Using Simulation as a Tool for Coker Troubleshooting and Predictions – the KBC approach

Presented By

Ranjit Bhise - KBC Advanced Technology Private Limited

Conference: Mumbai 2016

Delayed Cokers are the preferred technology for converting low-value heavy residue to middle distillates. Cokers are tough to operate and difficult to predict in a dynamic refinery operating scenario. Refinery planners find it difficult to ascertain Coker yields, distillate impurities, operating conditions compromise on a daily basis and matching them with refinery schedules.

Optimization of Coker operating parameters such as ‘combined feed rate’, heater skin temperatures and run length, drum maximum fill height utilization, cycle time variation, and continuous push for higher feed rate are just a few of a refiner’s concerns.
Solutions can easily be found when unit process engineers make extensive use of simulation tools. With time, simulation tools have developed to become efficient and very informative. KBC DCSIM™ is one such tool that answers most of the questions related to improving DCU performance.
DCSIM is a steady-state first-order reaction kinetic model in the family of reactor kinetic models under the KBC PetroSIM™ simulation platform. It is a well-proven simulation tool extensively used by our clients. It is backed by the operating knowledge base of our consultants with years of field experience.
The model precisely calculates process impact of variation of independent variables on dependent variables such as feed quality, and changes in operating parameters such as yields, and product properties at equilibrium state. In the process, it does rigorous heat and material balance including fractionator flash zone, so as to converge in order to satisfy unit equilibrium and provide vital prediction results.
Feedstock to Coker can be individually characterized on 50°F boiling range pseudo-components that help in distributing the yield shift between furnace, drum liquid, drum vapor and quench zone. Coking rates are calculated based on feed Conradson Carbon content, cracking vs. coking balance in terms of competing reaction rates inside the drums.

Drum dimensions as input data to the model help in estimating drum fill time, coke maximum height, drum vapor velocities and ‘Cfactor’ for the drum. These numbers form the basis of understanding drum limitations in terms of pushing more feed (unit feed rate), asphaltene carryover into distillate draw, and processing of lighter vs. heavier feeds The DCU drum kinetic model can be coupled with a detailed Coker fired heater model to precisely deliver heat balance, pressure drop numbers, coke buildup in tubes (thickness) and percentage vaporization across individual tubes. In doing so it can predict the rate of increase in tube metal skin temperature over heater run length. This becomes a very powerful tool for predicting time between tube spalling and decoking, so as to avoid unit run length surprises and capacity constraints.

Once calibrated and set up, the DCSIM reactor model becomes ready to predict the relationship between variation of independent and dependent unit parameters. The predictions are easier to work and compare between DCSIM Simulation and the Excel ® reactor workbook environment that is typical to the KBC SIM tools.
Active usage of DC-SIM can save refiners millions of dollars in terms of constrained unit capacity due to heater downtime and/or timely identification of predicted product qualities.

Benefits arising out of predicting and implementing workarounds for constraints using the Simulations model are plenty. A couple of examples:

  • At one of the refinery sites in India, DC-SIM was utilized to optimize drum operating pressure so as to minimize asphaltenes carryover in HCGO distillate draw. This helped in maintaining downstream hydrocracker operations, reducing issues associated with their feed filters and in turn increasing unit reliability. The accrued refinery-wide benefit for a period of six months was approximately $2025 million.
  • Optimization of velocity steam in the Coker furnace helps to reduce recurring utility costs. Based on unit capacity the savings can be anywhere between $0.1–0.5m/y. Additional benefits post optimization can be reduced pressure drop through coils, marginal improvement of fractionator overhead cooling, etc. This requires a solid base heater model that can be put to test through predictions. DC-SIM has this capability.
  • Variation in ‘Combined Feed Ratio’ for the unit and its positive effects on coke deposition inside heater tubes can be correctly predicted using DC-SIM and the Coker furnace model combination. This can help planners to extend scheduled spalling and decoking cycles based on the overall refinery turnaround plan.

The presentation of this paper will consist of the following key aspects:

  • Short introduction to KBC
  • DCSIM as simulation too
  • Coker heater modelling
  • Case studies:
    • Short synopsis on implementing model prediction output on HCGO quality by changes in unit operating pressure
  • Short Heater run length and velocity steam relationship
Refining Community