Home

Introduction to FLUINT

1. EXIT LOSS INTO RESERVOIR CALL STEADY CALL FOR A STEADY STATE Submodel WALL is now awakened using the routine ADDMOD and a new BUILDF is issued to add CHAMBRS to the list of active models before a transient is initiated BUILDF CONF CHAMBRS CIRC CALL ADDMOD WALL S AWAKEN WALL SUBMODEL CHAMBRS FK100 0 5 INITIALIZE INLET LOSS ON FIRST PATH CALL TRANSIENT S START A TRANSIENT ANALYSIS For completeness the new OPERATIONS block is provided below HEADER OPERATIONS BUILDF CON CIRC BUILD CON1 WALL CALL DRPMOD WALL SUSPEND WALL SUBMODEL AS BOUNDARY DEFMOD CIRC FK26 1 0 S EXIT LOSS INTO RESERVOIR CALL STEADY CALL FOR A STEADY STATE BUILDF CONF CHAMBRS CIRC CALL ADDMOD WALL S AWAKEN WALL SUBMODEL CHAMBRS FK100 0 5 INITIALIZE INLET LOSS ON FIRST PATH CALL TRANSIENT S START A TRANSIENT ANALYSIS Results Summary Because of the cooling provided by the recirculation loop the pressure and temperature limits are never exceeded the run continues to an end time of 59 seconds To continue modeling the system response the volume of Tank 1 could have been allowed to shrink to an arbitrarily small size at which point the VDOT volume rate could be set to zero to simulate the stopping of the piston as it reaches the bottom of vessel 1 The initial flow rate in the recirculation line is about 230 lb hr dropping quickly to about 65 lb hr as the sensible heat in the pipe wall which is t
2. The following registers are added to the model in preparation for later inputs dens 170 0 dond 95 0 cp 0 21 tinitw 300 0 outerd 0 6 12 0 apipe 0 25 pi outerd 2 diam 2 where apipe is the cross sectional area of the pipe for purposes of calculating axial conduc tion if applicable and nodal capacitance Other registers are more obvious dens for the den sity of the pipe material cond for its conductivity cp for its specific heat tinitw for the initial temperature of the pipe and outerd for the outer diameter OD of the pipe This list is in addition to the previously defined registers length 10 0 diam 0 5712 0 area 0 25 pi diam 2 throat 0 1 144 0 pinle 11 0 tiniti 200 0 tLinit2 400 0 Thermal Submodels and Submodel Usage Even though the pipe is externally insulated thermal interactions will exist between the wall mass and the fluid To model the mass of the pipe requires a different type of network a traditional SINDA thermal node and con ductor submodel Finite masses may be represented by diffusion nodes and conductive heat paths by linear conductors Models may be composed of collections of thermal and or fluid submodels Common uses of submodels include 1 Combined models Submodels enable SINDA FLUINT models to be combined with out internal numbering or control conflicts 2 Organization Even if a single engineer were building the entire model it is conve nient to us
3. USE STUBES TANKS DEFAULT Notice the repeated use of DEF subblocks to define and then redefine default parameters for subsequent inputs LINE is an HX like duct macro that generates no ties and hence is by default adiabatic although ties can be added manually if desired Note also that CZ is the coordinate location or elevation of each lump on the Z axis The acceleration along that axis is due to gravity which in units of ft hr yields ACCELZ 32 2 3600 0 2 or waccelz gray The above line may be placed in HEADER CONTROL DATA GLOBAL or in OPERATIONS before calling the transient solution routine TRANSIENT How can we calculate the initial flow through the recirculation loop and initialize that loop without disturbing the other submodels whose initial conditions are prescribed Using submodels we can build a configuration inside OPERATIONS in which CHAMBRS doesn t exist Then the wall submodel can be placed in a boundary state by calling the auxil lary routine DRPMOD or HTRMOD routines Using DRPMOD results in the following HEADER OPERATIONS BUILDF CONL CIRC BUILD CON1 WALL CALL DRPMOD WALL SUSPEND WALL SUBMODEL AS BOUNDARY Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 40 of 42 t csr TECHNOLOGIES Next a steady state solution which will act on CIRC using WALL as a boundary condition is invoked by a call to the steady state solution routine STEADY DEFMOD CIRC FK26 1 0
4. Variation 4 Mixtures In all previous analyses the working fluids were all pure substances or in the case of air could be treated as such in one or two phases Multiple constituents can also be tracked with associated gradients in concentrations Assume that the gas in vessel 1 in the original problem is initially air 23 by mass oxy gen 77 nitrogen while the gas in the second tank is initially 100 hydrogen Properties for these gases shown below remember that you don t normally create these yourself you will just use pre made files HEADER FPROP DATA 8728 SI 0 0 C N2 GAS NEAR 1 ATM C VALUES BELOW 77 36K ARE FOR VAPOR MOLW 28 01 TMIN 65 0 TMAX 1000 0 AT V 65 0 4 40E 6 77 36 5 44E 6 SU 0 5 59m 6 05 20 5 9E 6 90 0 61 22n 6 95 0 6 548 6 100 0 6 37E 6 LOS Ry LEER 110 0 7252E 6 Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 29 of 42 t csr TECHNOLOGIES 450 400 350 300 250 Temperature F 200 150 15 Time Since Compression Began seconds 20 25 30 Valve Inlet Node 5 35 40 45 Figure 5 Temperature Histories for Two Phase Model With Cold Wall Closed Valve Liss 130 220 200 480 0 7 838 6 O 9 TSE 6 0 12 SE 6 O 17 3B 6 pee Z AD Ky 65 0 6 1E 2 85 0 8 08F 23 105 130 200 320 480 AT CP 6S 500 as Qy 9 BEA syda 1E Uy 19 3B 3 0 27 4E 3 O 37 woh 3 0 1 03983 s0 1 056853 LAO
5. for hydraulic diameter etc Notice that register names have been substituted for various inputs expressions con taining references to registers could have been used as well Notice that the default temperature defined above is incorrect for vessel 2 That overgener alization can be corrected when Tank 2 is input Alternatively or a new LU DEF input can be issued prior to the input of that lump To define the lumps in any order using lump subblocks LU TANK 1 VDOT 1 0 60 0 VESSEL 1 SHRINKING WITH TIME LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL LU JUNC 10 LINE OUTLET VALVE INLET LU TANK 2 Th tinit2 VESSEL 2 INITIALLY AT 400 DEG F LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL Notice the use of a numerical expression for the definition of VDOT for Tank 2 converting ft min to ft hr SINDA FLUINT allows such expressions within blocks in order to make the models more self documenting These expressions can become quite complex perhaps including spreadsheet like interrelationships between inputs and outputs The designation of LSTAT STAG means that the two end tanks terminate the line and have negligible velocities that their pressure will be interpreted as total or stagnation and that an accelerational loss will be applied to any path that extracts fluid from such Since pressures are static by default in FLUINT a K factor of 1 0 need not be added to the exit Instead
6. 0 25 pi diam 2 throst 0 1 144 0 pinit 11 0 Limiti 200 0 Cinat2 400 0 dens 170 0 cond 95 0 Cp 0 21 tinitw 300 0 outerd 0 6 12 0 apipe 0 25 pi outerd 2 diam 2 INSERT 6070 water inc Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 25 of 42 t csr TECHNOLOGIES HEADER FLOW DATA CHAMBRS FID 6070 WATER IS THE WORKING FLUID C SET LUMP DEFAULTS LU DEP PL pinit 11 PSIA INITIAL PRESSURE 9 TL tiniti XL 1 0 S 200 DEG F INIT TEMP ALL GAS VAPOR VOL 1 0 1 CU FT VOLUME IN TANKS C PATH DEFAULTS PA DEF FR 0 0 ZERO INITIAL FLOW RATE DH diam 1 2 INCH DIA CIRC AREA UNLESS SPECIFIED C DEFINE MODEL LU TANK 1 VDOT 1 0 60 0 VESSEL 1 SHRINKING WITH TIME LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL C CENTERED 5 SEGMENT HX MACRO l REPRESENTING LINE C START WITH LUMP JUNCTION 100 PATH STUBE 100 TIE 1 ATTACHED TO C STARTING NODE WALL 1 GENERATE MACRO THAT CONNECTS LUMP 1 TO 10 M HZ 1 C 100 100 1 WALL 1 1 10 NSEG 5 DHS diam HYDRAULIC DIAMETER TLENT length TOTAL LENGTH LU JUNC PA STUBE USE JUNCTIONS AND STUBES THROUGHOUT S LU JUNC 10 LINE OUTLET VALVE INLET PA CONN 20 10 2 DEY LOSS FPK 20 0 AFTH throat LU TANK 2 TL tinik LSTAT STAG VALVE MODEL PATH 20 GENERIC LOSS VALVE LOSS THROAT AREA VESSEL 2 INITIALLY AT 400 DEG F NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL T UF UF Ur X y UF X Y Q HEADER CONTROL D
7. ALL GAS VAPOR VOL 1 0 1 CU FT VOLUME IN TANKS C DEFAULT CONCENTRATIONS OF HYDROGEN OXYGEN AND NITROGEN XGH 0 0 XGO 0 23 XGN 0 77 C PATH DEFAULTS PA DEF FR 0 0 ZERO INITIAL FLOW RATE DH diam 1 2 INCH DIA CIRC AREA UNLESS SPECIFIED C DEFINE MODEL LU TANK 1 VDOT 1 0 60 0 VESSEL 1 SHRINKING WITH TIME LSTAT STAG PA CONN 10 1 10 DEV STUBE NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL LINE PATH 10 FROM 1 TO 10 SHORT TUBE TLEN length LENGTH FK 0 5 INLET LOSS LU JUNC 10 LINE OUTLET VALVE INLET UY UF UF UF TU UF a PA CONN 20 10 2 VALVE MODEL PATH 20 Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 31 of 42 t csr TECHNOLOGIES DEV LOSS FK 20 0 GENERIC LOSS VALVE LOSS APTH throat THROAT AREA LU TANK 2 Th tinit2 VESSEL 2 INITIALLY AT 400 DEG F LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL XGH 1 0 XGN 0 0 XGO 0 0 ALL HYDROGEN INITIALLY In addition two new output routines are added to the OUTPUT CALLS block one to tabu late constituents by lump LMCTAB and another to tabulate data by constituent CNSTAB CALL LMCTAB CALL CNSTAB 9 9 9 When these changes are made to the original sample problem vessel 2 reaches 450 F in only 20 seconds at which time it contains by mass 18 oxygen 61 nitrogen and 20 hydrogen which is responsible for 78 of the partial pressure The system pressure has risen to only 14 2 psia Variation 5 Co
8. Oe Bs 140 0 9 240 0 15 340 0 19 300 0 25 AaOr Te L 0 0 92 5 110 0 10 150 0 12 220 0 19 340 0 28 500 0 38 320 0 1 HEADER FPROP DATA 8732 SI 0 0 C OXYGEN AT 1 ATM MOLW 31 9988 TMIN 80 AT V 80 0 6 27H 6 140 300 700 AT K 100 155 400 Dg 10 6E 6 20 20 7 6 0 358 5E 6 we bree See Wee 0 0 0142 0 0 03482 Introduction to FLUINT TO ye Ps L5E 6 4E 6 OR 6 gt TEG 1 9 E 6 Bess B 3 SBS y 9E 24 S 2 TE 3 6E 3 0398E3 0 TMAX 68E 6 153 0 11 68 6 400 0 25 120 0 0 170 070 300 070 9 65 0109 0156 D44342 T25 1809 260 440 0 8 0FE 6 0 11 8EB 6 D 16 0E 6 O 23 7EB 6 P3286 T 48 3 95407 849R 3 ries 160 240 380 380 700 Lo 170 500 130 200 600 V5 6 Rev 1 12 13 12 0 10 E 2 0 14 7E 3 0 21 B J SE s 0 1 042E3 20 O 9 12E 6 GO 14 TE 6 5 0 320 58 6 90 0119 0 0 0182 0 0 0480 126 2 8 65E 6 200 0 12 9BEB 6 280 0 16 SE 6 460 0 24 4E 6 60 0 7 6E 3 100 0 S 48 3 Lao Oy Ll TE 180 0 16 5E 3 3200 0 26 0B 3 400 0 32 5SE 3 460 0 1 050E3 130 0 9 5E 6 200 0 14 7E 6 600 0 34 7E 6 140 0 0 0128 300 0 0 0267 700 0 0 0544 Page 30 of 42 t csr TECHNOLOGIES AT CUP 80 0 909 8 LPO 0 909 38 200 0 910 2 300 0 918 4 400 0 941 5 500 0 970 9 600 0 1002 700 071031 C HEADER FPROP DATA 8702 SI 0 0 C PARA
9. TL static pressure PL density DL etc Lumps may represent the thermodynamic state of a finite volume within a fluid sys tem They may be used more abstractly to represent boundary conditions volume less inter faces or dead ends etc There are three types of lumps classified by their volume Tanks have a finite volume VOL Tanks may represent a vessel or a finite cell within a subdivided volume Their volume may grow or shrink with time according to a prescribed rate VDOT or they may stretch or contract with pressure changes according to a pre scribed volumetric compliance COMP Tanks may also share interfaces common bound aries with other tanks for modeling pistons bellows liquid vapor interfaces including those within capillary structures and for subdividing control volumes into quasi 2D and 3D flow networks Plena plural of plenum have an infinite volume and hence usually represent boundary conditions sources or sinks or huge vessels whose thermodynamic state is assumed con stant Junctions have zero volume both energy and mass flowing into an arithmetic node must bal ance the energy and mass flowing out at all times junctions represent a local steady state or time independent solution Junctions may be used to model interfaces joining points e g tees dead ends and negligibly small volumes You may apply a source or sink of power QL to tanks and junctions
10. This heat source may vary with time or temperature As with any variation in SINDA FLUINT these dependen cies may be defined by table look ups or by arbitrarily complex calculations and logical manipulations In addition to QL you can define additional heat transfer to a lump by using a connection to a wall element represented by a SINDA thermal node These connections between FLUINT lumps and SINDA nodes are called ties as described below Ties can be used to invoke single and two phase convection correlations with the heat rate between the fluid and solid element calculated automatically Many other optional lump variables or descriptors exist Lumps may be alternatively designated as stagnant meaning negligible velocities within the lump t SINDA FLUINT divides the world into the finite the infinite and the negligible Engineering judgment must be used to decide which volumes and time constants are important and which can be neglected in order to answer the question at hand Substituting a junction for a tank will speed up the solution but might also introduce sharp edged and potentially jittery responses since there is no lag for a junction Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 3 of 42 t csr TECHNOLOGIES Many lump variables are available to help describe the thermodynamic state such as qual ity XL void fraction AL enthalpy HL density of the liquid phase DF partial pressure f
11. air inc It can then be added to any input file using the INSERT directive within that input file INSERT air ine Where T must be placed in column 1 If the file air inc is not contained within the same sub directory or folder as the input file the full path name can be specified Otherwise the pro gram will automatically search for this file in your designated library locations INSERT files help promote the creation of user libraries of fluid and material property libraries although they have many other uses as well Once a fluid has been described and verified it may be used in one or more fluid submodels To invoke the above fluid description in the sample problem INSERT air inc HEADER FLOW DATA CHAMBRS FID 8729 C SET LUMP DEFAULTS LU DEP PL pani 11 PSIA INITIAL PRESSURE et cetera Variation 4 below explores working fluid mixtures Variation 2 Convective Heat Transfer Duct Macros and More For this sample problem extension readers are assumed to be familiar with basic SINDA thermal networks and or to have read the corresponding SINDA introduction What if the fluid in the line connecting the two vessels were not adiabatic Suppose the line consisted of an insulated aluminum p 170 lb ft k 95 BTU hr ft F Cp 0 21 BTU Ib pipe with an outer diameter of 0 6 in The initial temperature is 300 F Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 21 of 42 v CsR TECHNOLOGIES Registers
12. connecting line since the den sity changes turned out to be small Temperature histories are shown in Figure 6 Vessel 2 quickly cools to the temperature of the incoming fluid namely Vessel 1 and then as condensation takes over it tracks the temperature of the connecting line while Vessel 1 begins to heat up due to compression If data on the solubility of air in water were available or could be estimated by the program using Raoult s Law then the program would have included the effects of air dissolving into liquid water based on built in correlations for dissolution mass transfer that rely on flow regime information Extensive capabilities are available for dissolution evolution modeling that are beyond the scope of this introductory document The multiple constituent modeling feature is currently limited to mixtures of up to 26 perfect gases and or simple incompressible liquids with up to one real gas or condensible volatile species Vessel 1 Tank 1 Line Middle Node 3 Vessel 2 Tank 2 400 350 g 300 2 a 250 E 200 150 0 5 10 15 20 25 30 35 40 45 50 55 60 Time Since Compression Began seconds Figure 6 Temperature Histories for Model with Cold Wall and Gas Open Value Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 33 of 42 CsR TECHNOLOGIES Variation 6 Fast Transients Choking and User Logic In order to illustrate the application of FLUINT to fast hyd
13. developing your own fluid since it likely already exists Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 19 of 42 t csr TECHNOLOGIES T0 O cO AT8TET 8z L 0001 WAON O T O T Z OT SSOT oz aSWHd T XL00 0 O9 Z 0 46 68 S 8z L 0001 WAON O T O T OT E aanLS OT AWNI DAA HOVN SC IONATS sddd VI IHG ALVEIMOTA W4LSdN TX LLS r dnd I ana dWT T dWT ddAL HLVd SYdNVHO THCOWENS HOJ NOLLVINAVL WHLHIWV3SVd HLVd cO d6888e9 T CNAWIL SA O d6667L7 T NANIL AWIL WHTdOdd SO dcesoso 9 dALS ANIL ASVadAV 00000 0 O H8000000 T ANIWLG AXWNLG SA SO HST9ELO S dHLS ANIL LSVTI LIWIT H NVHO AWATOA NOSWHa T MNWL SNILIWIT SO0 H80ZZ06 S daLS ANIL XVW 8IL ON GINTA SYaWVHO HNWVN THGOWaNS QAINTA 3OdGM4 ANIT GHATVA Ad GCHLOYNNOD SYSaNVHO OML S8VGNIS IHdCOW STZ AOV d AOLVA ALNI CINTA HLIM WHZATVNV NIONAYAAAIA TVOIAANWAN AAAOYANWNI SWHLSAS VO dAvPST T 000 0 000 0 LOGT cO0 d8080 S 000 T 000 T SL E 9 GEE ONAC OT COTE 8z L 000 0 COGE cO d80TD D 000 T o00 T DL ee 8 O0S7 ANWYL Zz 6ZO0E B2L C 000 0 LOCT cO H0c80 S 000 T 000 T 9L ES 9 GEE ANWYL T ALVA A DAANA ALVA SSUN ALVA LVAH Ad TWHLNA ALISNAG LOVad GIOA ALITWNO aqanssadd dl L HddAL dIN0I7I SHqIWVHO IHGONS0S AOI NOLLVINEVL WMHLHMWV3AVd dWn l cO d6888e9 T CNAWIL SA cO d6667L0 T NAWIL ANIL WaTdOdd SO dceS0S0 9 dALS SWIL HDVHHAV 00000 0 0 H000000 T ANIWLG 4XWWLG SA SO HST9 L0 S d
14. enough for the liquid to stay liquid then a simple user defined liquid four digit fake ASHRAE number beginning with 9 can be used to represent the fluid Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 36 of 42 v CsR TECHNOLOGIES HEEL E LEE ETT N ae 4usuq SA JBA OJ 81P IAA0 d4 15 20 25 30 35 40 45 50 Time Since Valve Opening milliseconds 10 ATT TISSSNSNSNE TREE IWESENHEE 20 25 30 35 40 45 50 Time Since Valve Opening milliseconds 15 Figure 7 Flow Rate and Pressure Histories for Valve Opening Event Page 37 of 42 V5 6 Rev 1 12 13 12 Introduction to FLUINT v CsR TECHNOLOGIES Water Reservoir 200 F SSS SSS 10 x 1 2 line typ Vessel 1 Vessel 2 10 x 3 4 ID Annulus Figure 8 Schematic for Passive Circulation Shell and Tube Heat Exchanger The following simplified liquid water description was used HEADER FPROP DATA 9718 ENG 0 0 C PROPERTIES ARE ACCURATE FOR SATURATED LIQUID OVER FULL RANGE te TMIN 491 70001 TMAX 1150 0000 AT 5 491 70001 4 3486681 505 67493 3 3405128 22 5 7493 y 225328307 42 67493 1 9282620 566 67493 1 4802542 598 6067493 Islas 3355 646 67493 0 87072533 1150 0000 1l1 1691712 AT K 491 70001 33202705 1150 0000 0 61937493 AT D 491 70001 62 451759 71 67493 4 025072 1150 0000 30 815065 ATST 491 70001 4 81158635E 03 708 67493 3 69433896E 03 819 67
15. length of the segment Therefore in the above state ment the tie is made to path 10 and it will therefore add extract heat to the lump that is downstream of that path e g junction 10 If the flow rate in the line were ever to reverse which never happens in this model the HTNS tie would automatically jump to the cur rent downstream lump Tank 1 HTN ties and other finite difference style ties remain attached to the designated lump In addition to forced convection ties user defined ties may be made These allow you to spec ify and perhaps calculate the heat transfer conductance themselves Once again either finite difference style HTU or segment oriented HTUS ties are available Independent of the tie method chosen a few more inputs would be required to complete this model First a TIETAB tie tabulation call is made in the existing block which will print wall node information along with tie conductances and heat rates HEADER OUTPUT CALLS CHAMBRS S DESIGNATE OUTPUT OPERATIONS CALL LMPTAB LUMP TABULATION CALL TIETAB TIE TABULATION CALL PTHTAB PATH TABULATION IFP PL1 GE 30 0 OR LOGIC TO TERMINATE TRANSIENT TLI GE 450 0 OR TL2 GE 450 0 TIMEND TIMEN Ur U UW UW Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 23 of 42 t csr TECHNOLOGIES Next the existence of the new thermal submodel must be declared by a new build statement in OPERATIONS BUILD for thermal submodels and BUILDF for f
16. might be useful for calculating other inputs One register may be defined per line Note that registers may either be set to constants or to expressions which perhaps use built in functions such as sin x In x max x y etc or built in constants such as pi in the above expression for area and which perhaps contain ref erences to other registers such as diam in the above expression for area SINDA FLUINT input files are subdivided into blocks called Header blocks The above dec larations of registers are placed into a section titled HEADER REGISTER DATA A com plete subsection of the input file is therefore HEADER REGISTER DATA length 10 0 diam 0 5 12 0 area 0 25 pi diam 2 throat 0 1 144 0 pinit 11 0 tiniti 200 0 tinit2 400 0 Notice that the H in HEADER must be placed in column 1 This is one of the few column restrictions in SINDA FLUINT column 1 is reserved for certain top level commands Lumps and Paths The volumes of the vessels are important to the system response There fore each will be represented by tanks i e control volumes that are labeled 1 and 2 respectively The volume within the line and valve however is assumed negligible less than 0 013 ft and hence any lumps required there will be volume less junctions Does the line need to be discretized i e subdivided into strings of lumps and paths or can it simply be represented by a single path The an
17. model the reservoir Using the following new registers for convenience length2 10 0 diam2 0 5 12 0 od 0 75 12 0 id douter the FLOW DATA portion of the new submodel becomes INSERT water inc HEADER FLOW DATA CIRC FID 9718 E LU DEF Pl 100 0 S DEFAULT PRESSURE Th 200 0 DEFAULT TEMP Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 39 of 42 t csr TECHNOLOGIES PA DEF FR 100 0 DEFAULT FLOW RATE TLEN length2 DEFAULT LENGTH DH diam2 DEFAULT DIAM LU PLEN 1000 CZ length2 10 FT HIGH ELEVATION LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL PA CONN 10 1000 10 DESCENDING LINE DEV STUBE Wy Ur FK 0 5 INLET LOSS C LU DEF CZ 0 0 LOWER LEVEL ZERO ELEVATION C LU JUNC 10 INLET TO ANNULUS C ANNULAR PASSAGE ADIABATIC OUTSIDE CONVECTION INSIDE M HX 1 C 104 105 5 WALL 5 10 1 NSEG 5 MATCH CHAMBRS LUINC 1 PAINC 1 TIINC 1 NINC 1 NEGATIVE INCR TL 200 0 TLINC 10 0 SPEFICY INIT TEMP GRADIENT DHS od id HYD DIAMETER TLENT length TOTAL LENGTH AFS 0 25 pi od 2 id 2 FLOW AREA AFRACT id od id ONLY ID IS ACTIVE AREA FRACT PA STUBE USE STUBES TANKS DEFAULT LU DEF TL 250 0 CHANGE DEFAULT TEMP TO 250 LU JUNC 1 OUTLET OF ANNULUS M LINE 2 C 21 21 1 1000 NSEG 5 ASCENDING LINE DHS diam2 HYD DIA TLENT length2 TOTAL LENGTH CZ length2 5 CZINC length2 5 VARY ELEVATION ALONG IT PA STUBE
18. of transient integrations all starting with those initial conditions etc 4 output operations including the amount type and frequency of outputs control parameters to define or customize units physical constants solution meth ods or accuracy etc 6 supplementary logic such as convection correlations or simulation of electronic con trollers user defined devices etc Tanks WOE D STUBE short tube Lumps CAPIL capillary path i Junctions VOL 0 LOSS K factor loss _ LOSS2 directional loss Plena VOL CHKVLV check valve CTLVLV control valve TURBINE gas steam hydraulic turbine Connectors COMPPD positive disp compressor inertia 0 MFRSET set mass flow rate VFRSET set volumetric flow Paths PUMP centrifugal pump TABULAR head vs flow rate by tables Tubes ORIFICE orifice including regulating valve inertia gt 0 COMPRESS variable disp compressor NULL user defined device HTU HTUS user defined conductance HTN HTNS forced convection Ties HTNC forced convection centered HTP pool boiling and quasi stagnant convection USER user defined conductance between lumps Fties CONSTQ user defined constant heat rate between lumps AXIAL automatically calculated axial conduction within fluid FLAT pistons liquid vapor interfaces subdivisions etc OFFSET pistons liquid slugs etc Interfaces SPRING bellows actuators etc between SPHERE spherical voids tanks WICK li
19. or throat size ORIFICEs are commonly used to represent valves They can even be prepro grammed to open or close PA CONN 20 DEV ORIFICE MODA 1 AORI 0 0 AORI L 0 0 CONNECTING LUMP IDs NOT SHOWN OPENING VALVE MODELED AS AN ORIFICE INITIALLY CLOSED NO FLOW WHEN CLOSED MIN POSITION AORI_H AOPEN AOPEN WHEN OPEN MAX POSITION AORI_T TOPEN TOPEN IS TIME TO OPEN This ORIFICE description in FLOW DATA uses two custom user supplied variables named AOPEN the valve throat area when opened and TOPEN the time it takes to open the valve These auxiliary variables may be declared and initialized as additional registers which can then be used in input spreadsheet expressions or as Fortran variables in logic or both HEADER REGISTER DATA TOPEN 10 0E 3 3600 0 S TIME REQ D TO OPEN VALVE AOPEN throat S OPEN THROAT AREA FOR VALVE Registers are available for inspection or modification in all logic blocks and any changes to them will be reflected in any input expressions referring to them automatically One reason for declaring these constants as registers is because of the SINDA FLUINT spell checker or debug feature The spell checker makes sure that all variables and identifiers are declared and valid since Fortran otherwise does not enforce such a rule and will simply use Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 35 of 42 v CsR TECHNOLOGIES zero as values for misspelled
20. plots as a sinusoidal waveform superimposed upon the pressure and flow rate histories Actually the exact period of this oscillation varies over the course of the event because the sound speed changes as the steam in the line warms up and becomes pressurized Variation 7 More Fluid Submodels Heat Exchangers and Buoyancy Forces This variation starts from the model that included the wall thermal submodel Variation 2 Suppose that the connecting line had been actively cooled by a second working fluid perhaps pressurized liquid water Assume that the previous tube wall represented the inner tube annulus of a tube and shell counterflow heat exchanger The inner diameter of the new coolant annulus is 0 75 inches Furthermore assume that this heat exchanger were driven by natural recirculation a large tank 10 ft above the rest of the system and filled with water at 200 F is plumbed to the annular section by 10 ft long by 0 5 inch lines Figure 8 When the problem starts the fluid in this second loop has achieved equilibrium circulation with the wall initially at 300 F The flow has established itself such that the coolant water flows from the valve toward vessel 1 i e counterflow to the steam Although both fluid systems use the same working fluid and can therefore be contained within the same fluid submodel the best approach is to define a new fluid submodel which will be called CIRC As long as pressure in this new loop remains high
21. the acceleration loss from zero flow at the inlet is automatically included by designating the inlet tank as stagnant Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 12 of 42 t csr TECHNOLOGIES lumps This designation which is normally rarely needed except on plena the infinite vol ume lumps overrides the default assumption that the pressures are to be treated as static The paths connectors of varying device models might be defined as follows PA CONN 10 1 10 DEV STUBE TLEN length EE 0 5 PA CONN 20 10 2 DEV LOSS FK 20 0 AFTH throat LINE PATH 10 FROM 1 TO 10 SHORT TUBE DEVICE LENGTH INLET LOSS K FACTOR VALVE MODEL PATH 20 GENERIC LOSS DEVICE VALVE LOSS K FACTOR THROAT AREA TY UF UF UF TU Ur An UF Actually it is usually more convenient to arrange the inputs in schematic order mixing lump and path definitions together LU TANK 1 VDOT 1 0 60 0 VESSEL 1 SHRINKING WITH TIME LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL PA CONN 10 1 10 LINE PATH 10 FROM 1 TO 10 DEV STUBE SHORT TUBE DEVICE TLEN length LENGTH FK 0 5 INLET LOSS K FACTOR LU JUNC 10 LINE OUTLET VALVE INLET PA CONN 20 10 2 DEV LOSS FK 20 0 AFTH throat LU TANK 2 TL tinit2 LSTAT STAG All the data describing lumps paths and ties is placed in a header block called HEADER FLOW DATA Actually there may be more than one network or submodel in each model for pu
22. variables without warning While the spell checker can be turned off using it can help prevent errors that are difficult to trace One side effect of plac ing variables in REGISTER DATA is that they then become valid names in the eyes of the spell checker Otherwise their presence would trigger an error message if the spell checker were enabled Users can turn off the spell checker via a single command in OPTIONS DATA SPELLOFF It can also be turned on and off locally within a logic block When the valve first begins to open the pressure difference across it is large enough such that choking occurs In other words Mach number reaches unity at the throat and the flow through the valve cannot exceed the corresponding critical flow rate As the valve opens more and as the line pressure rises the flow rate will eventually become limited by fric tional losses rather than sonic constraints Finally the end time needs to be adjusted to account for the small time scale of the event of interest TIMEND 50 0E 3 3600 0 RUN FOR 50 MILLISECONDS Results Figure 7 presents the flow rate and pressure profiles for the resulting event Although the valve is fully opened by 10 ms it appears to have been choked for most of that time It takes about 30 ms for the pressure to reach 30 psia and for the flow through to valve to significantly diminish During the event the steam column oscillates with a period of about 5 to 10 ms visible in the
23. which position it exhibits a K factor head loss of 20 i e AP KpV7 2 where K 20 See Figure 3 1 0 ft min 1 2 in valve Fully open K 3 A 0 1in K 20 Vessel 1 Vessel 2 Ti 200 F Line 10 ft x 1 2 in dia Ti 400 F Pi 11 psia Pi 11 psia Vi 1 ft Vi V 1 ft Figure 3 Sample Problem Schematic How long after the piston begins moving does it take to either exceed 30 psia or 450 F any where within the system This problem will be first worked at a very simple level and will then be gradually expanded to include more details and design questions in later sections Model and Input Development Registers It is useful to start by defining a set of registers containing major problem vari ables properties dimensions etc Changes to these registers in subsequent runs will prop agate through the input file automatically facilitating updates These changes can also be propagated during a run facilitating parametric investigations Registers are the basis for many powerful options in SINDA FLUINT so their extensive use is strongly encouraged Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 10 of 42 t csr TECHNOLOGIES Using arbitrary user defined names the basic parameters of the model are defined as length 10 0 diam 0 8 12 0 area 0 25 pi diam 2 throat 0 1 144 0 pinit 11 0 tiniti 200 0 tinit2 400 0 where area is the flow area of the pipe which
24. 2 3 4 5 C GEN G G GINC NA NAINC NB NBINC cond X sectional Area length GEN 4 233 Asi Bats cond apipe length 5 0 Analogous to the GEN commands FLUINT features a single command subblock that gener ates lumps paths and ties representing a continuous duct segment Replacing STUBE 10 with a string of lumps and paths that corresponds to the nodes in WALL is achieved by the following command C CENTERED 5 SEGMENT HX MACRO 1 REPRESENTING LINE C START WITH LUMP JUNCTION 100 PATH STUBE 100 TIE 1 ATTACHED TO C STARTING NODE WALL 1 GENERATE MACRO THAT CONNECTS LUMP 1 TO 10 Cc M HX 1 C 100 100 1 WALL 1 1 10 NSEG 5 DHS diam S HYDRAULIC DIAMETER TLENT length TOTAL LENGTH LU JUNC PA STUBE USE JUNCTIONS AND STUBES THROUGHOUT This macrocommand generates junctions 100 through 104 STUBE connectors 100 through 105 and ties 1 through 5 default increments are unity Unfortunately this macrocommand is not completely equivalent to the replaced STUBE 10 since that path also had a K factor of 0 5 applied to represent the inlet losses Since adding FK 0 5 in the above macrocommand subblock would then add this factor to all six macro paths a single line of logic will instead be added to OPERATIONS before the transient rou Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 24 of 42 CsR TECHNOLOGIES tine is initiated This method is not necessary when using Sinaps or FloCAD since those codes enab
25. 493 2 838498915 03 1150 0000 5 88794137E 05 AT CP 491 70001 1 1114392 531 67493 1 0746689 FIZ 6 7493 4 L 0225220 1063 6749 1 4633179 1144 2500 2 1658936 Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 38 of 42 v CsR TECHNOLOGIES AT COMP 491 70001 2 81523830E 06 708 6097493 4 03973172E 06 850 67493 5 77156106E 06 where arrays of properties versus temperature include V dynamic viscosity K thermal conductivity D density ST surface tension CP specific heat and COMP compliance Surface tension is only needed when used in two phase mixtures and the compressibility is only needed for waterhammer choking or speed of sound calculations This information is placed in a file named water inc to be addressed by an INSERT directive The fluid sub model CIRC will therefore use FID 9718 within its HEADER FLOW DATA subblock The new submodel CIRC connects to the WALL nodes from the outside Neglecting the gra dient through the pipe wall the new fluid submodel can attach directly to the existing wall nodes Assuming that this water cooled jacket is itself adiabatic on the outside how should such an annular passage be modeled The flow area AF of the passage is t D D 4 where D is 0 75 inches and D is 0 6 inches the OD of the original connecting line The wetted perimeter is n D D By the definition of the hydraulic diameter DH DH 4A P D Dj An effective diamete
26. ALS AWIL LSVT LINIT SONVHO AWATOA NOSWHa T MNVL ONILIWIT S0 8807706 S daLS AWIL XWW 8TL ON CINTA SYaNWVHO AWVN THCOWaNS dIn A MOdadMa ANIT GHATVA Ad GHLOAYNNOD SYHENVHO OML VANIS TAON PIZ dDVd YOLVAOSLNI CINTA HLIM ASZATIVNY NIONAAAAAIA TWOIAANWNAN GHAOHSdNWI SWHLSAS Page 20 of 42 V5 6 Rev 1 12 13 12 Introduction to FLUINT v CsR TECHNOLOGIES CP 1 005E3 CPTC 0 03 SPEC HEAT AT TREF CP TEMP COEF K 26 1E 3 KTC 8 0E 5 COND AT TREF PLUS TEMP COEF V 18 5E 6 VTC 0 051E 6 VISC AT TREF PLUS TEMP COEF More complete variations may be specified for any or all properties in the form of bivariate arrays e g temperature property pairs such as Tj k4 To ko etc as follows HEADER FPROP DATA 8729 SI 0 0 MOLW 28 96 TMIN 90 0 TMAX 1000 0 90 0 6 35E 6 100 0 7 06E 6 110 0 7 7T5B 6 120 0 8 43E 6 130 0 9 09R 6 140 0 9 74E 6 160 0 1150R 200 0 13 4E 6 220 4 14 5E 6 320 0 19 5E 6 340 0 20 4E 6 380 0 22 1E 6 400 0 22 9E 6 BU0 20 26 8E 6 600 0 30 3E 6 AT Ky 90 0 9 37EB 2 100 0 32 bHs 2 Lig 0 IO 25 25 159 0 13 8R 3 160 0 14 6BE 2 180 0 16 4E 3 200 0 18 28 3 220 0219 BE 3 5 240 0 21 5E 3 300 0 26 1 B 3 5 400 0 23 18 3 200 0 39 58 3 600 0 45 6E 3 AT CP 900 1 002E3 200 0 1 00285 280 0 1 004E3 340 0 1 007E3 400 0 1 013E3 500 0 1l 029E3 600 0 1 051E3 AT V The above block might be placed into a separate file named
27. ATA GLOBAL C ENG UNITS ft F lb hr psia will be used UID ENG TIMEND 59 0 3600 0 RUN FOR 59 SECONDS HEADER OPERATIONS BUILDF ALL BUILD ALL CHAMBRS FK100 0 5 INITIALIZE INLET LOSS ON FIRST PATH CALL TRANSIENT START A TRANSIENT ANALYSIS gt HEADER NODE DATA WALL a GEN N N NINC Ti density Cp Length X sectional Area number GEN 1 5 1 tinitw dens cp length apipe 5 0 HEADER CONDUCTOR DATA WALL C GENERATE 4 CONDUCTORS 1 2 3 4 BETWEEN 5 NODES 1 2 3 4 5 C GEN G G GINC NA NAINC NB NBINC cond X sectional Area length GEN 1 4 1 1 1 2 1 cond apipe length 5 0 E Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 26 of 42 CsR TECHNOLOGIES HEADER OUTPUT CALLS CHAMBRS DESIGNATE OUTPUT OPERATIONS CALL LMPTAB LUMP TABULATION CALL TIETAB TIE TABULATION CALL PTHTAB PATH TABULATION IF PLI GE 30 0 QR STOP PROBLEM LOGIC TLI GE 450 0 OR TL2 GE 450 0 TIMEND TIMEN END OF DATA Figure 4 displays the results of this analysis in terms of temperature histories for several important points in the model Since the tube mass was initially warmer than the fluid in vessel 1 it acts to heat the fluid in this line during the transient event Because warmer gas is injected into Vessel 2 it takes even less time 39 seconds to reach the criterion of 450 F than when the pipe wall was neglected By the end of the event tank 1 has almost warmed by compression up to the wall temperature wh
28. ATH DEFAULTS PA DEF FR 0 0 Uy ZERO INITIAL FLOW RATE DH diam 1 2 INCH DIA CIRC AREA UNLESS SPECIFIED C DEFINE MODEL LU TANK 1 VDOT 1 0 60 0 VESSEL 1 SHRINKING WITH TIME LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL PA CONN 10 1 10 LINE PATH 10 FROM 1 TO 10 DEV STUBE SHORT TUBE DEVICE TLEN length LENGTH FK 0 5 INLET LOSS K FACTOR LU JUNC 10 LINE OUTLET VALVE INLET VALVE MODEL PATH 20 GENERIC LOSS DEVICE VALVE LOSS K FACTOR THROAT AREA VESSEL 2 INITIALLY AT 400 DEG F NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL PA CONN 20 10 2 DEY LOSS PE 20 0 AFTH throat LU TANK S Th tinik LSTAT STAG UY UF UF UF TU UF X y UF X y A UF XY HEADER OPERATIONS BUILDF ALL CALL TRANSIENT START A TRANSIENT ANALYSIS a HEADER CONTROL DATA GLOBAL C ENG UNITS ft F lb hr psia will be used UID ENG TIMEND 59 0 3600 0 RUN FOR 59 SECONDS G HEADER OUTPUT CALLS CHAMBRS DESIGNATE OUTPUT OPERATIONS CALL LMPTAB LUMP TABULATION CALL PTHTAB PATH TABULATION IF PLI GE 30 0 OR STOP PROBLEM LOGIC TLI CE 450 0 OR a TL2 GE 450 0 TIMEND TIMEN END OF DATA The final optional line enables even more comments to be appended to the file since SINDA FLUINT will not read past this command Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 17 of 42 v CsR TECHNOLOGIES Other than the first mandatory OPTIONS DATA block with few exceptions the remaining HEADE
29. F When a substance condenses in the presence of noncondensible gases as is expected in this problem noncondensible gas collects near the liquid vapor interface forming a diffusion bar rier that slows the condensation rate and therefore results in larger temperature drops and lower effective film coefficients This effect is automatically included by the code providing that information necessary to calculate the diffusion constants for user defined fluids has been provided in the form of DIFV While DIFV is relatively easy to find or calculate for var ious fluids as per guidance provided in the User s Manual further discussion of that Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 32 of 42 t csr TECHNOLOGIES parameter is beyond the scope of this introduction Otherwise the diffusion effects are usu ally small compared to the shifting of the saturation condition due to the partial pressure of the noncondensible gas Results The addition of the gas eliminates the severe condensation rates that would other wise occur in a pure substance yet the existence of condensation keeps the tanks from hit ting a pressure or temperature limit throughout the 1 minute transient event By the end of the compression event the quality of Vessel 2 is about 80 meaning it is 20 liquid by mass although the volumetrically Vessel 2 is almost dry and it contains about 12 5 gas by mass Junctions could have been used instead of tanks in the
30. H2 GAS NEAR 1 ATM MOLW 2 0159 TMIN 14 0 TMAX 1000 0 AT V 14 0 0 748E 6 18 0 0 988E 6 20 27 1 12980 06 22 0 1 2198 6 26 0 1 424E 6 30 0y 1 621 B 6 oo40 1 856H 6 40 0 2 028 6 50 0 2 498E 6 60 0 2 884E 6 80 0 3 585E 6 100 0 4 574E 6 120 0 5 408B 6 260 076 3H 65 200 0 6 901EF 6 240 0 7 419R 6 300 0 8 141E 6 350 0 8 7238 6 400 0 9 297H 6 500 0 10 426E 6 AT K 14 0 0 01254 18 0 0 01497 20 27 0 01692 22 0 0 O19595 28 0 0 02247 32 0 0 02534 35 0 0 02741 40 0 0 03088 50 0 0 037813 60 0 0 04481 80 0 00 06027 100 0 0 08954 140 0 0 13613 180 0 00 15565 220 0 0 16241 260 0 0 16914 300 0 0 17591 400 0 0 19745 500 0 0 22128 AT CP 14 0 10 54E3 20 0 10 4353 24 0 10 3983 22 0 10 35E3 40 0 10 36E3 50 0 10 49E3 60 0 10 62BE3 80 0 11 7253 100 0 125 2523 160 0 16 34E3 200 0 18 07ESy 2860 0 15 9B3 350 0 14 63E3 400 0 14 55E3 500 0 414 52B3 These fluids may all be combined into the submodel CHAMBRS All are named within the subblock formed by the HEADER FLOW DATA card using FIDx where x is a single letter designating a fluid constituent HEADER FLOW DATA CHAMBRS FIDO 8732 OXYGEN IS FIDN 8728 NITROGEN IS N FIDH 8702 S HYDROGEN IS H Initial concentration in the form of a mass fraction XGx are added to the lumps C SET LUMP DEFAULTS LU DEF PL pinit 11 PSIA INITIAL PRESSURE TL tinitl XL 1 0 200 DEG F INIT TEMP
31. NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL Notice that the FID 6070 that is found within the subblock formed by the HEADER com mand This statement corresponds to the a file that is inserted via the INSERT command shown below into the input file This file is named f6070_water inc and it contains a tabu lar description of fluid 6070 which is water with an incompressible liquid phase This file was generated by calls to REFPROP and is available as part of the installation package UY UF UU UY UF X y UU X y aA INSERT 6070 water inc INSERT files and user defined fluids are described in more detail as Variation 1 below Units and Top Level Control Normally as long as you provide inputs that obey a consistent unit system SINDA does not enforce a unit convention When fluid submodels are used however you must choose between US Customary old English or SI metric units Such model level decisions are placed in a data block called HEADER CONTROL DATA which applies to all submodels if the special name GLOBAL is used For the current problem HEADER CONTROL DATA GLOBAL C ENG UNITS ft F lb hr psia will be used UID ENG Within this unit system you may optionally choose degrees Rankine and psig SI users may choose degrees C or K etc Other top level data that is appropriate to include in this model level block is the problem end time TIMEND for transients This is chosen as 59 seconds since Ve
32. R blocks may be input in any order Execution Internally SINDA FLUINT follows a two step process as was shown in Figure 2 In the first step the preprocessor the data file is scanned and analyzed for consistency Any format errors or missing data will be flagged and will cause the run to terminate If no such errors are found the preprocessor will write out a Fortran file created from your inputs The For tran compiler will then be invoked and the SINDA FLUINT library will be linked with the resulting object code to create the processor which will be unique for each problem run The processor is then executed with the instructions defined in OPERATIONS completely defin ing its scope On Linux machines as an example the above sequence may be invoked as sinda cham inp gt pp out where cham inp is the name of the file containing the model and pp out is the name of the file to contain preprocessor messages which are normally discarded for successful runs Fortran compiler errors if any will be either displayed on the screen or written to a file depending on the operating system and the compiler Processor output will be directed to the file named within cham inp which is named chambers out in this case On PCs a Windows based utility called SINDAWIN is used to launch a run if neither Sinaps nor Thermal Desktop is used Results Without Sinaps or Thermal Desktop the presentation of the output must be made in tab
33. ble volatile substance with conservation of mass equations applied to each fluid species Phenomena such as reduced heat transfer coeffi cients due to diffusion limited condensation are modeled Within mixtures any number of gases may dissolve into any number of liquids with phenomena such as homogeneous nucle ation modeled Liquids can be assumed to be miscible or immiscible perhaps as a function of temperature Solubilities can be defined in a wide variety of ways to exploit available data or even to estimate solubilities in the absence of data Species can also react generating or absorbing heat in the process Pseudo mixtures of variable molecular weight fluids are also available for simplifying the modeling of the hot products of combustion Goal Seeking The value of any input variable can be found given a desired response revers ing the traditional solution sequence Design Optimization SINDA FLUINT can perform multiple variable design optimization sizing design synthesis with arbitrarily complex constraints Automated Data Correlation SINDA FLUINT can be used to automatically adjust the uncertainties in a model as needed to correlate to CFD data or test data steady state tran sient or a complex mixture of the two Worst case Design Scenarios Given a list of uncertainties and variations SINDA FLUINT can be tasked with seeking the worst case scenario to be used as a design case Reliability Engineering Statistical Desig
34. capillary action to yield a pumping action when heat is applied Two phase Flow Regimes and Slip Flow If two phases are present flow regime mapping options may be invoked to improve the pressure drop and heat transfer calculations These options automatically choose between one of four regimes bubbly slug annular and strati fied based on local flow conditions orientation with respect to body forces or accelerations etc If desired slip flow differences in velocity between liquid and vapor can be modeled Otherwise the default two phase flow assumption is homogeneous Nonequilibrium Two phase Control Volumes By default liquid and vapor phases are assumed to be in thermal equilibrium within control volumes Modeling options exist that remove this assumption enabling finite rate mass and heat transfer rates between phases Choking and Fast Hydrodynamic Transients Choking sonic limited or critical flows is automatically detected and modeled through out a fluid network Also options exist to mon itor fast transient analyses waterhammer and acoustic wave propagation events if desired Otherwise FLUINT s implicit methods normally enable short time scale events to be skipped when the time constants of interest are longer e g the focus of the analysis is on thermal rather than hydrodynamic events Fluid Mixtures Up to 26 real or perfect gases and or simple liquids may be mixed within a single fluid network along with a condensi
35. changed into the equivalent LINE macro which generates no ties Also the TIETAB call in OUTPUT CALLS can be removed or commented HEADER OPERATIONS BUILDF ALL CALL TRANSIENT START A TRANSIENT ANALYSIS BUILD ALL HEADER OUTPUT CALLS CHAMBRS DESIGNATE OUTPUT OPERATIONS CALL LMPTAB LUMP TABULATION CALL TIETAB LUMP TABULATION CALL PTHTAB PATH TABULATION If the line were modeled with junctions instead of tanks the line would instantaneously pressurize to 30 psia when the valve were opened Therefore to capture the pressurization of the line and any associated acoustic waves tanks will be used instead Tubes are required instead of STUBE connectors in order to resolve any acoustic waves The HX macro becomes C CENTERED 5 SEGMENT HX MACRO 1 REPRESENTING LINE C START WITH LUMP TANK 100 PATH TUBE 100 TIE 1 ATTACHED TO C STARTING NODE WALL 1 GO FROM LUMP 1 TO 10 M HZ 12 C 100 100 1 WALL 1 1 10 NSEG 5 DHS diam HYDRAULIC DIAMETER TLENT length TOTAL LENGTH CG LU JUNC PA STUBE USE TANKS AND TUBES THROUGHOUT DEFAULT Th 1nib1 FLUINT uses an implicit solution that is specifically designed to avoid resolution of fast transients enabling long time steps Nonetheless if the time step is appropriately con Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 34 of 42 lt csr TECHNOLOGIES strained then such events can be resolved A routine named WAVLIM exists that helps cal culate the maxim
36. condensation will occur in the connecting line gradually filling the connecting line volume with liquid This effect will help to reduce the pressure rise that occurs as Tank 1 is compressed Two phase flow will be modeled automatically by the code as the fluid temperatures drop below the local saturation condition Nonetheless two phase flow not only introduces many more modeling options and therefore decisions its existence may render invalid some of the previous decisions The purpose of this section is to discuss these options and assump tions rather than to describe the details of the resulting model By default a Rohsenow Traviss condensation correlation is invoked to calculate film coeffi cients within the line Since these coefficients are much larger than the previous single phase gas case the wall temperatures change more rapidly and larger gradients develop Also by default the two phase flow is assumed homogenous and flow regime mapping is applied Other two phase pressure drop calculations and correlations may be easily invoked via the IPDC descriptor for tubes and STUBE connectors which could be added to the HX subblock IPDC 1 McAdam s Homogenous TPDC 2 Lockhart Martinelli IPDC 3 Baroczy s IPDC 4 Friedel s IDPC 5 Whalley recommended combination of above IPDC 6 Default Flow regime map based estimates The last option pressure drop estimates based on flow regimes deserves a li
37. d knowledge of the engi neer is a vital ingredient in both arriving at a suitable model and an optimum solution approach While this strategy may frustrate the casual user who is looking for an easy joy stick approach it delights the thermal fluid engineering professional who understands that real thermal problems rarely lend themselves to such simplistic treatment or to hard wired assumptions The inputs to the program may include 1 network description a set of lumps paths ties interfaces fties and perhaps SINDA nodes and conductors describing the device or system 2 associated support data if needed fluid properties event profiles such as fluxes ver Heat transfer directly between lumps axial conduction back conduction is modeled using fties or fluid ties Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 5 of 42 v CsR TECHNOLOGIES sus time etc SINDA FLUINT offers a spreadsheet like feature for defining key parameters e g dimensions loss coefficients in a central control panel such that the remainder of the inputs can be defined indirectly on the basis of these parame ters This spreadsheet feature not only facilitates model building and upkeep it also enables parametric analysis optimization data correlation goal seeking statistical design etc 3 solution sequence operations to be applied such as finding a steady state solution for initial conditions followed by a parametric series
38. e submodels for improved organization and better self documentation This is the most frequent use of submodels 3 Dynamic model variations Submodels may be dynamically added or deleted from the solution as needed to model changing geometries materials boundary condi tions assumptions etc To change the current configuration a new BUILD and or BUILDF statement is issued defining the new set of active submodels Any submod els not currently defined as active are ignored by subsequent analyses and any con ductors or ties that extend to nodes or lumps in inactive submodels are also ignored Single Node If temperature gradients along the pipe can be neglected then a single diffu sion node submodel WALL node number 1 can be used to represent the isothermal pipe wall mass Nodes are defined in their own unique HEADER block as follows HEADER NODE DATA WALL Cc N Ti density Cp Length X sectional Area 1 tinitw dens cp length apipe Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 22 of 42 t csr TECHNOLOGIES where the above information represents the node number the initial temperature and the nodal capacitance product of density volume and specific heat Within HEADER FLOW DATA a tie or heat transfer connection can then be made to the wall node The valve inlet junction 10 is the logical place for such heat energy ingress or egress to take place An HTN convection tie can be made that will automatica
39. he driving force for the buoyancy driven flow is extracted The pipe wall cools to about 200 F in 30 seconds then slowly heats back up due to the warm gas flow Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 41 of 42 v CsR TECHNOLOGIES More Information If you have questions about the use or availability of SINDA FLUINT Sinaps EZXY Thermal Desktop RadCAD or FloCAD please contact C amp R Technologies Inc CRTech Boulder Colorado Phone 303 971 0292 FAX 303 971 0035 Web site www crtech com The web site contains demonstration versions on line hypertext user s manuals training materials and other announcements Additional working fluid descriptions are also avail able to users Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 42 of 42
40. ices valves and control valves etc The two most important and common types of connector devices are STUBEs and MFR SETs MFRSETs represent passages with prescribed mass flow rates and are called mass SetFlows in Sinaps and FloCAD for this reason They are often used as either boundary conditions or idealized pumps STUBEs are short tubes or the time independent instan Short is actually defined as a small value of pL D In other words long and or thin and or liquid filled lines have more inertia than do short and or wide and or gas filled lines Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 4 of 42 t csr TECHNOLOGIES taneous analog of tubes In other words STUBE connectors are in every way identical to tubes except that they have no inertia just as junctions are in every way identical to tanks except they have no volume Ties Ties describe the means by which heat flows between FLUINT lumps representing the fluid and SINDA nodes representing the duct wall Each tie has a single characteristic conductance UA inverse of resistance and heat flow rate QTIE There are six types of ties subdivided by whether you provide the UA conductance as with HTU and HTUS ties or the program calculates it based on convection correlations and the current pressures two phase qualities flow rates etc as with HTN HTNC HTNS and HTP ties Forced convect
41. ich has stayed relatively constant due to the low gas flow rates and low film coefficients Temperature History Vessel 1 Temperature Vesselt 2 Temperature Gss Inlet Node 1 Inlet Node 5 450 400 E g 350 Q 2300 Q 0 002 Time Since Compression Began seconds 0 003 0 004 0005 0 006 0 007 008 009 000 0 011 Figure 4 Temperature Histories for Model Adding Tube Wall Mass In retrospect because of the brevity of the event compared to the thermal mass of the pipe no significant temperature gradient is ever developed in the wall a single node would have sufficed in this case Nevertheless resolution was required in the fluid line since large tem perature gradients did occur there In this particular case the use of an HTNS tie may have eliminated even this resolution requirement Variation 3 Two Phase Flow Pure Substance If the initial pipe wall temperature tinitw in the above problem Variation 2 were 150 F instead of 300 F then condensation would occur in the connecting line and two phase flow would result For this variation the valve in the line remains shut at all times and the line Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 27 of 42 t csr TECHNOLOGIES is initially full of vapor at the same temperature as Tank 1 e g considerably warmer than the wall around it Thus rapid cooling and then
42. igh that sonic limits are reached within the inlet of the connecting line The subtleties associated with a simple change in initial wall temperature should serve to illustrate the extra care that is sometimes needed in fluid system modeling especially when two phases are present Nonetheless you shouldn t arrive at the conclusion that two phase analyses are intrinsically difficult since they are performed routinely using SINDA FLU INT Rather you should note that the violence of the resulting transient in this particular example is due in part to the artificially contrived initial conditions The above problem takes less than 10 CPU seconds to solve on most workstations Interest ingly the liquid volume fraction in the line never grows beyond about two percent so annu lar and stratified flows predominate The rapid condensation in the line is partially offset by the compression heating of the fluid in Tank 1 which is the source of fluid entering the con necting line The model runs for about 39 seconds at which point the upper temperature and or pressure limits are exceeded and the logic stops the analysis Figure 5 presents the results Notice that the valve warms up faster than the inlet despite the fact that the inlet fluid is substantially warmer than the fluid near the valve The reason is that two phase heat transfer coefficients near the valve are orders of magnitude larger than the single phase film coefficients near the inlet
43. ignated in HEADER CONTROL DATA GLOBAL In the above block PL1 means the pressure in lump 1 TL the temperature in lump 2 etc Such variables are translated by SINDA FLUINT into a reference to a cell in a Fortran array While the details of the translation are usually not important it is necessary to know that such translations occur and that they can be customized and controlled The last line of Fortran logic checks to see if the temperature or pressure limits are ever exceeded and halts the transient if so by setting the end time TIMEND equal to the cur rent time TIMEN By default OUTPUT CALLS is executed before and after each steady state solution and as requested in transients You can customize the calling frequency if desired In fact the value of the output frequency and other control constants may be changed within logic blocks and therefore during a solution procedure In the above logic the answer will as accurate as the output interval about 1 2 second in this case More accuracy could have been achieved by defining TIMEND directly in CONTROL DATA using a spreadsheet relationship and avoiding Fortran logic TIMEND max chambrs TL1 chambrs TL2 gt 450 timen 59 3600 The chambrs PL1 gt 30 clause has been omitted in the above relationship for clarity The above logic would be automatically invoked at least every time step as part of the self resolv ing spreadsheet that operates concurre
44. in the standard library or you ll require a more complete perhaps supercritical description of a library fluid or perhaps you ll need a simpler description than is provided by the library All these purposes may be addressed by specifying user defined fluids To illustrate the use of user defined working fluids assume that the working fluid in the above sample problem is air instead of steam Air while not contained within the standard library is easily described as a perfect gas Perfect gases are signaled by a fake four digit ASHRAE identifier starting with 8 such as 8000 or 8765 Fluid descriptions are meant to be developed once and reused many times promoting the development of user libraries For this reason fluid descriptions are made in separate HEADER blocks A simple description of air is as follows HEADER FPROP DATA 8729 SI 0 0 C SIMPLIFIED AIR 1 ATM NEAR 300K 540R IN SI UNITS DEGREES K C MOLW 28 96 MOLECULAR WEIGHT CP 1 005E3 SPECIFIC HEAT K 26 1E 3 CONDUCTIVITY V 18 5E 6 amp VISCOSITY In the subblock formed by the HEADER FPROP DATA you name your ID for the desired fluid along with the local unit system In this case the fluid description is provided using standard SI units with degrees Kelvin The zero 0 0 at the end of the first line is the value of absolute zero ABSZRO in SINDA FLUINT in the local unit system If degrees centi grade had been desired the last number
45. ion ties require one or more duct like paths G e tubes or STUBEs to be named such that the program can use the shape and flow rate of the duct as part of the con vection calculation The thermodynamic state of the associated lump is used to determine whether standard single phase boiling or condensation correlations are used Fties Fties are similar to heat transfer ties above except that they interconnect lumps directly In other words they are used to simulate heat transfer within the fluid which is often only important either for high conductivity fluids such as liquid metals and for slow moving or stagnant flows For example fties can be associated with ducts tubes and STUBEs to easily simulate axial backconduction in such lines Interfaces Interfaces ifaces are specialized network elements enabling any two tanks finite control volumes to share a boundary This boundary might represent a liquid vapor interface or a piston but it might also represent an imaginary subdivision of a volume Usage Overview SINDA FLUINT is user extensible providing you with complete control over inputs out puts and solution procedures The program assumes very little about the problem at hand or which details are important to you as the analyst To use SINDA FLUINT correctly you must have questions you want answered and you must pose them in a way it can compre hend There are no cook book methods available the experience an
46. is the current pressure of lump 2 There are two types of paths classified by whether or not they neglect fluid inertia Tubes are flow paths in which fluid inertia is taken into account In other words the flow rate through a tube cannot change instantly during a transient event the forces acting on the flow passage friction pressure velocity gradients etc must accelerate or decelerate the fluid over time To model waterhammer tubes must be used Usually tubes represent a duct or a segment of a duct although more complicated components can also be simulated using tubes Tubes are described by a hydraulic diameter DH a flow area AF and a length TLEN As with lumps many other optional descriptors exist from the simple e g FK for added K factor losses to the sublime e g AM for virtual or added mass coeffi cient that is only applicable to two phase slip flow simulations Few of these variables will be needed in any single analytic case but all of them are available just in case Connectors are flow paths in which fluid inertia can be neglected In other words the flow rate through a connector is always at equilibrium with the forces on that path connectors represents a local steady state or time independent solution Connectors are further subdi vided into devices Devices are generic representations of common fluid system components such as pipe segments pumps turbines compressors filters orif
47. issolution evolution and slip flow velocities of each phase differ In the original problem statement the comparatively small volume of the line led quickly to the use of junctions instead of tanks to model that line However since the line will now be Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 28 of 42 t csr TECHNOLOGIES filled initially with vapor and will eventually fill with condensate that previous assumption is much less valid Because of the orders of magnitude difference in liquid versus vapor den sity the changes in mass associated with the fluid in the line is no longer negligible com pared to the rest of the system and tanks must now be used When tanks are used in two phase systems tubes become preferable to STUBEs because of enhanced stability as dis cussed in the full User s Manual When the steam from vessel 1 enters the cold line it rapidly condenses This deceleration of fast moving vapor to slow moving liquid counteracts friction and can easily result in a net pumping action or pressure recovery Also the flow resistance of liquid is lower than vapor at the same mass flow rate During the relatively violent event that follows the flow rate in the line quickly grows to very high values despite having a dead end in the form of a closed valve In some respects the presence of the piston serves only to initiate this action In fact if the wall temperatures were much colder flow rates can become so h
48. le individual paths and lumps in a macro to be edited The new block becomes HEADER OPERATIONS BUILDF ALL BUILD ALL CHAMBRS FK100 0 5 INITIALIZE INLET LOSS ON FIRST PATH CALL TRANSIENT START A TRANSIENT ANALYSIS Since OPERATIONS is executed sequentially the value of FK for path 100 in fluid submodel CHAMBRS CHAMBRS FK100 is initialized to 0 5 before the transient solution begins CHAMBRS FK100 could be simplified to FK100 if SINDA FLUINT knew which sub model name applied The default submodel name can be set using the DEFMOD command as follows HEADER OPERATIONS BUILDF ALL BUILD ALL DEFMOD CHAMBRS FK100 0 5 INITIALIZE INLET LOSS ON FIRST PATH CALL TRANSIENT START A TRANSIENT ANALYSIS Notice that submodel prefixes and DEFMOD commands were not necessary in the logic that was listed at the end of the OUTPUT CALLS block e g PL1 etc The reason is because that block was already specific to the submodel CHAMBRS If a wall node temperature had been referenced within that block or some variable from another fluid submodel then sub model prefix would have been required e g WALL T3 Complete Input File and Results The complete input file for the expansion of the original problem to include the mass of the tube wall is as follows HEADER OPTIONS DATA TITLE TWO CHAMBERS WITH VALVED CONNECTING LINE OUTPUT chambers out C HEADER REGISTER DATA Length 10 0 diam 0 5 12 0 area
49. lly invoke convection heat transfer calculations Such a tie requires the specification of a tie identifier a lump and a node to which it should attach and a tube or STUBE connector which should be used to provide flow rate shape and size information for the heat transfer passage C Tie 99 connects lump 10 to thermal node 1 in submodel wall and C path 10 an STUBE defines the shape size and flow rate T HIN 99 10 WALG 1 10 Alternatively an HTNS convection tie can be made This type of tie effects an LMTD log mean temperature difference solution which when used properly can yield better results in single phase models with coarser spatial resolution 1 e fewer nodes and lumps C Tie 99 connects PATH 10 to thermal node 1 in submodel wall T HTNS 99 10 WALL 1 The difference between these two types of ties is subtle HTN ties represent a more true finite difference lumped parameter assumption in order to capture an important property gradient more elements must be used When too few elements are chosen however the total heat transfer rate is typically underestimated In contrast HTNS ties are segment ori ented they assume an exponential temperature profile within each fluid section With HTN ties lumps are treated as subsections of a line With HTNS ties lumps are treated merely as end points as an inlet state and an outlet state The node is assumed to represent the aver age wall temperature along the entire
50. luid submodels HEADER OPERATIONS BUILDF ALL BUILD ALL CALL TRANSIENT START A TRANSIENT ANALYSIS If this new statement were missing and the wall model were not built into the current con figuration any input ties would be ignored and the previous adiabatic case would result Similarly if the wall model were built but the fluid model were not built the wall node would simply progress through the transient event without any change in temperature BUILD and BUILDF statements can be made repeated times within a single OPERATIONS as needed to analyze various combinations of configurations Axially Discretized Line Unfortunately because of the large temperature ranges in the model and the long thin aspect ratio of the line the assumption of an isothermal line wall may not be valid The line will therefore be subdivided into 5 smaller nodes numbered 1 through 5 each connected axially by linear conductors representing axial conduction along the line Even if this term is small SINDA nodes must be have at least one conductor so adding the axial conduction satisfies this requirement SINDA like FLUINT provides for GEN commands to generate multiple elements The newly updated WALL network blocks therefore become HEADER NODE DATA WALL eo GEN N N NINC Ti density Cp Length X sectional Area number GEN 1 5 1 tinitw dens cp length apipe 5 0 HEADER CONDUCTOR DATA WALL C GENERATE 4 CONDUCTORS 1 2 3 4 BETWEEN 5 NODES 1
51. m perature of lump 100 TL100 ever exceeds 212 degrees Alternatively you could simply supply the following definition of TIMEND using spreadsheet expressions TIMEND TL300 gt 212 TIMEN 1 0E30 Logical instructions are also the method by which you define the solution sequence and the output operations Additional logic may be inserted before during or after each network solution i e steady state or transient analysis as needed to tailor the execution Entire libraries of reusable auxiliary routines are often generated by experienced users You do not need to know much Fortran in order to use SINDA FLUINT You can perform straight forward analyses using a few simple commands such as CALL STEADY to request that a new steady state solution be performed However if you already know some Fortran or are willing to learn a few simple manipula tions you will find few limits to your ability to pose new problems to SINDA FLUINT Advanced Features This section lists some of the more advanced features that are contained in SINDA FLUINT many of which will not be described in this document Macrocommands Input commands exist that allow you to generate multiple network ele ments i e lumps paths ties as well as strings of such elements that represent discretized continuous flow passages duct macros Duct macros not only represent convenient means of quickly generating discretized models of lines for heat e
52. n Inputs may be specified not just as determinis tic points but rather as ranges or probabilistic distributions The code can then predict the chances that failure criteria will be exceeded In fact it can be combined with design optimi zation to synthesize a design based on reliability constraints including perhaps defining what tolerancing is required Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 9 of 42 v CsR TECHNOLOGIES FLUINT Sample Problem Basic This section develops a simple FLUINT model using the traditional text input file approach Modern graphical methods are available using Sinaps or Thermal Desktop which are described separately A template input file is available in the installation set but this sam ple will start from a blank slate Variations of this basic model will be made in later sections This model is developed in English units SI units may be used alternatively in the code Problem Description Consider two adiabatic 1 ft vessels connected by a 10 ft long by 1 inch inner diameter adi abatic line The vessels are filled with steam at 11 psia The first vessel initially at 200 F slightly superheated is fitted with a piston that begins to compress that vessel at time zero at a volumetric rate of 1 0 ft minute The second fixed volume vessel is initially at 400 F A partially closed valve is located in the line near the second vessel The valve is partially closed throat area of 0 1 in2 at
53. nd 3D problems Given appropriate inputs FLUINT can produce answers that are the same as those produced by codes that are based on finite volumes or finite differences Most FLUINT models are much more free form than such codes allow enabling the creativity and experience of the engineer to be exploited without having to resort to writing specialized single use computer programs FLUINT is a network style fluid flow simulator Although it can be used by itself for purely hydrodynamic analyses such as pump matching manifolding flow distribution or water hammer it can also be combined with SINDA thermal networks to simulate combined ther mal hydraulic systems including vapor compression cycles and heat exchangers You pose a problem by creating an arbitrary network of thermodynamic points lumps connected by fluid flow passages paths You may also define heat transfer routes ties between SINDA nodes and FLUINT lumps to simulate convection In addition you define an arbitrarily com plicated solution sequence perhaps providing auxiliary Fortran style logic and chooses the desired output frequencies and formats Inputs may be defined indirectly as algebraic func tions making the code a cross between a spreadsheet and a thermal fluid network analyzer for easy parametric analysis Although such networks are intentionally analogous to traditional thermal SINDA networks fluid flow networks are spe cialized to handle the extra equa
54. ndensible Mixtures This variation starts from the model that included the wall thermal submodel Variation 2 In the previous variation the original working fluid water fluid 6070 was replaced by a noncondensible mixture In this variation the effects of adding a noncondensible gas namely air to the original condensible water used as the working fluid in Variation 2 are explored The initial concentration of air in the gas is 0 1 meaning 10 by mass Many of the variations are similar to those chosen in Variation 3 1 the initial wall tem perature is lowered to 150 F to attempt to force condensation to occur 2 the internal line will be modeled with tanks and tubes instead of junctions and STUBE connectors to help track density changes and 3 ACCELZ is again set to gravity perpendicular to the line to enable better flow regime predictions Unlike Variation 3 however the valve is not closed For the air description we will use the simple one presented earlier except that a new data value DIFV the diffusion volume has been added as will be explained later HEADER FPROP DATA 8729 SI 0 0 C SIMPLIFIED AIR 1 ATM NEAR 300K 540R DIFV 19 7 DIFFUSION VOLUME MOLW 28 96 MOLECULAR WEIGHT TREF 300 0 S REFERENCE TEMPERATURE 9 CP 1 005E3 CPTC 0 03 SPEC HEAT AT TREF CP TEMP COEF K 26 18 3 KTC 8 0E 5 COND AT TREF PLUS TEMP COEF V 18 5E 6 VTC 0 051E 6 VISC AT TREF PLUS TEMP COE
55. ntly with SINDA FLUINT execution Solution Sequence Like OUTPUT CALLS the entire solution sequence is specified as a logic block meaning that you have complete control over program execution from start to finish The solution sequence is specified in a header block called HEADER OPERATIONS Instructions placed in this block will become the main driver for the run to be made It will be turned into a once through subroutine meaning that once the operations contained in that block have been executed the program will stop In this particular sample problem a single transient run is needed Such solutions are requested by calling single routines such as CALL TRANSIENT Before starting a transient you must select the problem end time TIMEND input as a large value if unknown TIMEND 59 0 3600 0 RUN FOR 59 SECONDS This statement may be placed either in the logic block HEADER OPERATIONS prior to the call to TRANSIENT or it may be placed in the data block HEADER CONTROL DATA GLOBAL where initializations may be made This data block just happens to have a format that looks like a logic block This is atypical Most fluid transients must start from a valid initial condition which is usually provided by a prior steady state analysis Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 15 of 42 CsR TECHNOLOGIES If instead a steady state solution had been requested you must first provide the maximum number of iterations
56. quid vapor interface in capillary structure NULL user defined interface Figure 1 Hierarchy of FLUINT Network Elements Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 6 of 42 v CsR TECHNOLOGIES Outputs may be user defined but generally include temperatures pressures and flow rates In its traditional form SINDA FLUINT is a batch style code that accepts an ASCII text input file and returns binary and text results files A Fortran compiler is required since each SINDA FLUINT run builds and executes a custom program as shown in Figure 2 Network Description Ties Nodes etc Operation Sequence Perform analysis Steady State Transient Map Network Units etc Spreadsheet Thermal Desktop Relationships Restart ett Fortran Logic Compiling Output Procedures Pre Linking Post What Processing Processing Processing When O DATA C OUTPUTS O PLOTS etc Control Parameters Sinaps Error Tolerance aA EZXY or Concurrent Logic Initialization Customizing Wrap up User Data Arrays Constants Figure 2 Basic Data Flow While SINDA FLUINT may still be employed in the traditional manner two graphical user interfaces GUIs are also available Sinaps retains the geometry free nature of SINDA FLUINT you sketch thermal and or fluid networks on the screen launch a run and postpro cess color plot etc results using such sketches Thermal Desktop is a geometric CAD based tool tha
57. r DEFF may also be specified for highly noncircular cross sections but this detail is neglected for simplicity DEFF defaults to the hydraulic diameter DH When convective heat transfer calculations are invoked FLUINT assumes by default that the entire wetted perimeter is available for heat exchange Since this is not true in this case a tie area fraction is applied as a correction reducing the heat transfer area by D D Dj Since the wall will cool down once the transient begins the flow rate through the recircula tion loop will gradually change How fast does the liquid water move through this loop com pared to the event duration Based on preliminary analyses the flow rate of the recirculation loop stays above about 60 lb hour meaning that the entire loop will be swept clear at least every 1 5 minutes or so Since this time scale is on the order of the entire tran sient event tanks should be used to represent the volumes within the recirculation line so as not to neglect fluid lag effects Similarly the ascending line will be subdivided to better monitor these effects since temperatures in that line have the most effect on the buoyancy driving force Actually since the line descending from the reservoir never changes tempera ture its volume can be neglected and it can be represented by a single path Since hydrodynamic transients are unimportant STUBE connectors may be used in place of tubes A plenum is naturally used to
58. raction of gaseous species X PPGX etc You may optionally choose to define the 1 2 or 3D coordinate location CX CZ and CZ of each lump and the body force gravity or other acceleration vector component along those axes ACCELX ACCELY and ACCELZ Gravity effects including thermosyphoning and buoyancy effects stratification of two phase flows etc can thereby be included as can time and direction dependent vehicle accelerations All such variables are available for inspection output and manipulation within concur rently executed user logic blocks and spreadsheet expressions as will be described later FLUINT abounds with such variables most of which the casual user need never know exist since most are either not frequently used or are calculated automatically by the program Nonetheless they are available in order to make the program completely user extensible since fluid system analyses are often full of specialized or custom equipment Paths Paths describe the means by which fluid flows from one lump to another Each path has a single characteristic mass flow rate FR Paths may represent duct segments valves pumps leaks etc Usually the path flow rate is a function of the fluid state and pressure drop between the endpoint lumps FR1 PL Ply where FR is the flow rate of the fluid flowing from lump 1 to lump 2 and PL is the current pressure of lump 1 and PL
59. resent the path A junction number 10 must now be added in order to connect the paths in series This junction represents the valve inlet Since the pressures in the tanks has been assumed to roughly equal choking detection and simulation can be disabled otherwise it will be applied by default Units of ft F hr and psia will be used Units of m C s and Pa could have been used alter natively as could absolute temperature units R or K FloCAD permits many more unit choices and conversions In a traditional text input file it is convenient to state the default conditions for lumps and paths These defaults will be used by all subsequent lump and path definitions unless spe cifically overridden by those definitions or replaced by new default values The way in which these defaults are stated might appear as C SET LUMP DEFAULTS LU DEF PL pinit 11 PSIA INITIAL PRESSURE TL tiniti AL 1 0 200 DEG F INIT TEMP ALL GAS VAPOR VOL 1 0 1 CU FT VOLUME IN TANKS C PATH DEFAULTS PA DEF FR 0 0 ZERO INITIAL flow rate DH diam 1 2 INCH DIA CIRC AREA DEFAULTED Notice the use of C in column 1 for comments The dollar sign may also be used for in line comments The PA and LU must occur in column 1 since they designate subblocks or multi line zones in which path or lump data may be defined Within each subblock inputs are column independent and are usually keyword driven e g DH
60. rodynamic transients such as acoustic waves the original problem set up must be changed slightly In this variation Ves sel 2 is pressurized to 30 psia but the valve is initially closed The piston is in the closed position meaning that the connecting line which is assumed to be at 11 psia and 200 F leads to a dead end The line is assumed to be adiabatic during this event The valve is opened rapidly within 10ms at time zero Since little is known about the resis tance of the valve at each stem position an orifice with linearly increasing hole area will be used to model the valve This situation is common enough that such capabilities are built into the ORIFICE device model as will be shown below The transient pressure response of the system is desired Since the piston is in the closed position the volume of vessel 1 is zero This is affected sim ply by changing Tank 1 into Junction 1 LU JUNC 1 VDOT 1 0 60 0 VESSEL 1 CLOSED END The connecting line is now dead ended Since some resolution is needed within the line it is convenient to start from the model used to illustrate heat transfer G e the model with a wall network presented as Variation 2 However for the purposes of this model the WALL network is no longer needed This can be affected simply by moving the BUILD statement in OPERATIONS after the call to the tran sient Alternatively the thermal submodel can be removed and the HX macro in FLOW DATA
61. rposes that are described later Therefore even if there is only one submodel in this sam ple problem it must be given a unique alphanumeric name just as each node must be given a numeric identifier that is unique within its submodel Using CHAMBRO as the sub model name yields the complete input block VALVE MODEL PATH 20 GENERIC LOSS DEVICE VALVE LOSS K FACTOR THROAT AREA VESSEL 2 INITIALLY AT 400 DEG F NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL Ur UY UF OF UF fk UF UU Sy A UF A HEADER FLOW DATA CHAMBRS FID 6070 WATER IS THE WORKING FLUID see INC file C SET LUMP DEFAULTS LU DEF PL pinit 11 PSIA INITIAL PRESSURE TL tinitl XL 1 0 200 DEG F INIT TEMP ALL GAS VAPOR VOL 1 0 1 CU FT VOLUME IN TANKS C PATH DEFAULTS PA DEF FR 0 0 ZERO INITIAL FLOW RATE DH diam 1 2 INCH DIA CIRC AREA UNLESS SPECIFIED C DEFINE MODEL LU TANK 1 VDOT 1 0 60 0 VESSEL 1 SHRINKING WITH TIME LSTAT STAG NEGLIGIBLE VELOCITIES TREAT PL AS TOTAL PE CONN 10 1 10 LINE PATH 10 FROM 1 TO 10 DEV STUBE SHORT TUBE DEVICE Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 13 of 42 t csr TECHNOLOGIES TLEN length LENGTH FK 0 5 INLET LOSS K FACTOR LU JUNC 10 LINE OUTLET VALVE INLET VALVE MODEL PATH 20 GENERIC LOSS DEVICE VALVE LOSS K FACTOR THROAT AREA PA CONN 20 10 2 DEV LOSS FEFK 20 0 AFTH throat LU TANK 2 Th Hime VESSEL 2 INITIALLY AT 400 DEG F LSTAT STAG
62. s Enables complete descriptions of all fluid properties as Fortran instructions including calls to third party property data bases Many full range descriptions of cryogenic fluids fire retardants the newer green refrigerants such as R134a and accurate descriptions of water are available in this format All fluids available in NIST s REFPROP program are available in this format which can include variations such as gas only real gas and thermodynam ically compressible liquid phases Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 2 of 42 CsR TECHNOLOGIES Once defined fluid descriptions can be stored separately for reuse in other models Very few users actually write their own fluid descriptions relying instead on pre existing fluid prop erty files created by CRTech and others Working fluids can exist in one or two phases within any fluid system Compressibility of the fluids is included though it can be optionally excluded for the liquid phase Up to 26 real or perfect gases and simple liquids may be mixed within a single fluid system along with one two phase condensible volatile substance Gases can dissolve into and evolve out of liquids and species can chemically react as well reacting flows The mass of each such constituent is conserved in such cases Lumps Lumps represent a point at which energy and mass are conserved Each lump has a single characteristic thermodynamic state temperature
63. s to be used to treat analytic and systematic uncertainties calibrate models to test data etc Often FLUINT is used in conjunction with CFD codes For example one might wish to cal culate the frictional or heat transfer characteristics of a nonstandard component e g a braided cable inside a pipe forming an annular passage a using 2D or 3D point analyses and then use FLUINT to perform analyses of system level interactions dynamics sizing and optimization etc Extrapolation and interpretation of test results rather than CFD predic tions is also a common usage Working Fluids Thermodynamic and transport properties viscosity and conductivity are included in the SINDA FLUINT package for 20 common fluids including ammonia propane and various refrigerants However most of the time you will provide descriptions of alternate working fluids as separately provided files as one of the following categories 1 Perfect gases although not necessarily calorically perfect gases Specific heat and transport properties can vary as functions of temperature These gases can also be used as solutes within mixtures 2 Simple liquids Specific heat density compressibility and transport properties can vary as functions of temperature 3 Simple two phase fluids Enables engineers with either limited thermodynamics backgrounds or limited property data to quickly describe a two phase fluid 4 Arbitrarily complex single or two phase fluid
64. ssel 1 would shrink to an illegal zero volume in 60 seconds In the default units of hours TIMEND 59 0 3600 0 RUN FOR 59 SECONDS Output Specifications Despite their occasional Fortran like appearance all of the previ ously described inputs are data blocks since they specify network or control data rather than execution instructions Unlike previously described inputs the desired outputs are specified by a logic block Logic blocks are pseudo Fortran listings that are converted into real Fortran by SINDA FLUINT and then compiled and executed To specify the desired outputs you supply the output operations to be performed by the code at predefined intervals 1 of the problem end time by default using canned routines and or user supplied output instructions HEADER OUTPUT CALLS CHAMBRS DESIGNATE OUTPUT OPERATIONS CALL LMPTAB S LUMP TABULATION CALL PTHTAB PATH TABULATION IF PLI GE 30 0 OR LOGIC TO STOP TRANSIENT INTEGRATION TLL GE 450 0 JOR TL2 GE 450 0 TIMEND TIMEN Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 14 of 42 t csr TECHNOLOGIES The routine LMPTAB tabulates lump properties such as temperature density and pres sures If a submodel is not named then the call outputs properties by default for all fluid submodels Otherwise individual submodels may be specified by their name PTHTAB functions analogously for paths These operations will be performed every OUTPTF hours as des
65. swer to this question depends on whether or not any gradients in significant properties temperature density etc are expected within that line Since there is no heating of the line the question becomes whether or not a significant pressure difference is expected to exist within the line An assumption is made and later verified that the pressure difference between the tanks will be small at all times Therefore a single path is used to represent the line Although the line is long and thin and therefore might be represented by a tube the inertial path in FLUINT it contains low density gas and therefore little mass Also the time scale of inertial effects is on the order of hundredths of seconds whereas the duration of the imposed event is on the order of tens of seconds or minutes Therefore the inertia of the line will be assumed to be negligible An inertia less STUBE connector will therefore be used to represent the line rather than a tube The losses in the partially closed valve together with entrance losses say K 0 5 can be added directly to this STUBE connector Duct frictional losses are by default calculated Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 11 of 42 CsR TECHNOLOGIES automatically according to an internal curve fit to the Moody chart However it is desired to distinguish the valve losses from the line losses so a separate path is used to represent the valve A simple LOSS connector is chosen to rep
66. t eliminates cumbersome generation of geometric factors by hand and enables interfaces to CAD and FEM software The FloCAD module of Thermal Desktop allows 1D FLUINT circuits to be built in 3D and to interconnect with 2D or 3D thermal mod els Engineers new to SINDA FLUINT are strongly encouraged to use Sinaps or Thermal Desk top to avoid having to learn the traditional text input file formats An introduction to Sinaps is available separately User Logic In addition to its geometry independence the feature that sets SINDA FLUINT apart from other analyzers is its extensive use of user logic and spreadsheet like interrelationships to both define and customize the solution approach In essence SINDA FLUINT uses Fortran as its command language although spreadsheet like interrelationships can alternately be used Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 7 of 42 t csr TECHNOLOGIES Take the simple example of defining the end of a transient event Assume that the duration of an event is unknown and may itself be the goal of the analysis For example you might wish to know how long a hot water valve can be opened until a vessel somewhere down stream achieves a certain temperature With SINDA FLUINT a simple line of Fortran like logic might be used to detect such an event and terminate the solution IF TL100 GT 212 0 TIMEND TIMEN where the problem end time TIMEND is set to the current time TIMEN if the te
67. that each steady state call may attempt before either convergence is achieved or the program gives up Each iteration represents a single pass through the solu tion equations with each nodal temperature being updated once The maximum number of iterations is a global control constant named NLOOPS Generally NLOOPS should be set to the maximum size you can afford a number that is normally estimated based on prior expe rience and knowledge of each model Generally the larger the model the more iterations will be required to solve it Since SINDA FLUINT can apply multiple submodels to any problem you must define the list of submodels that will participate in the next solution even if only one such submodel exists This list of fluid submodels is declared by a BUILDF statement of the format BUILDF config smli sm2 smN where config is the arbitrary user name for the current configuration or active subset of the master model and sm1 through smN are the names of fluid submodels comprising this list Note that the B in BUILDF must be placed in column 1 whereas other instruc tions in this and other logic blocks follow the Fortran column conventions Using a short cut to build all fluid submodels the complete input block becomes HEADER OPERATIONS BUILDF ALL CALL TRANSIENT START A TRANSIENT ANALYSIS The control constants OUTPTF and TIMEND were already initialized in CONTROL DATA GLOBAL Other Inp
68. tions and phenomena associated with single and two phase flows Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 1 of 42 lt csr TECHNOLOGIES Detailed Geometry At Your Discretion Unlike CFD codes SINDA FLUINT need not be geometry based Without the use of tools such as FloCAD this lack of geometry may make FLUINT more cumbersome to use than geometry based codes for some problems with clearly defined simple geometry In most cases however actual geometries are much more complex than need be represented for fluid flow and heat transfer solutions Geometry based 3D meshes often produce unnecessarily large and inappropriately detailed models that are not only slow to solve they can obscure results Lacking geometry makes SINDA FLUINT more flexible and more appropriate for undefined or changing designs high level modeling e g entire vehicle what if and sensi tivity analyses and for optimization and correlation calibration and statistical design tasks On the other hand if the thermal structural geometry is detailed and CAD drawings or structural finite element meshes are available then FloCAD enables 1D fluid networks to be generated and linked to such 3D thermal geometry without resorting to a full CFD solu tion This makes FloCAD both easy to learn and rapid to use for applications such as air or liquid cooled electronics packaging The resulting flow solutions are fast which enables SINDA FLUINT s Advanced Design module
69. ttle attention By default one of four regimes bubbly slug annular and stratified is predicted based on current flow conditions void fractions flow rates etc and orientations with respect to a body force if any Since no body force vector nor lump coordinates which provide orientation data for flow regime mapping have been provided a zero gravity flow regime map will result By applying a body force acceleration vector without providing lump coordinates all tubes and STUBE connectors are assumed to be perpendicular to the vector i e horizontal To specify that the acceleration along the Z axis is due to gravity the following line is used ACCELZ grav OR ACCELY OR ACCELX Units ft hr2 The above line may be placed in HEADER CONTROL DATA GLOBAL or in OPERA TIONS before calling the transient solution routine TRANSIENT Use gravsi when using metric units Otherwise vertical and slanted lines are modeled by providing the 1 2 or 3D coordinate locations for the endpoint lumps see Variation 6 along with the direction and magnitude of the acceleration You may optionally override the mapping processing and select a regime directly perhaps applying test results or other knowledge of the regime Flow regime mapping is often a prerequisite for even higher fidelity two phase modeling automatic calculation of heat and mass transfer coefficients in nonequilibrium tempera tures and volumes of each phase differ and d
70. u lar form as shown in the rotated sample output page This output contains the final page of LMPTAB and PTHTAB information As can be seen the limiting factor is the temperature of the second vessel which exceeds 450 F in 53 seconds The competing effects of com pressed gas and injected cold gas result in only a modest temperature rise in that volume As assumed the pressure differences between the two tanks are small Losses in the valve exceed line losses keeping the Reynold s number in the line mostly in the laminar regime throughout the event This model requires much less than a minute of CPU time on most machines FLUINT Sample Problem More Details In this section the previously defined sample problem will be reworked in various details illustrating key FLUINT features Variation 1 More on User defined Working Fluids and INSERT Directives 20 simple fluids form the standard library descriptions cover the liquid vapor and dome regions up to the critical pressure Even if Sinaps is not used CRTech provides a free plotter on the PC called EZXY t Results including number of iterations can vary slightly from version to version and even from host machine to host machine because of minor changes in internal numerical approaches round off errors etc Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 18 of 42 CsR TECHNOLOGIES However you will almost always require a fluid that is not contained with
71. um time step that can be taken in order to resolve such events CALL WAVLIM CHAMBRS DTEST where DTEST is a temporary Fortran variable containing the returned time step limit Usu ally a value smaller than this should be used for conservatism say DTMAXF 0 4 DTEST where DTMAXF is a control constant or user specified maximum allowable time step for fluid submodels Where should these statements be placed If they are placed in OPERA TIONS they will only be executed once and the data upon which WAVLIM is based may vary during the solution As an aid in such cases SINDA FLUINT provides additional logic blocks which are called at prescribed times during the solution The most common such block for fluid submodel manipulations is FLOGIC 0 which is called before each steady state iteration and before each transient time step HEADER FLOGIC 0 CHAMBERS CALL WAVLIM CHAMBRS DTEST DTMAXF 0 4 DTEST The frequency of which such blocks are executed assures that the calculations they perform are relevant and that the data they manipulate is fresh While ORIFICE connector devices can be used to model their namesakes then are also han dly for modeling leaks and gaps or orifice like passages Because you will often find it neces sary to model a pressure regulating valve or a valve opening or closing event but will not be supplied much information about the valve itself other than perhaps an open resistance
72. ut Sections Other data blocks are used to control and customize program execu tion such as naming the files to be used by SINDA for outputs and other purposes One such block which must always occur first within the input file is the OPTIONS DATA block HEADER OPTIONS DATA TITLE TWO CHAMBERS WITH VALVED CONNECTING LINE OUTPUT chambers out In the above block the file to use for program output is specified along with a title to appear at the top of each output page Complete Input File The complete input file defining the above problem is as follows HEADER OPTIONS DATA TITLE TWO CHAMBERS WITH VALVED CONNECTING LINE OUTPUT chambers out e HEADER REGISTER DATA length 10 0 diam 0 5 12 0 area 0 25 pi diam 2 throat 0 1 144 0 Namely columns 1 through 5 are reserved for numeric labels column 6 for continuation characters and columns 7 through 72 for the statement itself t Again an historical misnomer Constants are actually Fortran variables that whose value may be changed during the course of processor execution Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 16 of 42 t csr TECHNOLOGIES pinit 11 0 tiniti 200 0 tinit2 400 0 9 INSERT f6070_water inc HEADER FLOW DATA CHAMBRS PID 6070 WATER IS THE WORKING FLUID C SET LUMP DEFAULTS LU DEF PL pinit 11 PSIA INITIAL PRESSURE 9 TL tinitl XL 1 0 200 DEG F INIT TEMP ALL GAS VAPOR VOL 1 0 1 CU FT VOLUME IN TANKS C P
73. v CsR TECHNOLOGIES Introduction to FLUINT SINDA FLUINT is a comprehensive software package used by thousands in the aerospace energy electronic automotive aircraft HVAC and petrochemical industries for design and simulation of heat transfer and fluid flow problems It is the NASA standard analyzer for thermal control systems This document introduces FLUINT the fluid flow network capabilities FLUINT represents only part of the complete SINDA FLUINT package The thermal node conductor network capabilities SINDA are introduced in separately available documents Familiarity with the basics of SINDA is helpful A nongeometric sketchpad graphical user interface Sinaps is available but is docu mented separately as is a dedicated plotting package EZXY Other supporting codes include Thermal Desktop a CAD based and FEM compatible geometric pre and post pro cessor to SINDA that includes radiation calculations via the RadCAD module Thermal Desktop also features FloCAD a geometric GUI for FLUINT Other CAD based model preparation tools include CRTech SpaceClaim and Mesh Generation for SpaceClaim What is FLUINT FLUINT is the most powerful and featured general purpose thermal hydraulic analyzer that is commercially available FLUINT is classified as a control volume or lumped parameter code Although normally lim ited to internal one dimensional piping networks it can be used to solve certain classes of 2D a
74. would have been input as 273 15 SINDA FLUINT will then convert the fluid properties into the master model unit set if required In other words the unit system within a fluid properties block need not match that of the model in which it is used MOLW is the average molecular weight of the gas CP is the specific heat V is the dynamic viscosity and K is the thermal conductivity Since they are input as singular val ues in the above block these properties will remain constant for all temperatures and pres sures In order to make the properties vary linearly with temperature a reference temperature TREF can be specified along with a slope or temperature derivative for each property at that temperature HEADER FPROP DATA 8729 SI 0 0 C SIMPLIFIED AIR 1 ATM NEAR 300K 540R cS MOLW 25336 MOLECULAR WEIGHT TREF 300 0 S REFERENCE TEMPERATURE A certain computational overhead exists for fluid descriptions the wider the range of valid properties the more expen sive the analyses that use that description even if the full range is not used within a particular model Therefore an all liquid nonvolatile or all gas noncondensible analysis of a standard library fluid would execute considerably faster if the fluid description were simplified t Actually a perfect gas description of air is included in the installation set This discussion is for illustration purposes only Always contact CRTech before
75. xchanger segments manifolds etc they also automatically invoke more physics e g spatial accelerations or momentum flux terms than would individually input elements since the program knows they represent a continuous passage Exploitation of Symmetry Parallel and identical flow passages and subcircuits frequently occur in fluid systems Examples include ideally manifolded parallel lines FLUINT features convenient means of modeling one such passage and then specifying the number of such pas sages in the system This duplication feature allows symmetries in the system to be exploited in order to reduce model size Turbomachines Whether defined by single performance curves or families of curves pumps compressors and turbines can be simulated at the system level Other features allow model ing of internal secondary flows including rotating and shearing flow passages Phase specific Suction Any FLUINT path may be directed to extract only liquid or only vapor from upstream lumps when both phases are present This feature may be used to sim ulate the blockage of vapor in filters and capillary restrictions liquid vapor separators or stratification of two phase control volumes Introduction to FLUINT V5 6 Rev 1 12 13 12 Page 8 of 42 lt csr TECHNOLOGIES Capillary Devices Options exist for modeling capillary passages filters wicks and other bubble barriers as well as capillary evaporator pumps or devices that utilize

Introduction to FLUINT

Contents

Download Pdf Manuals

Related Search

Related Contents