Modeling Helps Reduce Corrosion in Refineries

Shutdowns caused by corrosion can occur due to buildups in the columns.

Corrosion is a significant challenge for refiners in downstream operations. A company’s objectives include maintaining production, increasing yield, and growing profits, but doing this successfully involves understanding the reason for—and extent of—corrosion. This is a considerable problem in crude tower overheads. One of the key factors in causing this corrosion in crude distillation unit overheads is hydrochloric acid (HCl).

“In 2003, Saudi Aramco initiated a study to define the cost of corrosion throughout core operations, focusing plant, engineering, and research investment in corrosion control to the areas that had the largest economic impact on corporate performance. For Saudi Aramco’s five domestic refineries, 36% of maintenance budget was due to corrosion.”1

“Corrosion is kind of like the silent killer in the refining, chemical, and petrochemical industry in general,” says Ezequiel Vicent, senior application engineer at OLI Systems, Inc. (OLI) (Houston, Texas, USA). “It’s not a fast process—it’s not something that starts showing up today and tomorrow you have to bring down the unit for shut down. It can happen in as quick as a couple months, but it can also happen over years or even decades.”

Buildups in the Columns

Shutdowns caused by corrosion can occur due to buildups in the columns. “The organizations and companies think, ‘That column will last four or five years,’ and it’s actually gone within a year,” says Rasika Nimkar, client success consultant at OLI (Vancouver, British Columbia, Canada). “There’s a lot of investment and planning that goes into purchasing or replacing a column or repairing a column, specially repairing column internals, like trays. Replacing a column could cost upward of $500 milllion, including but not limited to parts, labor, peripherals, geographical location, etc. for a 100 kbpd refinery.”

Nimkar emphasizes that to do so, everything has to be shut down, and all the processes that revolve around it need to be halted. A shutdown could last for up to 14 days. In addition, plants may have to pay premiums for replacement trays and internals to ensure that the parts that need replacement can be made available as fast as possible to avoid extending the shutdown. In addition, if a part is 16 weeks out (which is considered fast!) a plant may have to run at a diminished capacity due to corrosion damages until they can shut down the unit, often operating at losses. This can mean millions of dollars in unrealized profits left on the table.

“Productivity and reliability are foundational to success,” notes Nimkar in an article.2 “More than ever, companies are striving to adopt the latest operating methods and technologies in order to enhance downstream processes.”

New Downstream Technologies

What do these technologies include? Thermodynamic modeling software for one. Software, such as that developed by OLI, including its Mixed Solvent Electrolyte (MSE) model and Aqueous (AQ) model, allows users to better predict where corrosion might occur. Rather than making an educated guess about the location of corrosion based on experience, modeling software can enable operators to make better decisions regarding prevention of corrosion. Instead of simply having a physical picture of what is taking place in a refinery, modeling software also provides a picture of the chemistry component. This can be validated through simulation.

Nimkar elaborates, “Once crude oil is transported downstream, it is run through a desalter to remove salts, solids, and other impurities; however, it is difficult to remove all contaminants, like salts (NaCl [sodium chloride], MgCl2 [magnesium chloride], and CaCl2 [calcium chloride]—these salts hydrolyze in the furnace to form HCL) and H2S [hydrogen sulfide], which then flow into the tower. Although hydrochloric acid and other corrosive liquids are common in refining processes, many companies lack the water chemistry analysis capabilities to identify this type of corrosion, as well as the precise location, cause, and treatment.”

The means of mitigating corrosion in refineries has evolved over the years. It initially involved the use of coupons, which were examined to determine the amount of metal lost over a particular length of time. These are still used today, but now, additional, newer physical technologies can help capture the thickness of the pipe or the metal, such as spot x-ray checks or permanently installed monitoring systems. However, notes Vicent, “You still have to guess where the corrosion is happening—you’re still using your experience to determine where you think corrosion will occur.” Now, industry-leading modeling and simulation software helps customers understand their unique processes, industry conditions, and chemical behaviors.

“Rigorous modeling and simulation tools enable companies to visualize their operating environments, providing precise measurements, real-time intelligence, and predictions to achieve their desired results,” Nimkar explains. “Ionic modeling is a groundbreaking development in modeling and simulation technology that leverages more comprehensive water chemistry analysis to provide even greater results. This software provides in-depth insights into the behavior of chemicals at every stage of production, including some conditions that are tough to observe.” These conditions include water dew point, ionic dew point, and salt point.

Benefits of Modeling Data

Operators may opt to change the overhead temperature in order to limit corrosion (Figure 1). However, by modifying the temperature, operators can also shift products. For example, a refinery that supplies jet fuel (kerosene) could increase the crude tower overhead temperature to save itself from corrosion (avoiding ionic slurries or salt deposition), but make less jet fuel (a more expensive product) and more gasoline (a less expensive product). In this instance, a tradeoff is made. An operator must choose whether it makes more financial sense to stop fouling or corrosion, or allow it to continue, but put an inspection program in place. By using data provided by the modeling, operators can make decisions about how to best avoid, prevent, and mitigate corrosion, plus increase their bottom line.

FIGURE 1 Overhead corrosion prediction in a crude distillation unit.

Ionic modeling gives “specific answers that they would not be able to get through regular modeling because not all the software is capable of water chemistry modeling,” Nimkar says. “That makes it very unique and distinguished.” The ability to predict the presence of ionic liquid through electrolyte modeling is exclusive to OLI.

Vicent and Nimkar confirm that OLI is on the brink of an innovative new feature in thermodynamic modeling. “Right now, we can tell you where a corrosive environment will form, but we cannot tell you what the actual corrosion rate will be,” says Vicent. However, that is expected to change within the next 18 to 24 months. “We’re working on bringing corrosion analyzer into MSE,” Vicent explains. “Not only are we going to be able to tell you where the corrosive environments form, but also the actual corrosion rate that we can calculate for that system.” It is something they have been working on for quite a while and are excited to bring to the market.

Through continued effort, such as the creation of what Nimkar calls “a massive private databank of amines used to neutralize impurities in process streams,” and a joint industry project with OLI and major oil refineries, data was collected to enhance corrosion prediction capabilities. These software improvements will, Nimkar believes, empower “downstream companies to optimize refining processes, select the right materials, increase efficiency and throughput, save money, and avoid catastrophic failure.”


1 “Cost of corrosion in oil production and refining,” ResearchGate web site, June 2006, (Nov. 13, 2020).

2 “Minimizing corrosion with ionic dew point process modeling to optimize downstream refinery operation,” OLI Systems Inc. web site, July 13, 2020, (Nov. 5, 2020).

Related Articles