Avoiding HTHA Failures: An Owner-User Perspective of the HTHA Risk Mitigation Process

The views expressed in this article are those of the author and do not necessarily reflect the views of Valero Energy Corporation.

Introduction

In my article published in the last issue of Inspectioneering Journal (Nov/Dec 2013) I began the discussion on what it takes to establish an effective High Temperature Hydrogen Attack (HTHA) risk mitigation program. I not only provided a brief introduction and historical summary of HTHA, but I also discussed how to properly define the scope of work, collect critical data, determine hydrogen partial pressure and process temperatures, and use this data to compliment your site’s Positive Material Identification (PMI) program. In Part II, I will be highlighting the importance of several other recommended aspects of an effective HTHA risk mitigation strategy. The intent of these two articles is to share lessons learned from recent experiences setting up a comprehensive HTHA review process across multiple refineries, and to help other operators define and mitigate the HTHA risk to an acceptable level.

Quantifying Risk

Now that scope is defined and data is collected and analyzed, what next? Not all the items in the original scope may be susceptible to HTHA and for those that are, there are varying degrees of HTHA likelihood, depending on how severe the service. API has developed methodologies for determining likelihood and priority using many of the items listed in the data section of this article. These methods are highlighted in both API RP 941 and API Technical Report 941. Perhaps the simplest, and one that is used by many owner-users, is to determine HTHA likelihood according to how close the operating conditions are to the Nelson Curve for a given material. Although the Nelson Curves have been developed based on real data accumulated over time and the confirmed presence of HTHA, a few failures have been documented for materials below their respective curves. Some of this may be due to incorrect assumptions of operational data and some may be due to the curve itself. Because of this, operating carbon steel just below the carbon steel Nelson Curve is not a guarantee that HTHA will not occur. One needs to assign some “cushion” below the curve for temperature, and to the left of the curve for hydrogen partial pressure to be not just conservative, but also realistic. Some owner-users have used a starting point of 25°F below and 25 psia to the left as the first likelihood hurdle, while others have used 50°F and 50 psia. Some apply the former to heat treated carbon steels and the latter to non-heat treated carbon steels. Each company developing their program will have to decide what works best for them, depending on the certainty of data and their risk tolerance.

Carbon - 1⁄2 Mo steels and Mn- 1⁄2 Mo steels are normally treated a little differently. Since 1970, several (27) accounts of HTHA have been reported on C - 1⁄2 Mo steels occurring below the 1977 published 1⁄2 Mo curve, which already had been lowered from the 1970 version. This resulted in the removal of the 1⁄2 Mo Nelson Curve from the other curves in 1990. It can still be located in Annex A of API RP 941, but the uncertainty of the HTHA resistance of C - 1⁄2 Mo steels makes determining the likelihood less clear. Many owner-users have referenced the position of the C – 1⁄2 Mo operating parameters above the carbon steel curve. The likelihood of HTHA increases as conditions plot further above and to the right of the curve. As with carbon steel, where the risk lines are drawn depend on the specific company’s protocol and risk tolerance. The most conservative approach is to identify any C - 1⁄2 Mo component operating above the carbon steel curve as having HTHA likelihood, with that likelihood increasing as you move up and right on the curve.

Note that per the latest edition of API RP 941, no reports of Mn – 1⁄2 Mo steel experiencing HTHA below the 1⁄2 Mo Nelson Curve have occurred, therefore determining the likelihood of HTHA for this material should take that fact into account.