How to use a computer process simulator to model a candidate distillation process
nSet up topographical flow sheet model
nSelect components from library, or create new components or pseudo components
nSelect VLE correlation options
nRetrieve VLE BIPS from library
nEvaluate TPXY relationships calculated from the selected VLE model
nAdjust BIPS or in-fill using UNIFAC until all binary pair TPXY are close to the truth
nInput temperature, pressure, state, & quantity data for feed streams, cut streams, side draws
Set up model constraints & input equipment & convergence specifications
nDefine equipment:
nFlash blocks for setting feed stream conditions, mixers where multiple feeds go to one column, feed preheaters, product coolers, interchangers
nDefine distillation column fixed parameters:
nNumber of stages, feed & product locations, condenser & reboiler type, column pressure profile
nDefine distillation column calculation constraints:
nOne constraint will determine the column bottom condition (use fixed reboiler heat input for the first trial run) & another constraint will determine the column top condition (use fixed reflux ratio for the first trial run)
nSet convergence criteria & damping factors
Provide estimated distillation column internal flow & temperature profiles
nColumn flow & temperature profiles must be provided, for the computer simulator to have a starting point for its calculations
nUsually it is sufficient to input a reasonable top & bottom temperature for the column, & reasonable top & bottom molar flows, telling the simulator program to interpolate & fill in intermediate stage data
nAfter one converged run has been made on the system you are studying, successive runs usually will converge more readily if the column profile computed during a previous run is used as the starting point, rather than the coarsely interpolated initial estimated profile
Run initial simulation, adjusting input parameters until one run converges with reasonable results
nIf your first set of input parameters runs to convergence with reasonable results (rather than zero overhead vapor flow or other correct but trivial converged solutions) you are very lucky:
nYou input a good guess for the initial tower profile
nIf your first run does not converge, heed the warnings given in the simulator output
nIt is very easy to under estimate the heat input required, or to broadly miss the correct column flows, on the high or low side, so change the input & try again
nIt is also very easy to mistake the engineering units used by the simulator model for other similar units, so when the process simulator results seem to be off by a large or familiar factor (like 1.8, 2.2, 454, or 1000), check the units, reinput, & try again
After one simulation has converged: Evaluate, adjust, rerun
nAfter one simulation has converged, evaluate the output product streams quantity & quality, & energy used
nAdjust the primary column input parameters to improve deficiencies in quantity or quality:
nMove the feed point up or down, increase the number of theoretical stages
nIncrease the reboiler heat input & increase the reflux ratio to improve product purity, but reduce the heat input & the reflux ratio to decrease hot & cold utility usage
nDo not be afraid to make many variations & run the simulator over & over again, to see what happens if this or that increases or decreases
nLimit the size of step changes made to input parameters, to increase the chance of new runs converging successfully
The opportunity window for optimization opens now, & closes when the final simulation run becomes the design basis for column, internals, & overall distillation system design
Make use of the process optimization opportunity now
nNow that the simulation gives close to the desired result, it is time to run more simulations in a parametric block
nOne parameter to reduce is the number of theoretical stages - try 10%, 20%, 50%, 80% fewer!
nAdjust the feed point because the result may surprise you:
nTry moving the feed up & down because either may yield improved purity
nStep all side draws up & down in position & in flow rate
Make use of the process optimization opportunity now
nIn the optimization runs, permit the more advanced column top & bottom convergence specification criteria their chance to decrease energy input or reflux quantity:
nTry using purity specifications or flow quantity specifications instead of heat input & reflux ratio as in the initial runs
nRun column tray or packing sizing software on these optimization run outputs to provide physical size input for cost estimation
nThe optimal design will balance fixed & variable life cycle costs
Column internal selection:
Which trays or which packing?
nUnless low pressure drop (< 0.1 psi per stage) is necessary, for column diameter > 2 feet (500 mm), consider valve trays first
nIf pressure drop of near 0.01 psi per stage is required, in non-aqueous low pressure systems, consider structured packing first
nFor low pressure drop in aqueous systems, consider random packing first at or above atmospheric pressure, consider structured packing first under vacuum
nFor high vacuum conditions (1 to 20 mmHg absolute) consider woven wire mesh packing first, knit wire packing second
Column internal selection:
Which trays or which packing?
nValves for valve trays are available round or rectangular, with directional patterns such as checkered, basket weave:
nAll will work, with variations in cost & maximum & minimum rate at a given diameter & tray spacing key among selection criteria
nFloating valves act as check valves on the tray deck, reducing liquid weeping at low vapor rates, improving performance during turn down
nPunched raised louvers, also called fixed valves, resist fouling & are mechanically very strong, good when turn down is not important
Rate the physical distillation column just selected for alternative operations
nOften a distillation system will be used for more than one process
nAfter simulating the selected primary process & optimizing the physical distillation column for that process it is time to simulate alternative processes
nIf the required alternative processes can be performed in the physical system selected for the primary process no design modifications will be needed, but their process information should appear in the design basis because it is not certain which system will demand the highest duty or dictate maximum size of the reboiler, condenser, coolers, pumps, instruments, or auxiliaries
nIf the alternative process requires more separation stages, a different feed point, different decanter or reflux provisions, or much higher or lower energy input, incorporate this information into the design basis specification, in addition to the complete primary design
Example of process design for a
multi-purpose distillation system
nThe following slides portray the computer simulation output from rating a newly designed distillation system to process five different feed streams on a campaign basis
nThe column & all of its auxiliaries were built with all features needed to handle all of these streams
nThe maximum design case for the column was for the most difficult fractionation:
nAcetone dehydration
nThe reboiler design considered two maxima:
nPeak heat input (thus peak heat flux) during methanol recovery, & lowest available temperature difference during toluene recovery
nThe condenser design considered two maxima:
nPeak cooling demand with methanol, & lowest temperature difference during methylene chloride recovery