An Introduction to Fouling in Fired Heaters – Part 1 | Interior Fouling of Heater Tubes

By Erwin Platvoet, Chief Technology Officer

Part 2 & 3 up next: Exterior and Burner fouling!

Introduction

Fouling is the accumulation and formation of unwanted materials on the surfaces of processing equipment. It is an extremely complex phenomenon and considered the major unresolved problem in heat transfer (Bott, 1995). Fouling in refineries and petrochemical plants has an impact on safety, reliability, operation, the environment and profitability. The cost of fouling is believed to be more than $2 billion per year in U.S. refineries alone (source: www.energy.gov) due to increased production costs, production losses, unit shutdowns and high maintenance costs. Fouling is further exacerbated by continued increase in use of heavy, unconventional oil sources, deeper residue conversion to light ends, tightening environmental demands and fuel standards and increased production complexity.

The goal of this article is to provide an overview of the most common types of fouling in fired heaters and preventive strategies. It addresses areas inside the tubes, on the outside of radiant tubes, convection section, burners, air preheaters and selective catalytic reduction (SCR).

Fouling is generally classified by six categories.  The first three are commonly found in fired heaters and related equipment.

Type Definition
Particulate fouling Accumulation of suspended particles in the liquid or gaseous process streams onto the heat transfer surfaces. Main drivers are concentration of particles and fluid flow velocity. Example: fouling of convection banks due to dust and soot.
Reaction fouling Chemical reactions on the heat transfer surfaces. The surface material itself is not a reactant but can be a catalyst. Main driver is temperature. Example: thermal cracking of hydrocarbons in the fluid film.
Corrosion fouling Accumulation of corrosion product on the heat transfer surface. The heat transfer surface material itself reacts to produce the corrosion products. The products of corrosion can cause particulate fouling downstream.
Precipitation fouling Crystallization of dissolved salts from saturated solutions due to solubility changes with temperature. Example: scaling in boiler tubes.
Biological fouling The attachment of living materials to the heat transfer surface.
Solidification fouling Freezing of high-melting components from a multi-component solution on subcooled surfaces. Example: wax formation on subcooled surfaces.

 

Fouling Inside Fired Heater Tubes

Particulate, reaction, and corrosion fouling occur inside fired heater tubes. The most common one is reaction fouling in the form of coking, which is driven by temperature, residence time, velocity, and feed composition. The heaters that are the most prone to this type of fouling are the ones that process crude feeds, due to the wide range of components they require. This is especially true for refineries that turned to alternative heavy feedstocks that are more cost-effective to process. For example, bitumen (asphalt) is a highly viscous semi-solid form of petroleum. Canada has the largest reserve of natural asphalt (“tar”) in the tar sands, which is a combination of clay, sand, water, and bitumen. Tar sands can be mined and processed to extract the bitumen, which is then refined into oil.

The first problem that occurs when heating Tar Sands Bitumen is the potential for particulate fouling by clay particles. Alumina Silicate clay particles that are normally dispersed in a colloidal system lose solubility and deposit upon heating. The silicate deposits have low thermal conductivities and form a significant resistance against heat transfer. This type of fouling usually occurs in the convection section or the top of the radiant section. Clay deposits can only be removed by pigging, not by spalling or steam air decoking.

This type of fouling is particulate fouling and therefore driven by concentration and velocity. To minimize particle deposits, it is recommended to keep cold oil velocity above 6 ft/s at a minimum and ideally above 10 ft/s. Consider adding velocity steam to the convection inlet to increase the tube side velocity.

A second problem is that bitumen typically contains 16 – 25% asphaltenes. Asphaltenes are heavy, polyaromatic molecules that contain sulfur, nitrogen, and heavy metals. Asphaltene molecular weight is in the range of 500 – 3,000 but the apparent molecular weight can be up to 300,000 due to association by polar constituents. Their weight and molecular structure make asphaltenes strong fouling precursors and must be kept in solution as much as possible to prevent excessive deposition.

On the other end of the spectrum, we have shale oil. The production of shale oil, also known as Light Tight Oil (LTO), has grown exponentially in the last 10 years. Shale oil has many features that are attractive to refiners; it is a light oil with a low viscosity, a low asphaltene concentration (typically less than 0.1 wt%), a low sulfur content, and due to its recent abundance, has become a very economic feedstock. Unfortunately, there are downsides to using LTO as well. It is highly paraffinic, with long chain alkanes of 20 to 50 carbon atoms). This has significantly increased the risk of wax deposition on cold walls of tanks and processing units.

Another downside for vaporization of LTO inside a fired heater is that its physical properties are very inconsistent. Day to day variations in density and solids content can be very wide, even for shale oil coming from the same basin. A high variability in vaporization potential can lead to excessive vaporization inside the tubes, which can lead to dry points. Dry points should always be avoided since they leave behind residue and cause excessive fouling.

Refineries in the US are not designed to process either heavy bitumen or shale oil. They are historically designed for medium crudes and cannot readily process very light or very heavy crudes without significant (and expensive) changes. For economic reasons, many refineries blend LTO and bitumen to achieve the characteristics of a medium type crude. This has introduced yet another fouling problem inside fired heaters. Blends of LTO with heavy asphaltenic crude can result in asphaltene instability and precipitation, resulting in a strong increase in coking rate when onset of precipitation occurs early. Crude, vacuum and delayed coker heaters that used to run for years on Arabian crudes without decoking now show runlengths of several months or less before the tubes are too fouled to continue. Studies show that the paraffinic character of LTO causes the ashpaltenes to lose solubility, an instability that is hard to manage with the daily changes in LTO composition, and even harder when sourcing many different heavy crudes to blend with the LTO. Blending will have to be done extremely carefully to maximize asphaltene stability. Toluene solubility tests according to ASTM D 7157 (“Standard Test Method for Determination of Intrinsic Stability of Asphaltene-Containing Residues, Heavy Fuel Oils, and Crude Oils”) help the refiner determine the optimum ratio of crude to shale. Since coking is a type of reaction fouling, it is strongly dependent on the process film temperature, which in turn is dependent on the incident flux profile from the flames.

See Figure 1 for two different kinds of flame interactions, leading to two very different incident flux profiles. The flames that merge (on the left side of the figure) create a long and uniform flux profile with a peak near the top of the firebox. The flame interactions on the right create a much more intense peak near the bottom of the heater due to flames impinging on the tubes. Note that the flux profiles are normalized and that the absolute flux values of the flames on the right are much higher.

Neither flame behavior is desirable; the long merging flames will lead to poor fuel efficiency, while the short impinging flames create hotspots and very high coking rates. The difference in peak film temperature between the two cases is well over 100°F even though they are the same heater and the same burner design. Flame behavior and flux profiles from burners can be studied using CFD and manipulated by changing burner type, quantity, and/or location. Figure 1 – Two types of flame interactions creating two distinctly different flux profiles

The Mechanism of Coking

There are two main types of coking mechanisms. The first one is catalytic coking which takes place at the tube wall itself. The shape (‘morphology’) of catalytic coke is filamentous, which means that a network of fine carbon threads is formed on the tube inner wall. Small metal particles can be found on the ends of these filaments. The process of catalytic coking is demonstrated in Figure 2.

Catalytic coke is formed by absorption and cracking of hydrocarbons on surfaces containing nickel. Hydrogen and solid carbon are formed in a reaction that is catalyzed strongly by nickel, and to some extent, by iron.
Some of the deposited carbon reacts with oxygen and steam to form CO. In ethylene plants the CO is a catalyst poison for downstream converters.
 

Some of the carbon diffuses into the material along intergranular boundaries forming chromium carbides, a process called carburization. It also precipitates on the backside of the nickel catalyst particle

 

The carbon precipitation grows into a filament. The catalyst particle at the end of the filament activates the filament into a site for “radical coking”. The filament growth continues as long as the catalyst particle remains uncovered.
The catalytic coke layer is rigid and branch-like in structure, creating “trapping” sites for other cokes particles
Solid particles like coke formed by the pyrolytic coking mechanism cover the catalytic sites and then takes over as the dominant mode of coke formation

Figure 2

 

Catalytic coke is the major form of coke formed in high temperature processes like gas cracking (ethane, propane) to produce ethylene. The radiant tubes in these heaters typically contain 35 – 45 % nickel. During operation, the catalytic coke layer continuously dehydrogenates and changes into a very hard graphite-like material that is difficult to spall and gasify. Due to its hardness and rigidness, it poses a risk of tube rupture during a thermal shock.

Pyrolytic coke, also called condensation coke, is softer and less structured than catalytic coke. It is formed in the bulk of the gas by several mechanisms, including dehydrogenation, polymerization, and condensation of aromatic and olefinic compounds. Pyrolytic coke is the major form of coke found in crude, vacuum, delayed coker heaters as well in liquid (naphtha, gas oil) crackers. It has an amorphous structure, is softer, spalls easily and fouls downstream equipment like transfer line exchangers.

Like any other kind of fouling, coking has a major impact on the heat transfer efficiency due to its low thermal conductivity. Coke thermal conductivity as a function of its porosity is illustrated in Figure 3.

 

 

 

 

 

 

 

 

 

 

Figure 3

Condensation / pyrolytic coke has a thermal conductivity in the range of 1 – 2 W/m*K (0.6 – 1.2 Btu/h°F*ft) whereas catalytic coke thermal conductivity is in the range of 3 – 4 W/m*K (1.8 – 2.4 Btu/h°F*ft) due to its structure and lower porosity. Compared with the conductivity this is an order of magnitude lower. The impact of coking on tube temperature can therefore be profound. For example, consider the outlet tube of a cracking coil with an absorbed heat flux of 50,000 W/m2. A coke layer of 5 mm thickness and thermal conductivity of 2 W/m*K will increase the tube wall temperature by

So, a modest coke layer of 5 mm (0.2 inch) thickness will increase the tube wall temperature by 125°C (225°F). A typical ethylene coil will see a temperature increase of 100 – 150°C between Start and End of Run when the tube has reached its allowable maximum temperature, all due to coking. This happens in a span of 30 – 90 days.

There are other issues associated with coking besides periodic decoking, like carburization, creep, and chrome depletion. These can all damage the tube, resulting in increased maintenance and replacement cost, and a risk of coil failure during operation if not properly mitigated.

Carburization: the process of carbon enrichment of the material and the subsequent formation of carbides. The presence of a coke layer greatly increases the rate of carburization. Carburization results in

  • Local volume increase, leading to tube bulging
  • Internal stresses due to volume increase, leading to intergranular cracking
  • Embrittlement, causing loss of thermal shock resistance
  • Reduced weldability
  • Rupture of the tube

 

 

 

 

 

 

 

 

 

Figure 4 – local tube bulging and cracking due to carburization

Creep: the elongation / stretching of a tube due to its weight. The creep rate depends on tube temperature, the load bearing cross-sectional area and the tube material. The presence of coke on the tube wall increases not only the tube temperature but also increases the total weight of the tubes dramatically. Deformation due to creep leads to creep voids inside the material:

 

 

 

 

 

 

 

 

Figure 5 – creep voids

Chrome depletion: the surface of the tube becomes depleted of chrome by the continuous process of chrome carbide formation and the removal of the chrome oxide layer. Over time this diminishes the capability of the tube to form new chrome oxide layers, which accelerates catalytic coking.

Tube failures: the combined effect of all the damage mechanisms can reduce the life of a tube to as little as four years in severe services and make it especially vulnerable to thermal shocks.

 

Factors That Impact Coking Rate

Temperature and velocity

Since coking is a reaction type fouling, the rate of coking strongly depends on temperature. A high process temperature

  • favors all reactions, coking included
  • produces more reactive species like dienes
  • increases diffusion of carbon into the coil material

In a crude heater, a 10 – 40° increase in outlet temperature can lead to 50 – 400% higher fouling rates.

The model of Ebert and Panchal was developed to predict fouling behavior of crude oils in heat exchangers. The combination of an Arrhenius type equation with a Reynolds dependent factor shows that heat transfer and process flow velocity play important roles in fouling:

In fired heaters, the process mass velocity in the radiant section should be kept in the range of 300 – 500 kg/m2s to limit fouling.

Feed composition

In cracking heaters, liquid cracking (naphtha, diesel, gas oil) produces more coke than gas cracking. Condensation coking is promoted when liquid feeds have a high End Boiling Point (EBP), also called a “heavy tail”.

In any service, coke formation is accelerated if the feed contains high amounts of asphaltenes, naphthenes (cyclo-alkanes) and aromatics.

Feed contaminants act as coking catalysts and change the coil surface characteristics:

Sodium: attacks the protective chrome oxide layer

Iron: coking catalyst

Potassium & Vanadium: attack radiant coil surface

Sulfur: the role of sulfur is complex. It promotes pyrolytic coke reactions but produces weaker coke structure. At the same time, sulfur passivates coil surface and prevents catalytic coking. At high levels it promotes carburization.

 

Steam

Steam is used to inhibit coke formation in various ways. It is used as ‘dilution steam’ in cracking furnaces to lower the hydrocarbon partial pressure and to reduce the residence time. Typical steam to oil ratios are 0.2 – 0.3 for gas cracking and 0.3 – 0.5 for liquid cracking.

In vacuum and coking heaters, steam is typically used as ‘velocity steam’. Typical steam quantity is 1 – 2% of feed flow rate. It is usually injected at the crossover to change the vaporization profile and to increase the process velocity.

Any steam in the feed also inhibits coke accumulation by gasifying the carbon that has been formed already.

Steam is sometimes used to passivate the coil surface prior to feed introduction by subjecting the coil to high temperature steam for several hours. A chrome oxide layer is formed under these conditions, which delays and prevents catalytic coke formation. This method is less effective for older coils that have been decoked several times and have a rough surface.

Decoking

The act of decoking also has an impact on coking rate. As mentioned, repeated cycles of coking and decoking (spalling, erosion) deplete the chromium content at the coil surface. Lack of chrome prevents the generation of protective oxide layers. The rough surface that is created by the spalling of coke has more sites for surface reactions.

There are three main decoking methods:

  1. Steam/air decoking
  2. Online spalling
  3. Mechanical pigging

Steam/air decoking is a process where a mixture of steam and air is introduced into the coils to gasify and burn the coke. It is typically applied in cracking furnaces (on-line), delayed cokers (off-line) and crude and vacuum heaters (off-line). The decoking process starts with high temperature steam to gasify the coke:

C + H2O  -> CO + H2

This reaction is endothermic, the process temperature must be higher than 1500°F (800°C) for the reaction to occur. This step makes the existing coke layer porous, which enables the next step of the decoking process. Air is added to oxidize the coke.

C + O2  -> CO2

This reaction is exothermic, so air must be added in a carefully controlled way to prevent a runaway reaction. The coke burn-off starts as soon as the oxygen and the coke meet, so this reaction proceeds from the beginning of the coil to the end. Coke ‘spalling’ is the fragmentation of the cokes layer into smaller particles. This part of decoking is an intentional part of the steam/air decoking process but needs to be controlled extremely carefully. The entire decoking procedure must be carried out carefully, to avoid

  • overheating the tubes, when the reaction proceeds too rapidly
  • breaking the coils, when thermal shocks cause the tubes to contract faster than the coke
  • erosion, due to excessive velocities of spalled coke particles
  • tube plugging due to excessive spalling
  • carburization, oxidation, tube bowing etc…

Online spalling is typically used in delayed coker heaters. This procedure uses high pressure (superheated) steam at high velocity (100 – 115 m/s). The intent is to vary the coil temperature to expand and contract the tubes to dislodge the coke. The benefit of this method is that it is faster and easier than steam/air decoke or off-line pigging, but the risk is that tubes can be plugged during excessive spalling.

Mechanical pigging is the process of propelling a “pig” through a coil with the help of a pig launcher. A pig is a studded device that travels through the pipe, driven by a motive fluid. Contrary to steam/air decoking or online spalling, there is no risk of overheating of overpressurizing the coils. The downside of this procedure is that it requires heater shutdown, and an external company is needed to execute the work. It is typically the most thorough decoking method unless tubes are oval. Likewise, special attention must be paid to heater coils with varying tube wall thicknesses.

Watch out for Part 2 on Exterior Fouling!  Have fouling related questions? Reach out to us at info@xrgtechnologies.com!

Please fill out the form below and we will be in contact as soon as possible.
GET IN TOUCH
Name(Required)
This field is for validation purposes and should be left unchanged.
Please fill out the form below and we will be in contact as soon as possible.
GET IN TOUCH
Name(Required)
This field is for validation purposes and should be left unchanged.
ERWIN PLATVOET
As CTO of XRG, Erwin is a true innovator, whose career spans more than three decades in heat transfer and combustion industries. Erwin is a graduate of Twente University in the Netherlands with a MS in Chemical Engineering. Erwin has served the industry around the globe in a variety of roles including Research and Development Engineer, Cracking Furnace Specialist, and Director of Engineering, and now CTO. Erwin holds eight patents in fired heat transfer and emissions control technology, has published numerous papers, and co-authored the John Zink Combustion handbook and Industrial Combustion Testing book. Erwin has been an active member of the API 560 and API 535 subcommittees and taken an active role in revising these standards.
BAILEY HENDRIX
Bailey graduated from Oklahoma State University with a Bachelor of Science in Mechanical Engineering. Upon graduation, she joined the private sector as an Applications Engineer in Tulsa, OK at a local combustion company where she managed the sales activities for the process burner refining market. She quickly accelerated her career, becoming the Refining Account Manager responsible for all business development and sales of process burners in North and South America. Her strong leadership skills and interpersonal qualities led her to a position as the Western Hemisphere Sales Director for the process burner business, leading a group of sales engineers in the areas of new equipment, retrofits and burner management systems. Her financial and commercial acumen drives the success of XRG Technologies’ business development.
ALLEN BURRIS
Allen’s background includes 10 years of experience in designing and selling process burners. Allen is a graduate of Oklahoma State University with a BS in Mechanical Engineering and is a licensed professional mechanical engineer in the State of Oklahoma. His knowledge and superior customer focus led him to a career change to process design, custom-engineered fired heater sales, and associated sub-systems for the petrochemical, refining and NGL industries. With more than two decades of experience in the combustion and fired heater industry, Allen has what it takes to overcome challenges associated with complex projects and possesses.
TIM WEBSTER
With over 25 years of experience in the combustion industry, Tim brings a wealth of industry experience and technical expertise to XRG. Tim graduated with a Bachelor of Science in Mechanical Engineering from San Jose State University and received a Master of Engineering from the University of Wisconsin. Tim began his career engineering custom combustion systems for a wide range of applications including boilers, heaters, furnaces, kilns, and incinerators. Tim is a licensed professional mechanical engineer in the states of California, Texas, Louisiana and Oklahoma, has authored numerous articles and papers, and has co-authored several combustion handbooks.
matt martin
As the Lead Scientist at XRG, Matt has over 30 years of experience in the combustion industry. He specializes in CFD of fired equipment, including UOP platforming heaters, burners in process heaters, thermal oxidizers and flares with over 300 simulations of installed, field-proven equipment. Matt received a Bachelor of Science in Computer Science with a minor in Mathematics from the University of Tulsa. He has written numerous publications, is listed as inventor or co-inventor on 27 patents and was awarded the title of Honeywell Fellow in 2011 for technical excellence and leadership.
gina briggs
Gina is a native Oklahoman and attended the University of Tulsa, graduating with a BSBA in Accounting. She is a Certified Public Accountant and Chartered Global Management Accountant. Gina began her career with the Tulsa office of Deloitte Haskins and Sells, providing audit and tax services. Since leaving Deloitte, she has held CFO positions with privately held companies in the manufacturing, construction and distribution industries. In 2013, she began a consulting practice providing contract CFO services to companies, one of which was XRG and joined XRG as CFO in 2019. Gina has always enjoyed working in the small business arena, helping business owners to profitably grow and manage their businesses.