Chapter17To define a convergence order:1.From the Data menu, point to Convergence, then Conv Order.2.Select a block from the Available Blocks list. Use the arrow to move the blockthat you want converged first to the top of the Convergence Order list.3.Select any other blocks that you want in the orderand movethem to theConvergence Order list. You can use the up and down arrows to rearrange theorder within the list. The first convergence block is converged first and nestedmost deeply.Specifying the Calculation SequenceYou candefinethe calculation orderfor all orpart oftheflowsheet.You supply anID for each sequence.To define a sequence:1.From theData menu,pointto Convergence, then Sequence.2.In the Object Manager click the New button.IntheCreateNewIDdialogbox, enter anID oraccept thedefault IDand3.click OK.Specifythe calculation sequence on the Specifications sheet. On each row of-the sheet, you can enter one ofthe following:The beginning of a loopThe end of a loopA block IDA sequence ID for part of a flowsheetFor the beginning and end of a loop, specify Begin or Return To in theLoop-Return field. Specify the block type in the Block Type field. The followingblocks begin loops:ConvergenceSensitivityDataFitFortranblocks can introduce loops onlyfor the special case of loop controlFortran blocks.Specifytheblock type andblock IDforthefollowingblocktypes:UnitoperationEquipmentUtilityTransfer17-17AspenPlusUserGuideVersion 10.1-0
Aspen Plus User Guide 17-17 Version 10.1-0 Chapter 17 To define a convergence order: 1. From the Data menu, point to Convergence, then Conv Order. 2. Select a block from the Available Blocks list. Use the arrow to move the block that you want converged first to the top of the Convergence Order list. 3. Select any other blocks that you want in the order and move them to the Convergence Order list. You can use the up and down arrows to rearrange the order within the list. The first convergence block is converged first and nested most deeply. Specifying the Calculation Sequence You can define the calculation order for all or part of the flowsheet. You supply an ID for each sequence. To define a sequence: 1. From the Data menu, point to Convergence, then Sequence. 2. In the Object Manager click the New button. 3. In the Create New ID dialog box, enter an ID or accept the default ID and click OK. 4. Specify the calculation sequence on the Specifications sheet. On each row of the sheet, you can enter one of the following: • The beginning of a loop • The end of a loop • A block ID • A sequence ID for part of a flowsheet For the beginning and end of a loop, specify Begin or Return To in the Loop-Return field. Specify the block type in the Block Type field. The following blocks begin loops: • Convergence • Sensitivity • Data Fit Fortran blocks can introduce loops only for the special case of loop control Fortran blocks. Specify the block type and block ID for the following block types: • Unit operation • Equipment • Utility • Transfer
ConvergenceFortranBalancePres-ReliefFor economic calculations,specifyEconomicforblock type.There is no blockIDforeconomiccalculations.Within a sequence you can insert a subset of the flowsheet that already has anID and a defined sequence. For large flowsheets it is useful to build up thesequence specification inthismanner.SpecifySequence intheBlockTypefield.Specify the sequence ID for the subset in the Block ID field.AspenPlus executesthesequencesexactlyas youenterthem,withtheseexceptions:If youAspen PlusCheck the Check Sequencefield ontheChecks whether all loops in a sequence are torn. If a loop is not torn,ConvOptions Defaults Sequencing sheetAspenPlusdisplaysanerrormessageSpecify Execute Before or Execute AfterinaInserts the Fortran block into your sequenceFortran blockAutomatically generates convergence blocks for design specifications andSpecify a Design-Specinserts them into your sequenceUsing Initial GuessesFor many simulations with recycle streams, initial guesses for the tear streamswill help convergence.This is especially true for recycle systems with closed loopsor recirculating solvent loops.You can often provide a reasonable initial guessfrom your knowledge of the process or through a simple mass-balancecalculation.The sequence is displayed in the left pane of the Control Panel. If the left pane ofthe Control Panel is empty, select Step from theRunmenu.Enter initial compositions and flow rates for the tear streams on StreamsSpecification sheets, and run the simulation. Or select your own tear streamsusingtheTearsheet,andprovideinitialestimatesforthem.17-18AspenPlus UserGuideVersion 10.1-0
17-18 Aspen Plus User Guide Version 10.1-0 Convergence • Fortran • Balance • Pres-Relief For economic calculations, specify Economic for block type. There is no block ID for economic calculations. Within a sequence you can insert a subset of the flowsheet that already has an ID and a defined sequence. For large flowsheets it is useful to build up the sequence specification in this manner. Specify Sequence in the Block Type field. Specify the sequence ID for the subset in the Block ID field. Aspen Plus executes the sequences exactly as you enter them, with these exceptions: If you Aspen Plus Check the Check Sequence field on the ConvOptions Defaults Sequencing sheet Checks whether all loops in a sequence are torn. If a loop is not torn, Aspen Plus displays an error message Specify Execute Before or Execute After in a Fortran block Inserts the Fortran block into your sequence Specify a Design-Spec Automatically generates convergence blocks for design specifications and inserts them into your sequence Using Initial Guesses For many simulations with recycle streams, initial guesses for the tear streams will help convergence. This is especially true for recycle systems with closed loops or recirculating solvent loops. You can often provide a reasonable initial guess from your knowledge of the process or through a simple mass-balance calculation. The sequence is displayed in the left pane of the Control Panel. If the left pane of the Control Panel is empty, select Step from the Run menu. Enter initial compositions and flow rates for the tear streams on Streams Specification sheets, and run the simulation. Or select your own tear streams using the Tear sheet, and provide initial estimates for them
Chapter17Flowsheet SequencingThe tearing and sequencing of a flowsheet is complex and can require user input.ThefollowinginformationoninteractingwiththeAspenPlussequencingalgorithmis intended for advanced users. It is recommended that other users accept thedefaultsequencing.Aspen Plus initially tears and sequences flowsheets in this sequence:1.The information flow (incidencematrix)of unit operation blocks,Fortranblocks,design specifications,constraints,optimizations,and cost blocksiscollected.2.Sequences you specify are checked for possible missing tears and are used togenerate a reduced incidencematrix.Inthereduced incidencematrix,subsequences you specify are collapsed and treated as a single block.3.Thereduced incidencematrix ispartitioned into independent subsystemsthat can be solved sequentially.4.Tear streams or Fortran block tear variables are determined for eachsubsystem,takinguser-specifiedTear,TearVariable,andConvergencespecifications into consideration. The automatic sequencing algorithm inAspen Plus selects tear streams by minimizing a weighted combination of thenumber of:TearvariablesTimes loops are tornAn initial sequence is determined as part of the tearing.For each subsystem,5.Convergence blocks are created for design specifications, tear streams, andtearvariablesthatarenotconvergedbyuser-specified convergenceblocks.Specifying Design Spec Nesting as Inside on the Convergence ConvOptionsDefaults Sequencing sheet generates one tear convergence block for all tearstreamsandtearvariables,andgeneratesanindividualdesignspecificationconvergence block for each design specification. See Specifying SequencingParameters, this chapter, for more information.Youcan affect theautomatic sequencing algorithm by:AdjustingtheVariableWeight and Loop Weight parameters ontheConvergence ConvOptions Defaults Sequencing sheet.Specifying initial estimates for possible tear streams on the Streams forms.The specifications for non-feed streams are used as initial guesses if possible.Streams with data are weighted in the sequencing algorithm, so they aremorelikelytobeselectedastearstreams.Specifying tear streams directly, using the Tear Specification sheet.Youshouldbecarefulnottospecifymoretearstreamsthanrequiredforconvergence. You can underspecify the number of tear streams, andAspen Plus will determine the additional tear streams needed.17-19AspenPlusUserGuideVersion 10.1-0
Aspen Plus User Guide 17-19 Version 10.1-0 Chapter 17 Flowsheet Sequencing The tearing and sequencing of a flowsheet is complex and can require user input. The following information on interacting with the Aspen Plus sequencing algorithm is intended for advanced users. It is recommended that other users accept the default sequencing. Aspen Plus initially tears and sequences flowsheets in this sequence: 1. The information flow (incidence matrix) of unit operation blocks, Fortran blocks, design specifications, constraints, optimizations, and cost blocks is collected. 2. Sequences you specify are checked for possible missing tears and are used to generate a reduced incidence matrix. In the reduced incidence matrix, subsequences you specify are collapsed and treated as a single block. 3. The reduced incidence matrix is partitioned into independent subsystems that can be solved sequentially. 4. Tear streams or Fortran block tear variables are determined for each subsystem, taking user-specified Tear, Tear Variable, and Convergence specifications into consideration. The automatic sequencing algorithm in Aspen Plus selects tear streams by minimizing a weighted combination of the number of: • Tear variables • Times loops are torn 5. An initial sequence is determined as part of the tearing. For each subsystem, Convergence blocks are created for design specifications, tear streams, and tear variables that are not converged by user-specified convergence blocks. Specifying Design Spec Nesting as Inside on the Convergence ConvOptions Defaults Sequencing sheet generates one tear convergence block for all tear streams and tear variables, and generates an individual design specification convergence block for each design specification. See Specifying Sequencing Parameters, this chapter, for more information. You can affect the automatic sequencing algorithm by: • Adjusting the Variable Weight and Loop Weight parameters on the Convergence ConvOptions Defaults Sequencing sheet. • Specifying initial estimates for possible tear streams on the Streams forms. The specifications for non-feed streams are used as initial guesses if possible. Streams with data are weighted in the sequencing algorithm, so they are more likely to be selected as tear streams. • Specifying tear streams directly, using the Tear Specification sheet. You should be careful not to specify more tear streams than required for convergence. You can underspecify the number of tear streams, and Aspen Plus will determine the additional tear streams needed
ConvergenceObtaining FinalConvergence SequenceTo obtain the final convergence sequence:1.All convergence blocks are ordered as they appear on the ConvOrderSpecificationformandthecurrentsettingofUsernestingontheConvOptions Defaults Sequencing sheet.Blocks not mentioned on theConvOrder Specification form are ordered according to the setting of DesignSpec Nesting on the ConvOptions Defaults Sequencing sheet and the span ofconvergence blocks in the initial sequence.2.AspenPlus obtains thefinal convergence sequenceby repeatedlyremovingtears and/or design specificationsfrom the outermost convergence block andpartitioning the reduced flowsheet.3.For Design specification nesting as Inside or Inside simultaneous on theConvergenceConvOptionsDefaultsSequencingsheet,youcandefineuserspecifiedconvergenceblocksfordesignspecifications,andtheywillbeinsertedautomaticallyintothesequence.Adding SpecialOptions to the SequenceAt the end of the final convergence sequence, special options are added:1.Blocks with Execute options are inserted into the sequence.2.Sensitivity,Balance,andDataFitblocks notalreadyin the sequenceareinserted.Because adesign specificationloop usuallyhas a small span,the sequencingalgorithm does not nest them (for example, a tear loop outside and manyindependent design specification loops inside).Since the algorithm does nottakenumericalvaluesintoaccount,itsometimesplacesdesignspecificationloopsinsidetearloopswhentheywouldperformbetteroutside.SpecifyingDesignSpecNesting as Outside on the Convergence ConvOptions Defauits Sequencing sheetwould alter the sequence but this often leads to deeply nested iteration loops forlarge flowsheets.Viewing the SequenceTo view the sequence along with the tear streams and the convergence blocksdetermined by AspenPlus:>From the View menu, click Control Panel.The sequence is displayed in the left pane of the Control Panel.If the leftpane of the Control Panel is empty,select Stepfrom theRun menu.17-20AspenPlusUserGuideVersion 10.1-0
17-20 Aspen Plus User Guide Version 10.1-0 Convergence Obtaining Final Convergence Sequence To obtain the final convergence sequence: 1. All convergence blocks are ordered as they appear on the ConvOrder Specification form and the current setting of User nesting on the ConvOptions Defaults Sequencing sheet. Blocks not mentioned on the ConvOrder Specification form are ordered according to the setting of Design Spec Nesting on the ConvOptions Defaults Sequencing sheet and the span of convergence blocks in the initial sequence. 2. Aspen Plus obtains the final convergence sequence by repeatedly removing tears and/or design specifications from the outermost convergence block and partitioning the reduced flowsheet. 3. For Design specification nesting as Inside or Inside simultaneous on the Convergence ConvOptions Defaults Sequencing sheet, you can define userspecified convergence blocks for design specifications, and they will be inserted automatically into the sequence. Adding Special Options to the Sequence At the end of the final convergence sequence, special options are added: 1. Blocks with Execute options are inserted into the sequence. 2. Sensitivity, Balance, and Data Fit blocks not already in the sequence are inserted. Because a design specification loop usually has a small span, the sequencing algorithm does not nest them (for example, a tear loop outside and many independent design specification loops inside). Since the algorithm does not take numerical values into account, it sometimes places design specification loops inside tear loops when they would perform better outside. Specifying Design Spec Nesting as Outside on the Convergence ConvOptions Defaults Sequencing sheet would alter the sequence but this often leads to deeply nested iteration loops for large flowsheets. Viewing the Sequence To view the sequence along with the tear streams and the convergence blocks determined by Aspen Plus: ➤ From the View menu, click Control Panel. The sequence is displayed in the left pane of the Control Panel. If the left pane of the Control Panel is empty, select Step from the Run menu
Chapter17Sequencing ExampleThis example describes steps to converge a simple flowsheet that does notconverge properly when automatic sequencing is used.It illustrates:Supplying initial estimates forrecycle streamsAltering the calculation sequencewithDesign specnestingAltering the calculation sequence with Conv OrderAdjustingtolerancestoaccountfornested loopsTheflowsheet consists of:BlocksTypeTOPCOL,BOTCOLInterconnected columnsHEATERPre-heaterCOOLERProduct coolerAtiProcessFlowsheetwindowC0区口P-TOPCOLCTOPCOLFEED>M-F-TOPCOLHEATERF-COOLERNPRODCOOLERREFLUXBOTCOLF-BOTCOLA>The mass flow of stream REFLUX,the inter-reflux stream fromBOTCOLtoTOPCOL, is manipulated to meet a purity specification of component THF instreamPROD.PRODisaproduct streamfromBOTCOL/COOLER,indesignspecification THF.PSPEC is the convergence block defined to converge THFAspenPlusUserGuide17-21Version10.1-0
Aspen Plus User Guide 17-21 Version 10.1-0 Chapter 17 Sequencing Example This example describes steps to converge a simple flowsheet that does not converge properly when automatic sequencing is used. It illustrates: • Supplying initial estimates for recycle streams • Altering the calculation sequence with Design spec nesting • Altering the calculation sequence with Conv Order • Adjusting tolerances to account for nested loops The flowsheet consists of: Blocks Type TOPCOL, BOTCOL Interconnected columns HEATER Pre-heater COOLER Product cooler The mass flow of stream REFLUX, the inter-reflux stream from BOTCOL to TOPCOL, is manipulated to meet a purity specification of component THF in stream PROD. PROD is a product stream from BOTCOL/COOLER, in design specification THF. PSPEC is the convergence block defined to converge THF