INFERENTIAL CONTROL OF NARROW CUT DISTILLATION COLUMNS
GDS is a semi-rigorous short cut model of a distillation tower section.The model is suitable for a large variety of distillation columns, from simple debutanizers & depropanizers to aromatics superfractionators to alkylation unit isostripper, and many others.
GDS estimates product purities as function of
Temperatures, especially tray temperatures
Pressure
L/V
Tray efficiency
Typical debutanizer configuration
Example debutanizer bottom GDS
Four unknowns: Xbot4, Xbot5, Xbot6, Xbot7
Equation 1.Bottom mass balance
Σ (Xboti) = 1
Equation 2.Bottom equilibrium
Σ (Xboti * Kboti) = 1
Equation 3.Section 1 separation
Σ (RNi * Xboti) = 1
Equation 4.Tray equilibrium
Σ (RNi / Ktrayi * Xboti) = 1
Coefficients Kboti, Ktrayi, RNi, are calculated by a thermodynamic model
This results in four linear equations, solved every minute.
Debutanizer inferences vs. Lab
Depropanizer example trend
GDS Features
Responsive to column feed composition changes.
Distillation column feed typically comes from upstream equipment, and feed quality varies. Composition change causes column temperatures to shift, and GDS, reading the new conditions quickly corrects its inferences. There is no need for the operator to input any data. Detection of feed changes is completely automatic.
Easily understood.
The model follows standard API procedures for heat balance, temperature / pressure correction and distillation thermodynamic calculations. Process engineers can easily understand and make model corrections when necessary.
Simple to calibrate.
GDS is calibrated by running historical plant data through the model and comparing the results against historical lab data. Calibration, in the way of adjusting one or two coefficients is simple.
Reliable enough to replace distillation analyzers.
GDS inferential precision is similar to analyzer repeatability, but at a fraction of the cost. That said, GDS can also work with analyzers, resetting a bias in the calculation via a dead time compensator.
Calculates vapor and liquid traffic and related constraints.
Able to infer flooding and/or L/V for fractionation efficiency. Flooding is a useful inference because alternative ways of detecting flooding by pressure difference is essentially a “post mortem” measurement.
Multi variable compatibility.
The inferential calculations can be used by DCS controllers as advanced regulatory application or integrate with MVC (multi variable control) packages available on the market.
GDS Benefits
Repeatability is similar to that of onstream analyzer at a fraction of the cost
The model has no significant dead time, making it easy to use in distillation applications with complex interactive dynamics
Understood by process engineers, it is not a black box
Useful process calculations, e.g., internal vapor and liquid flow calculation
Reduced lab support requirements
Reference literature:
APC of debutanizers make use of inferential predictions, Hydrocarbon Processing Journal, June 2017, later presented at IRPC conference, September 2018 2017_LOR_DeC4_HPJ
Prediction for APC in refinery gas plant, PTQ magazine, gas issue 2016. 2016_GasConPaper_PTQ
First-principles inference model improves deisobutanizer column control, Hydrocarbon Processing Journal, March 2003. 2003_DIB.pdf
First principles distillation inference model for a toluene – xylene separation column. ERTC Computer Conference, June 2002. 2002_Toluene_Inference_ERTC.pdf
First-principles distillation inference models for a superfractionator product quality prediction. Hydrocarbon Processing Journal, February 2002.HP0202BenzeneColumn.pdf
Experience with GDS, a first principles inference model for distillation columns. Presented ERTC Computer Conference, June 2001 and NPRA computer conference, October 2001, published in Petroleum Technology Quarterly, Autumn 2001. 2001_GDS.pdf
Simulation based inferential controls. Paper presented at the AICHE spring conference 1995, Houston TX. 1995_Inferentials_AIChE.pdf
