PTQ Q3 2025 Issue

Digital twin corrosion monitoring for CDU overhead systems

Integrating first-principles simulation with plant data equips engineers with an early warning system and actions to avoid risk and damage

Jagadesh Donepudi, Michelle Wicmandy, and Ashok Pathak KBC (A Yokogawa Company)

C orrosion in the overhead systems of crude distilla- tion units (CDUs) is a persistent and costly challenge for refiners. These overhead lines and condensers are regularly exposed to acidic components like hydrogen chloride (HCl) and ammonium chloride (NH₄Cl) salts, which can lead to aggressive corrosion if improperly managed. The shift towards processing higher-sulphur, opportunity crudes and blends has exacerbated the problem by intro- ducing more chlorides and corrosive compounds into the system.1 To address this corrosion problem, refiners are turning to digital twin technology – a virtual model that mirrors plant operations and updates in real time. In refining, the adop - tion of digital twins is growing as companies aim to improve safety, predictive maintenance, and asset reliability. The impact of corrosion is more than operational; it is also financial. According to the National Association of Corrosion Engineers (NACE), corrosion costs the oil and gas industry $1.372 billion annually, which includes $589 million for surface pipelines and facilities, $463 million for downhole tubing, and $320 million for corrosion-related capital expenditures.2 With effective corrosion manage - ment and predictive monitoring, corrosion-related costs could be cut by 15-35%.³ Safety risks Beyond financial loss, safety risks are also significant. Leaks in CDU overhead systems can release hazardous hydro- carbons or acids, leading to fires or personnel exposure. Traditional spot-checking methods such as monitoring boot water pH are often insufficient. This single metric can be misleading. For instance, corrosion has occurred even when boot water was maintained at pH 7, because early condensation events can produce localised, highly acidic droplets that go undetected (as HCl dissolves faster than neutralisers).4 Likewise, ammonium chloride salts may deposit in cooler areas above the monitored zones, trigger - ing under-deposit corrosion that escapes early detection. Traditional corrosion monitoring in CDUs tends to be lag - ging, discontinuous, and localised, leaving the refiner vulner - able to unexpected failures. Traditional methods in CDUs, such as corrosion coupons or manual ultrasonic thickness measurements, often provide only intermittent snapshots

of asset integrity. These techniques detect damage after it has already occurred (lagging). Data is collected at discrete intervals rather than in real-time (discontinuous). In addition, measurements represent conditions at only a few points, which may not reflect broader system health (localised). This reactive approach increases the risk of unexpected failures, unplanned shutdowns, and safety/environmental incidents. Consequently, there is growing interest in pre - dictive, continuous corrosion monitoring systems that can: • Offer real-time insights into corrosion rates and mechanisms. • Integrate with process conditions (for example, tempera - ture, flow, and sulphur content). • Support predictive maintenance through data analytics and machine learning. • Enable better risk-based inspection (RBI) strategies and asset life extension. Traditional corrosion monitoring in CDUs tends to be lagging, discontinuous, and localised, leaving the refiner vulnerable to unexpected failures This has created a strong demand for a predictive and continuous monitoring solution. Within this context, a cor - rosion-focused digital twin for the CDU overhead system offers a compelling opportunity to enhance reliability. By integrating first-principles simulation with plant data, such a system equips engineers with early warnings and actions to avoid risk and damage. Problem Traditional approaches to corrosion monitoring in the CDU overhead system often rely on periodic measurements such as corrosion coupons, electrical resistance probes, ultra - sonic thickness surveys, and lab analyses. These methods provide only periodic snapshots of corrosion activity. By the time thickness reduction or high iron content is detected,

PTQ Q3 2025