogen at 50 Hz 18 Threshold values for fatigue crack growth in hydrogen are discussed in Ref
42、erence 17. Figure 27 and Figure 28 present threshold data for 2.25 Cr 1Mo steel and X70 pipeline steel. These data support a threshold value between 3 and 4 MPam (3.3 to 4.4 Ksiin) at R=0.75. The data in Figure 20 support an even lower threshold of 2.0 MPam (2.2 Ksiin). In agreement with Ritchie, et
43、 al 18, the threshold decreases as R-ratio is increased. Figure 27 - Fatigue Crack Growth in 2-1/4 Cr 1 Mo Steel Tested in Moist Air and Dry Hydrogen at Atmospheric Pressure 19 STP-PT-064 Evaluation of Fracture Properties Test Methods for Hydrogen Service 22 STB-X-2005Style GuideFigure 28 - Fatigue
44、Crack Growth in X70 Pipeline Steel Tested in Moist Air and Hydrogen 19, 20 4.4 Summary The limited available literature was reviewed for information on the effects of test frequency, shape of the load-time cycle, and R-ratio on threshold fatigue crack growth rate. The appropriate test frequency and
45、R-ratio to be used in da/dN measurements are suggested. The test frequency and the R-ratio should adequately reflect the actual loading conditions expected in vessels during service, but are also be realistic for laboratory test conditions. No data was found for shape of the load-time cycle. The eff
46、ect of R-ratio on da/dN has been evaluated. A value of KTH for quenched and tempered steels has been identified for inclusion in ASME BPVC VIII-3 for hydrogen vessel design. Fatigue crack growth rates in hydrogen are affected by loading frequency. The frequency selected for fatigue crack growth rate
47、 testing in hydrogen must balance the conflicting issues of test duration and data reliability. The effect of frequency on crack growth rates diminishes at low K levels, so near the threshold, higher frequencies can be used. From the data above, a frequency in the vicinity of 0.1 Hz appears to be a
48、reasonable value for testing of quenched and tempered steels in hydrogen. Increasing R-ratio results in increased fatigue crack growth rates when plotted versus K. Data from various sources are in agreement that testing at higher R-ratios produces faster crack growth rates than testing at low R-rati
49、os. An R-ratio of at least 0.8 is needed to produce conservative crack growth data. This is also true in the threshold region. Crack growth rates in the treshold region are higher in hydrogen than in moist air. However, crack growth rates in argon are also higher than those in air in the threshold region. Crack growth rates in hydrogen and argon are comparable. It is likely that hydrogen and argon produce faster crack growth rates and lower KTH bec
