a CDU overhead system. This effort, Phase 1 of a larger corrosion management programme, aimed to confirm cor - rosion mechanisms, quantify corrosion rates, and define integrity operating envelopes. Despite the use of standard mitigation strategies, includ- ing ammonia injection and boot water pH control, periodic inspections revealed localised thinning in vertical runs, elbows, and non-insulated lines. This prompted the deploy- ment of a digital twin built using an electrolyte-based dig- ital twin combining Petro-SIM and the OLI engine. This rigorous simulation environment modelled the overhead system using real operating data, including lab analyses, corrosion probe readings, historian values, chloride content, and ammonia injection rates. The objective of the CDU overhead corrosion monitoring digital twin was to continuously monitor and optimise key operating parameters that directly influenced corrosion risk. These parameters included several corrosion-relevant KPIs: • Ionic dew point temperature and pH indicate when and where the first acidic water may condense in the overhead vapour. Low pH at this point can cause severe localised corrosion. • Salting point temperature is when ammonium chloride or similar salts begin to crystallise. Falling below this point can lead to under-deposit corrosion. • Aqueous phase condensation and pH reflect the overall water condensation behaviour and acidity, which influence general corrosion risk throughout the overhead system. • Ammonia injection rate measures the effectiveness of HCl neutralisation. Inadequate dosing can lead to acid cor- rosion if HCl remains un-neutralised. • Boot water pH is a traditional field measurement of the water phase in the accumulator. While useful, this can be misleading unless contextualised by dew point and initial condensation conditions. • Wash water rate indicates if sufficient water is injected to dissolve deposited salts and prevent salt-induced corrosion. • Corrosion rate, from the OLI Studio, is calculated based on actual stream chemistry and metallurgy, providing a pre- dictive view of expected metal loss. These KPIs were tracked and visualised via Petro-SIM dashboards, which provided alerts when values deviated from safe limits. In parallel, corrosion rate estimates were calculated using the OLI application, based on stream chemistry and metallurgy, especially under conditions below the ionic dew point. Implementation and observed results The twin was configured using historical and real-time plant data, including chloride concentrations, neutraliser dosing, and field inspection results. Initial modelling sug - gested corrosion should be minimal, but refinement of the model revealed: Boot water pH was artificially elevated due to the influence of sour stripper water, giving a false sense of protection. Ammonia injection was below optimal levels , allowing un-neutralised HCl to persist in the vapour phase. This led
to a dew point pH as low as 2.9, triggering localised acid corrosion in early condensate, particularly at elbows and vertical pipe sections. Phase change corrosion (shock condensation) was identified as a likely mechanism in areas with sudden cool - ing or inadequate insulation. Corrosion rates were estimated to be in the range of 30-50 mil/year, consistent with field observations despite limited initial data. Armed with these findings, the refinery implemented tar - geted adjustments: • Ammonia injection rates were increased to better neu- tralise HCl before condensation. • Continuous wash water strategy was introduced to dis- solve salt deposits and reduce under-deposit corrosion risk. • Digital twin technology implemented by engineers to simulate ‘what-if’ scenarios, testing crude blends, injection rates, and temperature adjustments before making real- world changes. Outcomes and benefits The deployment of the corrosion digital twin led to clear, measurable improvements: • Enhanced predictive accuracy : The model aligned closely with field inspection data, enabling early warning months in advance. • Faster root cause analysis : Engineers quickly identified low dew point pH and shock condensation as key drivers, without needing shutdowns or invasive investigation. • Improved process control : Operators maintained safe margins above dew and salting points, guided by real-time KPI tracking. • Reduced corrosion and downtime: Probe data showed lower corrosion rates post-implementation, and no unplanned overhead failures occurred. • Cost avoidance : Avoiding a single major failure prevented millions in expensive repair and lost production, supporting broader industry findings that predictive corrosion monitor - ing can cut maintenance costs by up to 35%.3 Beyond technical improvements, the digital twin increased visibility and trust in the corrosion control strat- egy. Operators and engineers used the dashboard as a daily tool, treating it as an extension of their digital control and historian systems. The success of this deployment val- idated the value of real-time, simulation-driven corrosion monitoring and established a framework for expanding digital twin applications across other units. Conclusion The deployment of a corrosion digital twin in CDU overhead systems marks a significant shift from lagging, periodic monitoring to predictive, continuous control. By combining rigorous chemical simulation with real-time operational data, refiners can pre-empt corrosion, protect equipment, and optimise operations. This approach aligns with broader digital transformation goals in refining, offering a scalable, proactive strategy to improve safety, reliability, and efficiency. The corrosion dig - ital twin not only delivers immediate value but also paves
50
PTQ Q3 2025
www.digitalrefining.com
Powered by FlippingBook