Our #EpicFail series continues, describing some of the most interesting and obscure failure mechanisms that we have encountered through our failure investigation work at R-TECH. This week we take a look at High Temperature Hydrogen Attack (HTHA)

HTHA is a phenomenon which occurs across a variety of industries, particularly in the petrochemical industry. HTHA is associated with a loss of tensile strength and ductility and is preceded by a period of time where no noticeable changes in properties are detectable by non-destructive techniques, consequently catastrophic failure can occur without prior warning signs. A recent explosion and fire in April 2010 at a refinery in Washington where 7 lives were lost was a consequence of HTHA in a heat exchanger in the Naphtha Hydrotreater unit.

Figure 1 – Ruptured heat exchanger which led to an explosion and fire at a refinery in Washington

HTHA occurs in carbon steels and alloy steels upon exposure to elevated temperatures and pressures. At high temperature, hydrogen gas molecules dissociate at the steel surface to atomic hydrogen which readily diffuses into the steel. The hydrogen reacts with dissolved carbon and with metal carbides to form methane gas. The methane molecule is much larger than the hydrogen atom and therefore too large to diffuse through the steel. As a result, internal methane pressure builds up, leading to bubbles, cavities and fissures which may coalesce to form cracks (see Figure 2). At higher temperatures, the dissolved carbon diffuses to the steel surface and combines with atomic hydrogen to form methane. This leads to surface decarburisation which affects the overall material strength.

Figure 2 Cracking due to HTHA

Early stages of HTHA damage is usually only noticeable using destructive metallography since the phenomenon generally occurs sub-surface and is too fine to be detected using conventional NDT methods. This is recognised by voids/fissures at the grain boundaries associated with decarburisation, see Figure 3.

Figure 3 Early stages of HTHA

Various industries use Nelson Curves to predict the susceptibility of various materials to HTHA damage. Elements that stabilise carbides, such as chromium and molybdenum, increase resistance to HTHA and this is reflected in the Nelson curves. It’s important to recognise that while the Nelson Curves serve as a guideline, there have been instances where HTHA occurred below these curves which was the case for the incident at the refinery in Washington in 2010.  Therefore, operators should exercise caution and consider additional factors beyond the Nelson Curves when assessing HTHA risks. The most recent edition of the Nelson Curves is included in the 8th edition of API Recommended Practice 941 (API RP 941), published in February 2016. Notably, the 8th edition introduced a new curve for non-post weld heat treated (PWHT) carbon steel, which is approximately 10°C lower than the previous curve used for non-welded or PWHT carbon steel. Time of exposure is an aspect not yet considered in the curves. The time variable is expected to be introduced into future editions of RP 941. Furthermore, operators should be mindful that if their understanding of the mechanism improves, the curves are likely to evolve further, resulting in more constraints.

Susceptibility to HTHA increases in areas of high stress or stress concentration since hydrogen preferentially diffuses to these areas. This can lead to HTHA damage in components operating below the Nelson curves. Isolated areas of decarburisation and fissures are often found adjacent to weldments. Fissures tend to be parallel to the edge of the weld rather than the surface due to the presence of residual stress. Fissures in this direction can form into through-thickness cracks and are therefore significantly more detrimental.

The following measures can be implemented to reduce the likelihood of HTHA ⁽¹⁾:

1. Using alloy steels with chromium and molybdenum will increase carbide stability, thereby minimising methane formation.
2. A common design practice is to use a 15 °C to 30 °C and 170 KPa to 345 KPa hydrogen partial pressure safety factor approach when using the API RP 941 curves.
3. The C-½Mo curve was removed from the Nelson curves in 1990 because of a number of cases of HTHA in C-½Mo steels in refinery service under conditions that were previously considered safe. This material is not recommended for new construction in hot hydrogen services. For existing C-½Mo equipment, the concern about its unpredictable resistance to HTHA has prompted refiners to perform reviews of inspection effectiveness and cost vs replacement with a more suitable alloy.
4. A stress relieving heat treatment can reduce the likelihood of HTHA.
5. The use of NDE methods to detect internal HTHA requires highly specialised training, skills, and experience. It has had mixed results and is an area of ongoing development. API RP 941 provides details on what NDT methods are suitable for detecting HTHA.

Reference:

1. API RECOMMENDED PRACTICE 571, 2020. DAMAGE MECHANISMS AFFECTING FIXED EQUIPMENT IN THE REFINING INDUSTRY. High-temperature Hydrogen Attack. Page 189.