API TR 934-F PART 4
The Effects of Hydrogen for Establishing a Minimum Pressurization Temperature (MPT) for Heavy Wall Steel Reactor Vessels
|Publication Date:||1 November 2018|
Hydrogen, dissolved in the thick wall of a steel pressure vessel during steady-state operation in elevated temperature high-pressure H2, can cause both slow-subcritical crack advance, as well as unstable catastrophic fracture during shutdown and startup. This behavior is defined in Section 2. It follows that modern fracture mechanics assessments of the minimum pressurization temperature (MPT) and fitness for service (FFS) must include the deleterious effect of H on both subcritical and unstable internal hydrogen-assisted cracking (IHAC). Two approaches are in draft stage to develop standard procedures that address this need: an API 934-F recommended practice and a WRC Bulletin 562 basis for ASME/API 579.
The objective of this technical report is to establish the technical basis necessary to enable and validate these best practices for quantifying the effects of hydrogen on (a) the MPT, and (b) FFS of a thick wall hydroprocessing reactor. The approach entails two parts. Part 1 emphasizes critical assessment and collection of two primary H-cracking properties: the threshold stress intensity for the onset of subcritical H cracking under slow-rising stress intensity (KIH), and the critical stress intensity for the onset of unstable catastrophic cleavage-like crack growth promoted by H (KIC-H). Part 2 focuses on the methods to use these data to quantitatively to predict an MPT that precludes H cracking during shutdown and startup. The sum of these two parts-validated-exte
Section 3 documents extensive KIH and KIC-H data that conservatively characterize IHAC in 2¼Cr-1Mo weld metal and base plate. The effects of critical variables are documented; including the degree of temper embrittlement in terms of the FATT after thermal exposure (FATTThermal), total H concentration, and stressing temperature. KIH data are aggregated for three classes of steel purity: Database A (low purity/high FATT) with FATTThermal > 50 °C; Database B (intermediate purity/intermediate FATT) with −30 °C < FATTThermal < 50 °C; and Database C (high purity/low FATT) with FATTThermal < −30 °C. These three steel composition categories were defined to both recognize the critical interaction of temper embrittlement with hydrogen cracking, and to optimize the combination of existing multiple IHAC data sets from different laboratories. [Alternately, the user can combine databases B and C to quantify IHAC in 2¼Cr-1Mo steels fabricated before and after (Database A) impurity-chemistry control.]
Subcritical H cracking (Section 3.2) is eliminated below a critical-dissolved H concentration and above a critical temperature, which are related through H-trapping theory to a single-critical parameter. The beneficial effect of increasing temperature is affirmed by fracture mechanics experiments with several specimen geometries, and provides the basis for MPT definition to eliminate subcritical H cracking. Fracture mechanics experiments (Section 3.3) clearly establish that dissolved H can reduce the unstablefracture toughness of 2¼Cr-1Mo weld metal and base plate, from KIC to KIC-H, consistent with the deleterious effect of H on Charpy impact energy and Charpy FATT. However, previous studies have not correctly eliminated those data that were improperly interpreted to yield a false KIC-H (e.g. due to the occurrence of innocuous pop-in events).
Validated KIC-H experiments covering a range of H-free FATT values establish that the occurrence of trueunstable crack growth correlates with (T-FATTThermal), essentially independent of dissolved H concentration and showing a distribution of behavior for a given temperature. H-promoted unstable cracking is eliminated; that is, KIC-H approaches the H-free KIC above a critical temperature equal to (Charpy FATTThermal + 66 °C) for base plate and above the Charpy impact FATTThermal for weld metal. (Hpromoted unstable cracking was never observed at absolute temperatures above 86 °C for base plate and above 25 °C for weld metal for the levels of temper embrittlement represented by existing data.) Specific KIC and KIC-H versus temperature data from this analysis are modeled in Section 4 to provide a basis for quantitative MPT determination. Additional data for high-FATTThermal temper embrittled steels, particularly stressed at temperatures above 50 °C, are required to refine the correlated temperature for the elimination of unstable-H cracking. The factor(s) that control the temperature dependence of KIC-H are not well understood, suggesting the need for improved understanding of the mechanism for H-promoted unstable cracking.
Section 3.4 presents results from a recent API-sponsored study that established the following materialproperty data relevant to MPT and FFS assessment for V-modified Cr-Mo steel. First, unstable H-cracking is unlikely at ambient temperatures (and above) given the high purity and low FATTThermal, which are typical of modern Cr-Mo-V, as well as the nil-to-small shift in Charpy FATT due to relatively high dissolved H concentration. Second, V-modified steel does exhibit very slow-stable IHAC for H concentration typical of the high H solubility of this modern steel, but only at relatively high KIH and below a relatively low-critical temperature. Cr-Mo and Cr-Mo-V steels are similarly and significantly susceptible to H cracking when stressed in moderate- to high-pressure H2 at near ambient temperature (Section 3.5). However, neither steel exhibits a deleterious interaction between IHAC and HEAC. Limited IHAC data have been reported for 1Cr-0.5 Mo, 1.25C-0.5Mo, and C-0.5Mo steels, but the results are insufficient to support an MPT or FFS assessment.
Section 4 develops the technical basis for MPT and FFS assessments using the fracture mechanics data presented in Section 3. The effects of H concentration, temperature, and cracked body geometry on KIH are effectively modeled based on the fundamental concept of crack tip H concentration similitude: Equal H damage (equal KIH) is created by equal localized crack tip H concentration. For such subcritical crack growth, an H-trapping-based model developed by Al-Rumaih and Gangloff (AG), as well as the engineering model by Anderson and Brown (AB), each use this concept to develop master curves that effectively correlate the large amount of KIH data for each of the three databases of 2¼Cr-1Mo steel (Section 4.1). The AG theoretical model fully justifies the AB engineering model, which provides the optimized engineering approach for MPT and FFS assessments aimed to minimize subcritical IHAC.
A scientific model is not available to describe H-promoted unstable cracking. Rather, the FFS procedures for avoiding unstable fast fracture upon shutdown and startup entail a combination of the API 579 Level 2 crack-like flaw assessment, combined with the Wallin Master Curve approach for fracture toughness in the transition region (Section 4.2). In the latter case, the KIC-H versus (T-To) Master Curve is conservatively described at the 95 % to 99 % confidence level using a 50 °C (90 °F) temperature increase in the H-free index temperature (ToH), resulting in ToH = FATTThermal. This temperature shift is based on an examination of the distribution of validated fracture toughness data collected and curated in Section 3.3, and is equally relevant to both base plate and weld metal. These data are insufficient to support a systematic effect of dissolved H concentration on the level of this temperature shift. The H-based shift in Charpy FATT is not used to develop this effect of H on KIC-H versus temperature. For the Level 1 fast fracture assessment, a default pressure-temperature
Section 5 lays out the proposed architecture for the MPT determination, which is defined by both stable IHAC and unstable fast fracture criteria. There are three levels of assessment for each criterion. Level 1 constitutes the simplest and most conservative method, while Level 3 is the most complex procedure and contains the least conservatism. The FFS user may mix assessment levels, depending on the situation. For example, if the MPT is limited by the stable IHAC criterion, the user may combine the Level 1 fast fracture assessment with a Level 2 or 3 stable IHAC assessment.