Damage Control: High-Temperature Hydrogen Attack Assessment

Editor’s Note: This regular column offers practical insights into various damage mechanisms affecting equipment in the O&G, petrochemical, chemical, power generation, and related industries. Readers are encouraged to send us suggestions for future topics, comments on the current article, and raise issues of concern. All submissions will be reviewed and used to pick topics and guide the direction of this column. We will treat all submissions as strictly confidential. Only Inspectioneering and the author will know the names and identities of those who submit. Please send your inputs to the author at damagecontrol@inspectioneering.com.

Introduction

The previous installment of Damage Control provided a historical overview as well as the fundamental damage characteristics associated with high-temperature hydrogen attack (HTHA). Additionally, the underlying chemical reaction that leads to HTHA damage in pressure equipment, where dissociated hydrogen atoms react with carbon in steel to form methane, was described [1,2]. Moreover, a summary of a significant, catastrophic industry HTHA failure was offered, and commentary on modern inspection methods that can be leveraged to reliably detect and meaningfully characterize HTHA damage was outlined. This article represents Part 2 of this 3-part series on HTHA. In this installment of Damage Control, a high-level overview of HTHA damage predictions and remaining life assessment methods are covered.

While HTHA has been recognized as a longstanding damage mechanism in the oil and gas, chemical, and fertilizer industries, mechanistic HTHA damage models for common pressure equipment have only been established and well-documented within roughly the last decade. As explained herein, any robust HTHA damage prediction model should account for volumetric HTHA damage (that is, material void coalescence due to internal methane formation) and the potential for crack-like flaw propagation. Historically, the Nelson Curves (as documented in evolving editions of API 941) have been used as an approximate HTHA damage screening tool to offer perspective on damage susceptibility based on available operating data from the oil refining industry [3]. However, the Nelson Curves are not inherently time-dependent, and they introduce uncertainty and numerous limitations, especially when characterized HTHA damage is accurately described in pressure retaining components using modern nondestructive examination (NDE) methods [3]. In these instances, utilizing more advanced engineering assessment approaches may be warranted to better quantify remaining life. To this end, a synopsis of an HTHA assessment case study for a vintage C-0.5Mo hydrotreater reactor is provided herein.

As outlined in Part 1 of this Damage Control series, HTHA is a complex damage mechanism that has been studied for many years. Collectively, while significant time and effort have been expended to better understand the behavior of materials in HTHA environments, accurate and well-documented HTHA damage prediction models remained elusive for decades. After the fatal 2010 Tesoro Anacortes HTHA failure, a renewed industry focus on HTHA inspection and engineering assessment methods has led to the formation of different joint industry projects (JIPs), novel material testing programs, and published HTHA damage models (developed by engineering service providers and owner-users) [4-11]. These methods range from purely empirical in nature to rigorous physics-based/mechanistic approaches. To be effective, empirical approaches often require extensive datasets while mechanistic models require validation and comparison to practical damage examples and documented failure case studies.

These industry research and testing efforts have been aimed at advancing HTHA inspection technology and quantifying material resistance to HTHA via damage prediction models. The latter efforts have attempted to increase industry’s collective understanding and ability to reasonably predict the rate of volumetric HTHA damage, manifesting in the form of methane-driven micro-voids that coalesce over time and eventually accelerate macro-scale crack-like flaw initiation and propagation. Some recent analysis methods have also been leveraged to generate recommendations for implementation of time-dependent HTHA screening curves in API 941. For the sake of brevity, details associated with and a comparison of these published HTHA damage prediction models are not provided herein. Additionally, damage progression is known to be influenced by material composition (alloy), operating metal temperature, hydrogen partial pressure, applied/residual stress, and heat treatment. Furthermore, damage progression rates are usually exacerbated by the presence of original fabrication/weld defects.

An example of HTHA-driven damage in a pressure vessel is shown in Figure 1. Specifically, a photo-macrograph of a cross-section of a longitudinal seam weld from a C-0.5Mo heat exchanger channel is depicted. HTHA damage (cracking) was observed in the weld heat-affected zone (HAZ) and adjacent base metal. Figure 1 also shows magnified images illustrating characteristic intergranular cracking in the coarse-grained HAZ. In this case, the depth of the observed cracking extends approximately 0.20 inches (5 mm) from the inside surface of the vessel. Internal decarburization adjacent to the cracks resulted in the elimination of most bainitic and/or martensitic laths formed in the coarse-grained HAZ. Ultimately, the cracking terminated at the overlapped region in the HAZ, where the original coarse-grained HAZ grain size was reduced. Hydrogen-driven cracks were also observed in the fine-grained HAZ near the inside surface and extended to a depth of 0.06 inches (1.4mm) [12]. Note that the weld fusion line (denoted FL) separating the weld deposit (denoted WM for weld metal) from the HAZ is shown in Figure 1.