F. Gilbert and O. JACOB, Suncor Energy Products Partnership—Montreal Refinery, Montreal, Quebec, Canada
A thermal mix-point is a known location that can create thermal stress that leads to failure. Usually, this stress is generated by two streams that have a high-temperature differential. This is often seen at a heat exchanger bypass; however, the heat exchanger itself can also be subject to the same damage mechanism and it is important to review this possibility at the design stage. This article presents a review of a failure that happened in a hairpin heat exchanger type on the shell side in a hydrogen (H2) reformer unit. This mechanism was not identified during the design phase and led to a failure after only two years of operation. This article also reviews the situations encountered during the design stage and how they affected the material selection.
Feed effluent exchangers can be exposed to a high-temperature differential between the shell side and tube side, depending on the design type (concurrent or countercurrent flow design). In the case presented here, a hairpin feed effluent exchanger was subjected to carbonic acid condensation and required replacement. The original construction was made of 1.25Cr-0.5Mo on the shell side and tube side. The tube material was Sa-213-TP304L. The exchanger had been in operation since 2004, and a replacement was planned for 2014. FIG. 1 shows the various parts of a hairpin exchanger.1
The original recommendation was to replace the shell with an Sa-335 P11 cladded with stainless-steel (SS316L), which is known to be resistant to carbonic acid.2–5 However, due to the complexity of completing this exchanger design, it was decided to build the shell with SS316L. The reason for this selection will be reviewed in the following sections. TABLE 1 shows the design information and typical operating temperature and pressure of the equipment.
The installation was completed in 2014, and a failure occurred two years later. FIGS. 2 and 3 present the location of the failure in the field. The unit was shut down and the failure investigated.
Failure investigation. The first inspection was a liquid penetrant inspection when the equipment was open. FIG. 4 presents the cracking morphology. It was observed that the cracks were located at a 6 o’clock orientation and the cracking was continuously present from the 6 o’clock to the 9 o’clock position. A sample was taken at the leak location (6 to 7 o’clock position). It is important to mention that most cracks were in the base metal and almost none were in the weld area.
It was possible to take a surface deposit sample when the exchanger was opened. This sample was analyzed by an energy dispersive X-ray spectroscopy (EDS) and provided the composition presented in TABLE 2.
The sample contained high concentrations of chloride, potassium, iron, oxygen and sodium. The sources of those contaminants will be reviewed later in this article.
The first evaluation was to confirm that the shell material met the fabrication drawing specifications. TABLE 3 provides the chemical composition of the shell sample using spark optical emissions spectroscopy. The analysis provided the information that the shell material met the stainless-steel 316L composition.
A microscopic examination was then completed using a stereomicroscope. FIG. 5 shows the location where the crack that caused the leak was sampled, while FIG. 6 presents the fracture facies. The presence of beach marks can be observed, which is a sign of fatigue phenomenon. Two small ratchet marks show the initiation location of the crack, and the yellow arrows show the propagation direction.
A scanning electron microscope (SEM) examination was performed on the same location where the macroscopic beach marks were present. FIG. 7 shows the images of the SEM observation. The beach marks are present, and it is also possible to see the propagation orientation. In addition, transgranular cracking morphology is present (FIG. 8).
An EDS microprobe installed with the SEM was used to complete a semi-quantitative analysis to evaluate the contaminants in the failure zone; FIG. 9 provides the spectrum obtained by this analysis. It is possible to see that chloride, oxygen, potassium and calcium elements are again present.
Then, a microscopic examination was completed following ASTM E3-2011 practice.6 The samples were given full metallographic preparation and examined before and after etching with modified glyceregia. This etching solution was selected in relation to reveal the carbides at the grain boundaries, general grain structure and ferrite stringers.7
FIG. 10 shows an austenitic structure with a small presence of chromium carbides at the grain boundaries. In addition, transversal metallography was completed at the crack location. FIG. 11 shows transgranular cracking.
Based on all observations, it is possible to see that a fatigue phenomenon caused a crack propagation. The beach marks show the presence of this damage mechanism. Also, considering the high presence of chloride in the shell deposit, the chloride concentration in the crack analysis completed with the EDS sensor with the SEM and the transgranular morphology of the crack, it is a possibility that chloride stress corrosion cracking (CL SCC) was present. The chromium carbides did not appear to influence the cracking mechanism.2,8
Depending on the fatigue phenomena (mechanical or thermal fatigue), the morphology of cracking can cause a single crack or multiple cracks—the crack morphology can also be transgranular. At this point, it is unclear if CL SCC initiated the cracking or not.9
Additional investigation. Following the failure event, a fast repair was completed, and the unit was put back in service—it required replacement one year later with the original design. The temperature profile of the shell exchanger was further investigated by means of a thermography scan of the uninsulated shell. FIG. 12 provides the thermography image that was completed.
With this inspection, a high delta temperature was observed on the shell. Depending on the location, the temperature delta between the hot and cold section was 200°F–300°F (93°C–149°C). At this temperature delta with a stainless-steel material, it is known that thermal fatigue activity can be aggressive.2,10–13
Also, the shell temperature on the barrel tubesheet section was 290°F–310°F (143°C–154°C). Considering the composition of the process and the operating pressure, it is possible to conclude that water was present in this dead zone of the exchanger. The estimated water dew point was ~400°F. The past inspection history of the exchanger also proved that water phase was present because there was an activity of carbonic acid corrosion on the shell. FIG. 13 provides a sketch to explain the phenomena. The interesting fact is that the failure happens directly at the temperature transition (cold/hot) presented on the thermography picture.
With the barrel tubesheet design acting like a dead zone and the inlet tube temperature, it is possible to condensate water in the shell side even if the operating temperature on the shell side is 600°F (316°C). The water accumulates and is exposed to the elevated temperature of the shell side. The evaporation of the water created a higher thermal gradient at the internal diameter (ID) surface. Therefore, the thermal gradient at the ID surface is higher than what the thermography reported. Thermal fatigue is well known in the industry at the thermal mix-point. Many studies have documented this phenomenon,2,10,11,14,15 including that a two-phase flow would cause more damage than a liquid-liquid or a gas-gas mix-point. With the water evaporation, this creates additional local turbulence; the phase transition also influences the local temperature, which then increases the thermal stress. It is also believed that in an exchanger where shock condensation is present, the same phenomena can happen.
With the thermography scan, it is possible to conclude that the high chloride concentration was built based on a condensation/evaporation phenomenon.
Sheldon9 presented the phenomenon that happens on a thermal mix-point. The article asserts that the cracking morphology can be misinterpreted to CL SCC rather than thermal fatigue. A testing protocol is presented to simulate the same environment as the thermal mix-point experimented by the authors. The results indicate the presence of thermal fatigue morphology, but CL SCC was not present. Abri, et al.16 did a failure analysis of an SS321 related to thermal fatigue. Again, the morphology of the failure crack showed a presence of transgranular with branch cracking morphology, but it was not related to CL SCC. Based on this literature, thermal fatigue can provide a similar CL SCC morphology. Furthermore, the bundle that is in SS304L did not experiment any cracking and the shell welds were not the location of the through wall failure. Typically, if CL SCC is present, this would happen in higher stress areas—this was not the case here. Also, if CL SCC was present, the tubes and barrel tubesheet should have failed or cracked.
It is difficult to rule out chloride cracking interaction in this failure case. There was a very high chloride concentration near the failure location; however, it is possible to conclude that the predominant mechanism is thermal fatigue.
Also, considering the temperature delta and the water presence that is evaporating, the original material design (1.25Cr-0,5Mo) would be susceptible to thermal fatigue based on API-574 Table 4 criteria.14 This type of design would still need to be inspected for cracking.
The question that all the readers would ask is, “Why choose solid stainless-steel shell for this environment?” Knowing this information, the decision would have been different. However, the situation and the validations that were made to complete this selection must be considered. Part 2 of the article (November issue) will review the design selection. HP
DISCLAIMER
Suncor Energy Inc. and its affiliates (collectively Suncor) do not make any express or implied representations or warranties as to the accuracy, timeliness or completeness of the statements, information, data and content contained in this paper and any materials or information (written or otherwise) provided in conjunction with this paper (collectively, the information). The information has been prepared solely for informational purposes only and should not be relied upon. Suncor is not responsible for and is hereby released from any liabilities whatsoever for any errors or omissions in the information and/or arising out of a person’s use of, or reliance on, the information.
LITERATURE CITED
American Petroleum Institute (API) Standard 663, “Hairpin-type heat exchangers,” 1st Ed., May 2014.
American Petroleum Institute (API) Standard 571, “Damage mechanisms affecting fixed equipment in the refining industry,” 3rd Ed., March 2020.
Chawla, S. L. and R. K. Gupta, Materials selection for corrosion control, ASM Intl. (formerly American Society of Metals), 1993.
Dillon, C. P., Corrosion control in the chemical process industries, 2nd Ed., MTI Publication No. 45, Materials Technology Institute of the Chemical Process Industries Inc., 1997.
Nickel Institute, “The role of stainless steels in petroleum refining: A designers’ handbook series,” Paper #9021, 2020, online: https://nickelinstitute.org/media/1781/roleofstainlesssteelinpetroleumrefining_9021_.pdf
ASTM International (formerly American Society for Testing and Materials) Standard E3-11, “Standard guide for preparation of metallographic specimens,” 2017.
ASM International (formerly American Society of Metals), ASM Handbook, Volume 9: Metallography and microstructures, 2004.
ASM International (formerly American Society of Metals), ASM Handbook, Volume 13: Corrosion: Fundamentals, testing and protection, 1987.
Dean, S. W., “Chloride SCC of stainless steel? No - Cyclic strain cracking!” September 2000.
Association for Materials Protection and Performance (AMPP – formerly NACE International) 34101, “Refinery injection and process mixing points,” March 2001.
Association for Materials Protection and Performance (AMPP – formerly NACE International) SP0114-2014, “Refinery injection and process mix points,” 2014.
American Petroleum Institute (API) Recommended Practice 574, “Inspection practices for piping system components,” 4th Ed., Novermber 2016.
American Fuel & Petrochemical Manufacturers (AFPM) HAZ002.00, “Hazard identification: Injection point and process mixing point hazards,” Annual Meeting, 2016.
Donnelly, C., K. Bagnoli, L. M. Gustafsson and A. Skoulidas, "Strain based modeling of thermal fatigue at mix points," NACE CORROSION 2012, Salt Lake City, Utah, March 2012.
Al Abri, N. and J. R. Nair, "Case studies of thermal fatigue damage in duplex and stabilized stainless steel," NACE CORROSION 2019, Nashville, Tennessee, March 2019.
ASM International (formerly American Society of Metals), Corrosion in the petrochemical industry, 2nd Ed., 1997.
François Gilbert has been a Material Engineer for the last 13 yr at the Suncor refinery in Montreal. He is responsible for the risk-based inspection program and participates in fitness for service, developing inspection plans, completing root cause failure analyses, completing corrosion studies and implementing/reviewing operational integrity limits. He is also involved in a group of material specialists covering all Suncor sites. Gilbert earned BS and MS degrees in material engineering, is certified in API-510, API-571 and API-580 standards, and is a Professional Engineer in the Quebec province.
Olivier Jacob has been a Unit Inspector for the last 7 yr at the Suncor Montreal refinery. He participates in developing inspection plans, evaluating equipment integrity and coordinating turnaround inspection teams. He has a professional degree in materials and welding and a BS degree in material engineering. Jacob is certified by API-510, API-570, NBIC-CI/R and is a Professional Engineer in the Quebec province.