Case study on failures of duplex stainless steel heat exchangers

By Immad Ovais Rohela, Olayan Descon Industrial Company Jubail, Saudi Arabia & Ahmad Raza Khan Rana, Integrity Products & Supplies Inc. Sherwood Park, AB, Canada

Introduction & background

Duplex stainless steel (DSS) is widely used in the chemical and hydrocarbon industry where the corrosive nature of service demands high corrosion resistance. Despite its high corrosion resistance, DSS is challenged by the service as well as operating conditions. Formic acid is a corrosive medium (with a scale of 0.76 of HCl) and is often used as a softening agent in the rubber Industry. The following article discusses a series of failure events that happened in a twin pass heat exchanger with tubes made from super DSS (Grade 2507) and a tube sheet made from austenitic stainless steel (Grade 316L). It also discusses the mechanical and metallurgical outlooks of various failure events.

Case study

A twin pass heat exchanger made from SS316 containing 10% formic acid on the inside of the tubes revealed some pressure drop on the tube side. Upon inspection, some formic acid deposits on the shell side were found. In addition, tubes revealed significant thinning, whereas 100 tubes (out of 531) were plugged within the first two years which made this exchanger unable to achieve its required heat duty. Consequently, the tube bundle was replaced with a new one. Despite the deployment of the new tube bundle, the tube thinning events continued. Within a period of two years, 61 tubes manifested severe thinning and even leaks. Since tube thinning was suspected to come from the corrosivity of formic acid, it was decided to retube the heat exchanger with super duplex SS Alloy 2507 (UNS S32750), which has a significantly higher pitting resistance equivalent number (PREN) in comparison to SS 316L.¹ The new tubes were assembled in the tube bundle followed by expansion and strength welding. Despite the upgrade to a newer tube metallurgy, the existing tube sheet and expansion joint were re-used where the welds between dissimilar materials (SS 316L & Alloy 2507) were made using ER 2209 filler wire as per the available welding procedure specification (WPS). Table 1 shows the physical and mechanical properties of the materials for the tube, tube-sheet, and filler wire.

After the upgrade, no leaks were observed during the first few months of operation. At the time of decommissioning for the first planned outage (since the tubes were upgraded), in addition to abnormal pressure reduction, formic acid traces were found on the shell side, which was a tell-tale sign of leaks in the tube bundle. The exchanger was dismantled, followed by thorough visual and non-destructive examination (NDEs). Eddy current testing (ECT) was performed to inspect for thinning damage to the tubes, especially at the tube-tube sheet expanded portions. Dye penetrant examination at the tube-tube sheet strength welds revealed cracks on the majority of tubes (> 20%) that resulted in the leakage of formic acid from tube side to shell side portions. The following discussion is focused on the mechanical and metallurgical aspects. Figure 1 shows the arrangement of tube-tube sheet welds that were subjected to cracks.

Table 1. Physical and mechanical properties of SS316 L, Alloy 2205 & Alloy 2507²

Mechanical design outlook

Referring to Figure 1, since the longitudinal expansion of the tubes was restrained by the strength of the weld(s) between tube-tube sheet, these welds were subjected to compressive forces. The super DSS tubes, standard DSS weldment and austenitic SS 316L tube sheet, have entirely different mechanical properties. Table 1 summarises the mechanical properties of the base material (tube, tube sheet) and weldment. Considering the length of the tube bundle (7 meters) and thermal expansion coefficient (α) from the table 1, the tube elongation (δ) and stresses (i.e., P/A) can be found via the below equations:²

The stresses from compressive forces (alone), as calculated from Eq (2), was 226 MPa. In addition to thermal elongation and consequential stress, high temperatures may also result in a reduction in the overall strength of DSS. Expansion joints are known to compensate for the thermal expansion-induced stresses. However, the expansion joints were made from the SS 316L, which itself has a 22% higher thermal expansion coefficient than alloy 2507 (refer to Table 1), which may not compensate for the compressive stresses (due to thermal elongation). There were no engineering reviews to check the suitability of existing (i.e., SS 316L) expansion joints with Alloy 2507 tubes, which provided no clarity on the stress value/ stress field around the tube–tube sheet weld joints. Since DSS and austenitic SS don’t follow any endurance limit, (unlike carbon steel, low alloy steel, and titanium) they manifest fatigue failures even below yield point. This implies that even at the lower thermal stresses (i.e., well below the tensile strength) the cracking may still happen depending on the temperature swing and the number of cycles.¹

Metallurgical outlook

The welding between the tubes and tube sheet was performed using the approved and well-practiced welding procedure specification (WPS) deploying 1/16” (or 1.6 mm) thick filler wire ER 2209. The inter-pass temperature was kept at 320 F (or 160°C) to avoid any metallurgical defects (sensitisation, sigma phase embrittlement, etc.) as well as thermal distortions.¹ NDEs and pressure tests were performed at maximum allowable working pressure (MAWP) for the bundle with upgraded tube metallurgy. Considering the inter-pass and operating temperatures, the risk of oxidations and cracking due to σ-phase and 885 F (or 475°C) embrittlement was remote. API practices recommend a maximum inter-pass temperature of < 150°C for the standard DSS and up to 120°C for super DSS.2 The higher inter-pass temperature than the recommended limits (for super DSS) may trigger some change in the weldment behaviour. However, the filler wire in this case was made from standard DSS (i.e., ER 2209) and the tubes were from super DSS, so there were two governing (as well as conflicting) inter-pass temperature limits. This may lead to a heterogenous nature with the weldment (though not fully investigated in this study).

Since the tube sheet had a significantly higher thickness (45 mm) than the tubes and the weldment, this may have resulted in the rapid cooling of the weldment. A review of the WPS revealed that it specified an inter-pass temperature of 160°C for the tube thickness ranging from 10 mm to unlimited thickness. This may have implications in terms of higher cooling rate(s) of weldments, especially at the tube-tube sheet welds.² The rapid cooling deprives the weldment from forming the austenite, which is the key reason behind the fracture toughness in a DSS microstructure.² From the review of the metallurgical compositions, it was evident that Alloy 2205, which is a standard grade DSS, has a nitrogen content between 0.14 – 0.20 whereas Alloy 2507, which is a super DSS, has a higher nitrogen content (i.e., 0.24 – 0.2) and a better ability to form austenite than standard DSS Alloy 2209. The API practices and ASME Section-II Part 3 (A8.49 & A8.52) recommend super DSS grade filler wire (i.e., ER 2594) for the welding of super DSS grades such as Alloy 2507. The ferrite number of ER 2209 (which is a standard DSS) is 54-60, whereas for the super DSS grade ER 2594 it’s 45-55. This implies the higher propensity of ER 2209 to retain the dominant ferritic microstructure (instead of transition to austenite) that generally results from the presence of nitrogen during the slow cooling of weldment(s).² In summary, the higher nitrogen content and lower ferrite in the ER 2594 super DSS filler wire (than standard DSS) will have enhanced ability to form the austenite and hence the increased toughness of weldment than ER 2209.

Figure 2 shows a schematic of the weld between tube and tube sheet.

So, the failure of the weldment (via cracking) resulted from several factors (or their combined effects) as below:

• High tensile stresses from thermal elongation of tubes, since the SS 316L expansion joints could be incompatible with the Alloy 2507 due to the 22% higher coefficient of thermal expansion
• The reduced toughness of the weldment due to the reduced austenite caused by the rapid cooling of the weldment
• Internal stresses due to dissimilar welds may have contributed to dissimilar metal weld (DMW) cracking¹
• Fatigue failure due to thermal cycling caused by multiple start-up and shutdowns

Summary

Under heat transfer conditions, corrosion rates from formic acid are higher than in the non-heat transfer condition. This factor was not accounted for during the material selection for the twin pass heat exchanger in the first place. The cracking that occurred even after the tubes were upgraded to Alloy 2507 stemmed from high tensile stress at the weld-tube sheet junction due to the tube expansion force (due to dissimilar thermal expansion coefficient). In addition, internal stresses due to dissimilar welds may have also contributed to dissimilar metal weld (DMW) cracking. Finally, fatigue due to thermal cycling was caused by multiple start-ups and shutdowns. The repairs involving dissimilar metals and corrosive service should be accompanied by Engineering reviews.

About this Tech Article

Appearing in the April 2025 issue of Stainless Steel World Magazine, this technical article is just one of many insightful articles we publish. Subscribe today to receive 10 issues a year, available monthly in print and digital formats. – SUBSCRIPTIONS TO OUR DIGITAL VERSION ARE NOW FREE.

Every week we share a new technical articles with our Stainless Steel community. Join us and let’s share your technical articles on Stainless Steel World online and in print.