April 2019

Heat Transfer

Assess remaining life for heater tubes

Refineries and petrochemical plants form some of the major sections of the hydrocarbon industry and consume much of the fuel that is burned in furnaces (heaters and specialty furnaces, such as reformers and crackers) to meet process heating requirements.

Patel, R., Singh, J., Ashutosh, Engineers India Ltd.

Refineries and petrochemical plants form some of the major sections of the hydrocarbon industry and consume much of the fuel that is burned in furnaces (heaters and specialty furnaces, such as reformers and crackers) to meet process heating requirements.

Heaters consist of three major components: the steel structure, the refractory, and the tubes and pipes that carry the process fluid. The steel structure is durable and can generally remain in service for 30 yr–40 yr without any major maintenance (e.g., painting or some other relatively minor repair that does not call for major cost, or that is normally less than 1% of the total heater cost). Refractory repair and replacement in heaters is part of regular maintenance checks, and generally repairs are required between turnarounds every 3 yr–4 yr. An analysis to evaluate the economy of replacement vs. fuel savings to compensate heat losses due to a damaged refractory has been conducted,1 and the optimum replacement period has been suggested as 5 yr–6 yr.

The lifetime of heater tubes varies from 5 yr–15 yr, depending on the material of construction and the severity of conditions they are subjected to during heater operation. The approximate cost ratios for heater tube materials—i.e., CS/AS/SS/special alloy for unit length—can be derived as 1/5/10/21 considering carbon steel (CS) cost as unity.2 This work lists the major parameters that affect the remaining life of heater tubes—or the time to reach allowable minimum wall thickness—and elaborates on the understanding of the existing provisions of heater design codes.3 It includes references of past works in condensed form on the subject that can be used for remaining life calculations, with the necessary precautions.

Equipment

Process heaters are used to provide requisite heat to the feed in an enclosed box (box or cylindrical). They are internally lined with refractory using single and multiple burners where fuel is fired for generating heat. The process feed flows inside the tubes and pipes and picks up heat from the hot flue gases generated from burners, mainly through radiation and/or convection. Definitive guidelines are available for most process services regarding the calculation of tube thickness and the selection of tube material of construction (MOC). FIGS. 1, 2 and 3 show typical heater and process coils.

FIG. 1. Typical fired heater.
FIG. 1. Typical fired heater.
FIG. 2. Horizontal radiant tubes.
FIG. 2. Horizontal radiant tubes.
FIG. 3. Vertical radiant tubes.
FIG. 3. Vertical radiant tubes.

There has always been an inherent competition to achieve increased user acceptance for any product, and that can be achieved in two basic ways: offering the same quality (service life and minimum specification requirement compliance) for a lower cost, or by offering improved quality (improved service life or improved specifications) at the existing cost.

To extend heater tube life, improved exotic metallurgies are being evolved almost every year with higher allowable stresses for higher temperature applications. However, the development of new, exotic MOCs only will not ensure the optimized and economic utilization of any tube in service until methods for estimation for the reliable remaining tube life are in place. This estimation becomes critical for the following cases, as well:

  • The correct estimation of remaining tube life will help ensure safe heater operation for a predefined period, eliminating:
    • Unplanned shutdowns and loss of profit, production, personnel and property
    • Cost (including opportunity cost) of material and resource mobilization required to restore the heater in service.
    • For pyrolysis furnaces, mechanical issues have been identified as one of the top seven causes for loss of ethylene production.4
    • Refiners in certain geographic regions and countries that do not have tubes readily available for unplanned replacements depend on imports, which may be a long process and requires planning for procurement for even low-grade materials. It becomes imperative to depend on a remaining life estimation of the tubes in various heaters to ensure safe operation until the next planned turnaround in the absence of the availability of the required tube spares.

Based on the time required for tube procurement for the next planned shutdown, maintenance personnel prefer first to “re-rate” the heater based on the available life of the heater tubes. This not only buys time until the next planned shutdown with the heater under optimum operation, but also ensures the safety of the system and personnel during heater operation by either restricting the operating parameters before the next planned shutdown, or by scheduling the next shutdown with the present heater operating parameters.

Cost factors and impact

The approximate initial cost ratios for radiant tubes vs. heater cost for heater duty and tube MOC for typical process heaters (FIG. 4), reformers and crackers in an Indian refinery have been tabulated in TABLES 1, 2 and 3.

FIG. 4. Graphical representation of % cost data, coil vs. heater systems.
FIG. 4. Graphical representation of % cost data, coil vs. heater systems.

The cost of a refinery shutdown due to a heater-dependent outage and the subsequent rebuilding may range from few days to several weeks, depending on the extent of damage. For a few cases, losses due to heater outage and rebuilding for units that are solely dependent on heater operation are also calculated (to be read in the context of the Indian refinery):

  • For a typical naphtha hydrotreater (NHT)/continuous catalytic reforming (CCR) unit with a capacity of 4.5 MMtpy, repair costs range from $150,000–$200,000. Operational losses can be $7 MM/d–$10 MM/d.
  • For a typical vacuum gasoil (VGO)/catalytic hydrotreating (HDT) unit with a capacity of 3 MMtpy, repair costs range from $300,000–$400,000. Operational losses can
    be $2.5 MM/d–$3 MM/d.
  • To determine the losses, the following assumptions were considered:
  • Suitable buffer based on typical intermediate product storage
  • Appropriate number of days, taking notice of the general plant shutdown/startup time involved
  • Plant turnaround to match the buffer, considering the number of shutdown days for repair work
  • To arrive at total repair work figures, approximately 10% of radiant coil replacement has been considered along with necessary repair work, including emergency manpower and machinery deployment, as required.

Coils are critical for heater operation and, based on the above data, it is evident that a heater outage caused by general coil failure can cost significantly more that its own material cost. Any operator decision to replace or retain the coil becomes a choice of “cost vs. catastrophe,” and the need for establishing optimum utilization to safely extract maximum tube life becomes pertinent.

Angle of impact

FIG. 5. Angle of impact.
FIG. 5. Angle of impact.

The simplest way to demonstrate the impact of any event for any engineering execution is to draw the angle of impact regarding the stage (phase) (FIG. 5) of identifying and correcting the flaw, which decides the magnitude of its impact in terms of cost and time.

Although the angle of impact is constant, the magnitude of impact (in terms of cost and time) varies significantly depending on whether the flaw is determined and corrected at design stage (minimal), construction stage (significant) or at post-commissioning stage (catastrophic). The guiding curve may or may not be linear in actual cases.

For heater tube failure cases, due to thickness loss, the first two stages are already completed; the intent is to avoid the catastrophic impact of tube failure by determining the safe operation life with maximum integrity operating window (IOW) parameter limits. Experience has shown that out of a production-first mentality to achieve higher crude processing rates, process heaters are subjected to hard firing that results in overshooting safe operating limits.5

Parameters affecting remaining tube life

Many variables directly and indirectly affect remaining tube life and are linked to the severity of operating conditions and fluctuations within the tubes. Although the main parameters are temperature, pressure and variations such as severity and cyclic changes, the life of the tube is affected by the phenomenon occurring both on the tube’s external surface (in contact with flue gas) as well as its internal surface (in contact with process fluid—i.e., feed).

External surface interactions (flue gas interactions) include:

  • Fuel-fired
  • External corrosion, oxidation, scaling, sulfidation (due to flue gas, high temperature, presence of excess or insufficient oxygen, etc.)
  • Other metallurgical interactions, such as carburization, decarburization, spheroidization and grain growth
  • Flame impingement, hot spots
  • Burner firing.
  • Internal surface interactions (feed interactions) include:
  • Type of feed (process fluid inside the tube)
  • Internal corrosion and erosion (fluid composition, temperature, velocity, tube MOC)
  • Sulfidic corrosion, naphthenic acid corrosion
  • Coking
  • Decoking/pigging.

Tubes can also fail due to design, operation and maintenance issues (MOC suitability for the service and design conditions, proper expansion gaps, etc.). However, it is assumed for the purpose of this article that the designers have handled these issues.

FIG. 6. Coil sagging.
FIG. 6. Coil sagging.
FIG. 7. Flame impingement.
FIG. 7. Flame impingement.
FIG. 8. Coke laydown and thinning.
FIG. 8. Coke laydown and thinning.

The effects of heater operation beyond IOW parameter limits, such as operating above permitted temperatures, excessive thermal fatigue/thermal shock during startups, process upsets resulting in creep failure (identified by sagging, bowing or bulging, as shown in FIG. 6); and other deterioration mechanisms, such as liquid metal cracking and embrittlement, polythionic acid stress corrosion (PTASCC), metal dusting, flame impingement (FIG. 7), hot spots and coking also impact the local tubes thickness (FIG. 8).

The remaining tube life under consideration focuses on the general phenomenon (mainly corrosion and erosion) that affect the complete tube life over an entire service period. Local tube failures (part length or in pieces) can normally be remedied with the help of mandatory spare heater tubes. Recently, there have been developments in the form of online spray methods that help eliminate heater tube surface deposits and scales and reportedly can lower tube surface temperatures for the same duty.6

Assumptions and limitations in theoretical estimation

Annexure A of API Standard 530 mentions that the formulae used to calculate the thickness for new tubes can “also be used to help establish operating skin tube metal temperature (TMT) limits and answer rerating and retirement questions about operating tubes.”

Theoretical estimations must be guided by the consideration of ideal conditions by making certain realistic assumptions to arrive at the desired value. Some of these assumptions and limitations of the code calculation procedures are:

  • The allowable stresses are based on a consideration of yield strength and rupture strength; plastic or creep strain were not considered. No considerations are included for adverse environmental effects, such as graphitization, carburization or hydrogen attack.
  • The design procedures have been developed for seamless tubes. They are not applicable to tubes that have a longitudinal weld.
  • These design procedures have been developed for thin tubes (tubes with a thickness-to-outside diameter ratio, δ min/Do, of less than 0.15).
  • No considerations are included for the effects of cyclic pressure or cyclic thermal loading. Limits for thermal stresses are provided in Annexure C. Stresses imposed by tube/fluid weight, supports, end connections, etc., are not discussed.
  • The procedures have been developed for systems in which the heater tubes are subjected to an internal pressure that exceeds the external pressure.
  • Transient or other nontypical events are not captured in the assessment (since these events are obviously not normal practice, unplanned and impossible to predict) and will affect correct estimation of life based on the gross parameters, such as average operating temperature and pressure (i.e., assumptions of constant operating conditions during any cycle).
  • For these estimates, it is assumed that the outside diameter remains constant. In an actual case, neither the operating pressure nor the metal temperature are uniform. Nonetheless, for the calculation, they are assumed to be uniform during each period.

Because of the uncertainties involved in these calculations, decisions about tube retirement should not be solely based on the results of these calculations. Other factors, such as tube thickness or diameter-strain measurements, should also be considered in decisions about tube retirement.

Past feedbacks on remaining tube life estimation

The API Standard 530 code statement that “design life is not necessarily the same as the retirement or replacement life” indicates that both can be different. Further, refiners have noticed that actual operational results and theoretical estimation procedure results do not converge initially for the remaining life assessment (RLA). Most likely, the theoretical output has yielded highly conservative results to ensure the safe operation of heaters.

Three primary areas of uncertainty exist in these calculations:

  1. It is necessary to estimate the accumulated tube damage (the life fraction used up) based on the operating history—i.e., the influence from the operating pressure, the tube-metal temperature and the corrosion rate of the tube. The uncertainties in these factors, particularly the governing temperature, may have a significant effect on the estimate. For example, heater designers in the 1960s initially used to consider a high margin for design tube metal temperatures as 200°F over the operating value. This margin has gradually reduced to approximately 25°F (as per the latest API Standard 530 code) over the past few decades, based on actual operating feedbacks.
  2. Knowledge of the actual rupture strength of a given tube MOC is ever-changing due to database updates based on operational feedback. The data of allowable stresses/strengths of material incorporated in codes are based on certain tests, and a change in material strength will result in variations in the estimated remaining tube life. These strength data are dependent on the tube suppliers and are continuously updated based on feedback and technological advances. For example, the rupture allowable stresses for P9 MOC have increased up to almost 140% from the previous version of API Standard 530. On the other hand, the rupture allowable stresses for stainless steel SS347 have been reduced 67% from the previous levels in the latest revisions.
    These are tabulated in TABLES 4 and 5 and FIGS. 9 and 10.
  3. API Standard 530 emphasizes the linear damage rule as per Annexure G of the code, although it also mentions that the limitations of the hypothesis are not well understood. Simplification for the wide variations in tube temperature range and rate of corrosion have been proposed in the form of equivalent temperature concept and corrosion fraction.
FIGS. 9 and 10. Graphical representation of variations in allowable stresses for P9 and SS347.
FIGS. 9 and 10. Graphical representation of variations in allowable stresses for P9 and SS347.

The uncertainty during the initial stages of a theoretical evolution can be noticed in a typical case of a crude heater in a Middle East refinery designed in the late 1950s. The crude heater was designed with carbon steel tubes (based on the 1958 API Standard 530) and revalidated in 1978 (based on the 1978 version of API) for more than 200,000 hr of tube life in place of the initial design life of 100,000 hr, since the heater had already been running for 20 yr. The remaining life-check calculations were then validated by other designers based on the remaining thickness and actual rate of corrosion from the operator’s in-house data. Since the tube deterioration had not happened as expected in the earlier design, the coil design temperature allowance was further reduced by 50°F (from 1,150°F to 1,100°F) during revalidation in 1978, based on actual damage rate feedback.

Improvements incorporated in estimation procedures

Initially, no provision existed for corrosion fraction in the previous versions of API Standard 530. However, since the actual operation feedback continued to show longer running lengths in the creep rupture range than those estimated by the code procedures with a corrosion fraction of 1, the guideline framers were prompted to reexamine and update the existing code provisions. This was done in the 1988 version of API Standard 530 by inclusion of the corrosion fraction graph for various cases to bridge the gap between code-predicted life of the tubes and actual operational feedback for various metallurgies.

Improved understanding of the rupture phenomenon for heater tubes and its calculation method has helped in the derivation of the corrosion fraction as described in Annexure G API Standard 530. It is recognized in this derivation that stress is reduced by the corrosion allowance; correspondingly, the rupture life is increased. The method takes into consideration the effects of stress reductions produced by the corrosion allowance.

Under the assumption of constant temperature, the rate of using up the life increases as the stress increases. In other words, the tube lasts longer if the stress is lower. If the tube undergoes corrosion or oxidation, then the tube thickness will decrease over time. Therefore, under the assumption of constant pressure, the stress in the tube increases over time. An integral of this effect over the life of the tube was solved graphically in the 1988 edition of API Standard 530 and developed using the linear-damage rule. The result is a nonlinear equation that provides the initial tube thickness for various combinations of design temperature and design life.

Present code provisions

The present corrosion allowance provision (as per API Standard 560) in heater tubes does not vary with regards to the corrosion rate and design life of the tube. However, a fixed minimum corrosion/erosion allowance value for various groups of MOC is recommended for designing and tube thickness calculations, as per API Standard 560 para 7.1.2.

The design life for the heater coils is normally 100,000 hr. The corrosion allowance is defined as being equal to the corrosion rate times the design life. Therefore, it is obvious that under present provisions, a certain margin is available in new tube adopted thickness on due to this fixed minimum corrosion allowance, which will result in lower than calculated design stresses in the beginning and facilitate longer actual creep life for the coil. The corrosion is measured in mm/yr or mills (1/1,000 in./yr) and can fall in the range of 0.5 mpy–5 mpy or higher.

Contribution of physical tests and measurements

Including the first inspection during equipment construction as per API RP 573, all inspections are recorded and compared with the preceding inspection for the same specific purpose. The determination of the physical condition, and the rates and causes of deterioration in the various components makes it possible to schedule repairs or replacements prior to compromising mechanical integrity and resulting failure. This is also applicable to a heater tube.

API Standard 530 Annexure A further states that “assessment of inspection data is collected in accordance with API RP 573 and API 570 and, assuming the normal or worst-case conditions, may be used to quickly assess the fitness for service of individual tubes. It is recommended that tubes, return bends or coil sections that fail the fitness for service assessment be further evaluated by performing a rigorous Level 1 or 2 assessment of metal loss and/or creep damage following the standard provided in Parts 4, 5 and 10 of API 579-1/ASME FFS-1. Tubes that pass this evaluation approach should also pass the rigorous API 579-1 assessment.”

Value additions in-house operating data

Refiners usually maintain a record for active and potential deterioration mechanisms based on shutdown or online measurement techniques, as well as previous operating history for any heater that helps in predicting a future deterioration rate. With decades of experience, refiners are building their own furnace corrosion/tube reduction rate database, which can be referenced during the basic calculations for RLA in case no major changes exist in the operating conditions/profile of the subject furnace. A common, shared database for such operating details can also be a useful tool for all refiners and help in arriving at a best remaining tube life.

Takeaway

Annexure A and Annexure G of API Standard 530 deal with guiding procedures for the prediction of the remaining life of a tube that has undergone stresses in the creep rupture range. Answers to the following questions can be found utilizing these provisions with a fair degree of accuracy:

  • What is the estimated life at a given operating pressure, metal temperature and corrosion rate?
  • For a specified operating pressure and corrosion rate, what temperature limit should be imposed for the tube to last a minimum time period?
  • How much should the operating pressure or metal temperature be reduced to extend the expected life by a given percentage?

Guiding calculation procedures in API Standard 530 can be converted into ready-made worksheets for the purpose of estimation of answers to these questions.

Operational history is divided into time periods during which the pressure, metal temperature and corrosion rate are assumed constant. For each of these periods, the life fraction used up is calculated. The sum of these calculated life fractions is the total accumulated tube damage. The fraction remaining is calculated by subtracting this sum from unity. Finally, the remaining life fraction is transformed into an estimate of the expected life at specified operating conditions. The extent of margin for conservative calculation of life is considered by selecting the desired operating parameters, as suggested in API Standard 530 Annexure A.3 (i.e., the maximum/normal operating pressure or maximum/normal temperature). Computer programs in VC++ or simplified Excel spreadsheets that are prepared following the code provisions for an estimation of remaining tube life are very useful.

A life management strategy for a typical gas heater tube has been presented with useful guidelines.7,8,9 This can be further simplified. MICA is an easy-to-remember approach toward a remaining tube life estimation:

  • Monitor and record the measurements (online and offline during shutdown)
  • Interpret and analyze the data, identifying the deterioration mechanism
  • Calculate as per the code provisions utilizing data and interpretations
  • Adjust or adapt the shutdown frequency of possible operating parameters.

Considering the high severity of heaters, the recommended offline inspection frequency can be considered as 3 yr, or 50% of the predicted remaining life, whichever is less. HP

LITERATURE CITED

  1. Bueno, A. and V. Regueira, “Energy losses from the radiant firebox wall of pyrolysis furnaces,” Hydrocarbon Processing, April 2018.
  2. Farrar, J. C. M., The alloy tree—A guide to low-alloy steels, stainless steels and nickel-base alloys, Woodhead Publishing Ltd. and CRC Press LLC, 2004.
  3. “Calculation of heater-tube thickness in petroleum refineries,” API Standard 530, 7th Ed., (including Annexures A and G), April 2015
  4. Cagnolatti, C. L., “Top seven causes for lost olefin production,” Hydrocarbon Processing, April 2015.
  5. Dutta, H., “Building blocks of process safety,” Hydrocarbon Processing, October 2018.
  6. Kumar, R., “Reported reduction in tube metal temperature in CDU heater application claiming up to 100°C,” Refining India.
  7. Vicente, F., “Defining the optimal life management strategy for gas heater tubes,” Inspectioneering Journal, May/June 2013.
  8. Upadhyaya, K., P. Haribhakti, J. Patel and V. K. Bafna, “An integrated approach for RLA of reformer tubes by NDT (ARTiS),” https://www.ndt.net/article/apcndt2013/papers/186.pdf
  9. Singh, S. P., “Remaining life assessment of outlet radiant heater tube of naphtha cracking unit—a case study,” CORCON 2017, Mumbai, India.

The Authors

Related Articles

From the Archive

Comments