User's Manual for US1 Module of the NUFT Code, Version
Contents
1. 36 decay Contaminant decay oo oona a 37 7 References 2 2 a 38 A The Mathematical Model 2002002 0 200 000 000000000000 39 A l Variable Switching oaa a 40 A 2 Numerical Discretization oaa aaa 40 B Sample Problems 2 aaa a 41 B 1 Flow Problem oaoa aaa e 41 usl tex US1 NUFT User s Manual March 20 1996 1 CONTENTS B 2 Transport Problem usl tex US1 NUFT User s Manual March 20 1996 1 INTRODUCTION 1 Introduction This manual describes the US1 module of NUFT which is a model for solving the flow and transport equations under isothermal conditions for unsaturated flow of a liquid phase The flow of the gas phase is neglected and is assumed to be at constant pressure The resulting governing equation often called Richard s equation The transport of a single dilute non volatile contaminant species in the aqueous phase can be optionally modeled The liquid density is assumed to be approximately a constant incompressible fluid with regards to gradients of mass fractions with regards to calculating molecular diffusive fluxes It is also assumed that the concentration of the contaminants are so low that the density of the liquid phase is essentially unaffected by the concentration of the contaminant Because of this assumption the flow and transport portions of the calculation in the code can be decoupled and performed sequentially NUFT Nonisothermal Unsaturated Saturated Flow and2. read restart file lt file name gt time lt restart time gt optional lt file name gt string name of the restart file created by the option output restart lt restart time gt real Time used to search in the restart file the first restart record with time lt restart time gt will be read in The initial time of the simulation run is set by time and overwrites the time read in through the restart file If time is not present the initial of the run is set to the time read in Overwriting Restart Data overwrite restart The overwrite restart statement is used to overwrite restart information read in from the read restart statement for selected elements or variables The overwrite restart options are identical to those for the state statement Restart Backup The model will periodically write out restart records so that the model can be restart in case of system failure The model alternately writes out to two restart files named lt prefir gt re0 and lt prefit gt rel where lt prefiz gt is set by output prefix Each file contains a single record previous records are overwritten The reason for using two files instead of a single one is to prevent losing a record if the system fails during a write The user must check the two files to see which has the record with the latest simulation time The model writes to a file at periodic intervals based on the wall clock time backup lt option gt optio
3. User s Manual for US1 Module of the NUFT Code Version 1 0 Single phase Isothermal Unsaturated Model John J Nitao Earth Sciences Department Lawrence Livermore National Laboratory date March 20 1996 time 3 30 PM file usl tex CONTENTS Contents 1 Introduction 2 2 2 2 3 2 How to Install the Model 2 2 0 200200 00200000 000000000020 5 3 How to Run the Model 2 2 0 200200 000000000002 ee 7 4 Input Format 2 20 00 00 000000 02 ee 8 5 Input Data for usip Flow Model 0 0 020000 0 0000000 10 time tstop dt dtmax stepmax Time step control 11 tolerdt reltolerdt Time step control 2 0 2 0 0000000000 12 tolerconv reltolerconv Newton Raphson convergence tolerances 13 genmsh Generate mesh 2 2 a 14 state Initial Conditions 0 0 0 0 0 17 restart Read and write restart information 0 0 0 0 0 000 18 rocktab Material properties 2 0 ee 20 srctab Sources and sinks 20 0 23 betab Boundary conditions aaa aaa a 24 output Output options 2 aaa 25 Other options oaaae 30 5 1 Notes Table Time Values a aaa aaa e 31 6 Input Data for usic Transport Model aaa aaa aa a 32 state Initial conditions ouaaa aaa 32 rocktab Material properties 2 auaa aaa ee 33 srctab Sources and sinks 20 0 34 betab Boundary conditions 0 aaa a 35 compprop phaseprop Component and phase properties
4. Newton Raphson convergence tolerances tolerconv lt tolercony S gt lt tolerconv H gt optional Sets tolerance for Newton Raphson iteration convergence criteria The model checks for convergence using both the tolerdt and reltolderdt parameters Conver gence is assumed if either one of the criteria is true for each primary variable lt tolerconv S gt real lt tolerconv H gt real Absolute criteria for convergence of the Newton Raphson iterations for so lution of the non linear implicit in time discretized balance equations is acheived if the magnitude of the changes in the primary variables satura tion S and head H meters from one iteration to the next is less than or equal to these values The default tolerance is 0 1 m for the head and 0 001 for saturation reltolerconv lt reltolerconv S gt lt reltolerconv H gt lt reltolercony S gt real lt reltolerconv H gt real Criteria for Newton Raphson convergence based on the maximum relative magnitude of changes in the respective primary variables Default values are 0 001 for pressure and 0 0 for saturation usl tex US1 NUFT User s Manual March 20 1996 13 5 Input Data for usip Flow Model genmsh Generate mesh genmsh coord lt mesh type gt down lt Vr gt lt Vy gt lt Vz gt dx lt dz 1 gt lt dz 2 gt dy lt dy 1 gt lt dy 2 gt dz lt dz l gt lt dz 2 gt mat lt el name prefiz gt lt mat type gt lt i0 gt
5. SlowbLiq lt slow lig gt usl tex US1 NUFT User s Manual March 20 1996 30 5 Input Data for ustp Flow Model Notes Table Time Values 5 1 Notes Table Time Values Currently the model may in some cases choose time steps which are so large that it overshoots sharp changes in a time table This may be a serious problem if for example we want to model a flux that is turned off suddenly A solution is to use the forcetimes command in output which forces the model to hit specified times in this case the times at which the boundary condition changes suddenly The last time value in table must be greater than or equal to the ending time of the simu lation as set by tstop or the model will abort usl tex US1 NUFT User s Manual March 20 1996 31 6 Input Data for usic Transport Model state 6 Input Data for ustc Transport Model Initial conditions state lt state var name gt lt method gt lt data gt lt state var name gt lt method gt lt data gt lt state var name gt The valid state or primary variable name is C which is concentration The concentration is in units of kg m mass of contaminant unit volume of phase if usip is used for providing the transport velocities and kg kg mass of contaminant unit mass of liquid phase if usnt is used usl tex US1 NUFT User s Manual March 20 1996 32 6 Input Data for usic Transport Model rocktab Material properties rocktab lt rock ty
6. The installation script then creates a directory with the name xxx equal to the NUFT usl tex US1 NUFT User s Manual March 20 1996 5 2 HOW TO INSTALL version number and copies the appropriate files there A file called nuft bat is copied to the nuft directory This is the file that your system executes when you type the nuft command which runs the model NUFT sample input files are also placed into the installed directory The script does not remove the directories containing older NUFT versions The user may delete older versions if they wish The installation script will give a message indicating whether the installation was successful After successful installation the model is ready to run See the next section in order to run the model usl tex US1 NUFT User s Manual March 20 1996 6 3 HOW TO RUN 3 How to Run the Model Steps to Run the Model 1 Install the code See the previous section on how to install the code 2 Specify the mesh The mesh can be created through either of the following methods e Using the genmsh option in the input file e Using an external program that creates a mesh file see the NUFT reference manual for description of the mesh file format The first option is recommended for new users 3 Create the input data file The input file is created using any ascii text editor An editor such as vi or emacs which signals matching parentheses is preferable You will need to read section
7. el var0 gt lt el var gt list of words list of element variables that will be outputted model specific lt con var gt word lt con var0 gt lt el varl gt list of words list of connection variables e g fluxes that will be outputted model specific lt outtimes gt list of t reals list of times at which output will be performed Model will reduce the time steps to meet these times lt index crange gt list of lists of integers specifies a list of connections Comprised of a list of list of integers with each list specifying a connection Each list of integers is of the form lt id gt lt j0 gt lt k0 gt lt il gt lt jl gt lt kI gt and denotes the connection between element i lt 710 gt j lt j0 gt k lt k0 gt and element i lt 11 gt 7 lt jl gt k lt k1 gt Positive flux denotes flux from the first element to the second lt out var name gt word name of output variable to be outputted by history option model specific lt element name gt string usl tex US1 NUFT User s Manual March 20 1996 27 5 Input Data for ustp Flow Model output name of element whose state or primary variable will be outputted lt index con gt list of integers specifies the single connection between the element i lt i0 gt j lt j0 gt k lt k0 gt and the element i lt il gt j lt jl gt k lt kI gt Positive flux denotes flux flowing from the first element to the secon
8. equal to the value lt real gt by replacing the above form kr lt kr func gt lt kr parameters gt by kr lt real gt US1 NUFT User s Manual March 20 1996 22 5 Input Data for ustp Flow Model srctab Sources and sinks srctab compflux comp water range lt elem range gt lt elem range gt table lt luz table gt compflux comp water range lt elem range gt lt elem range gt table lt luz table gt specifies the component mass flux into an element or range of elements through a table of source fluxes at specified points in time Linear interpolation is used for time intervals between the table values Positive flux is flux into an element negative flux is out of an element Flux is specified as volumetric flux cu m s at ambient conditions usl tex US1 NUFT User s Manual March 20 1996 23 5 Input Data for ustp Flow Model betab Boundary conditions bctab lt be name gt basephase liquid range lt elem range gt lt elem range gt tables S1 lt var table gt H lt var table gt factor water lt factor table gt lt be name gt basephase liquid range lt elem range gt lt elem range gt tables S1 lt var table gt H lt var table gt factor water lt factor table gt specifies boundary condition information for flow model usl tex US1 NUFT User s Manual March 20 1996 24 5 Input Data for ustp Fl
9. gt file ext lt out ert gt outtimes lt t0 gt lt t gt cumulative optional forcetimes outtimes lt t0 gt lt t gt restart file ext lt out ert gt outtimes lt t0 gt lt t gt extool file ext lt out ert gt index range lt i0 gt lt il gt lt j0 gt lt jl gt lt k0 gt lt kIl gt variables lt var0 gt lt varl gt outtimes lt t0 gt lt t gt specifies what variables will be dumped to files and the output format Any of the above output options field flux field etc are optional or can be present any number of times in any order including with regards to different variables See below regarding other options and important notes Output options are field output variables of selected elements at specified times flux field output fluxes between selected elements at specified times history output variables of selected elements at all times flux history output the fluxes between selected elements at all times not applicable to USNT model srcflux output component fluxes for a flux specified in srctab beflux output component fluxes flowing out of a set of boundary elements specified by bctab forcetimes force the model to hit the times specified in outtimes restart write out a restart record to a file that can be read in by restart option extool write out data in eatool format lt out ert gt string suffix of output file file ext is optiona
10. lt il gt lt j0 gt lt jl gt lt k0 gt lt kI gt lt el name prefiz gt lt mat type gt lt i0 gt lt il gt lt j0 gt lt jl gt lt k0 gt lt k gt anisotropic wrap around optional specifies the mesh geometry element material types and names lt mesh type gt word specifies the type of mesh that will be generated Valid options rect rectangular mesh cylind cylindrical mesh Let the three coordinates of our mesh system be x y and z If the mesh type is rect then z y and z are along the coordinate axes of a rectangular system If mesh type is cylind then the first coordinate is the radial distance r the second coordinate y is angle and the third coodinate z is the longitudinal axis lt Vir gt lt Vy gt lt Ve gt reals are the components of the vector pointing downward in the direction of the gravity vector The program will internally normalize the vector to unity Setting the components to all zero will turn off gravity in the model The vector is always with respect to a rectangular coordinate system X Y Z For a rectangular mesh the coordinate system coincides with the rectan gular coordinate system a y z of the mesh If the mesh is cylindrical the vector is with respect to a coordinate system X Y Z where X is the axis defined by 0 z 0 Y is the axis defined by 6 90 z 0 and the axis Z is defined by r 0 lt dx Il gt lt dr 2 gt reals are the subdivisions of the
11. on the input data syntax to understand the general format of the types of input data that goes into the input file For first time users it is easiest to edit the preexisting sample input files provided with the code distribution 4 Run the model Type nuft input file usl tex US1 NUFT User s Manual March 20 1996 7 4 HOW TO RUN 4 Input Format Before going further the user should read the sections on the input data syntax in the general users manual for NUFT The model can be used either to solve for flow of the liquid phase only or for simultaneous flow including transport of a contaminant Form of Input Data for Flow Only usip title run title output prefix prefix of all output files time initial time tstop ending time dt initial time step size dtmax Maximum time step stepmax Maximum no of time steps genmsh mesh specification state initial conditions rocktab soil property type output output specification srctab source term tables bctab boundary condition tables The denote data which will be explained later All input past a semicolon on a given line is treated as comments The above data units can occur in any order except for the word usip which must come first Either state or read restart must be present but not both The data items output srctab and bctab are a
12. 01im end output set material properties rocktab sandy material HI porosity 0 25 Kx 1 5e 5 Ky 0 Kz 0 pc peVanGen Sr 0 4 m 0 6 alpha 5 0 Sj 0 41 kr krVanGen Sr 0 4 m 0 6 moderate permeability material MOD porosity 0 3 Kx 1 5e 6 Ky 0 Kz 0 pc peVanGen Sr 0 4 m 0 3 alpha 1 0 Sj 0 41 usl tex US1 NUFT User s Manual March 20 1996 41 APPENDIX B SAMPLE PROBLEM kr krVanGen Sr 0 4 m 0 3 clayey material LO porosity 0 3 Kx 1 5e 7 Ky 0 Kz 0 pc peVanGen Sr 0 4 m 0 2 alpha 0 5 Sj 0 41 kr krVanGen Sr 0 4 m 0 2 pseudo soil type for source element above surface SRC porosity 0 99 Kx 1 e 2 Ky 0 Kz 0 pc 0 0 zero cap pressure kr 1 0 rel perm set to unity end rocktab set boundary conditions bctab put flux into the source element src basephase liquid range S 33 turn source on from 0 to 200 min and off after that factor water 0 0 1 0 200m 1 0 20im 0 0 1 e30 0 0 3 set element to constant head of 3 5 meters tables S1 0 1 0 1 e30 1 0 H 0 3 5 1 2630 3 5 keep water table elements at saturated conditions WT basephase liquid range W tables S1 0 1 0 1 e30 1 0 H 0 0 0 1 2630 0 0 33 end bctab 33 initial conditions state S1 by key S 1 0 I 1 0 H 0 2 M 0 3 L 0 8 W 1 0 H by key 0 0 note that bctab will o
13. March 20 1996 40 APPENDIX B SAMPLE PROBLEM B Sample Problems B 1 Flow Problem Description This sample problem is a two dimensional infiltration problem in a cylindrically symmetric system with point constant head source The model has eight layers with the top layer simulating the boundary condition at the top of the model Elements at the topmost layer are all inactive except for the innermost element which is the source kept at constant head The source is kept at a constant head of 3 5 meters for 200 minutes and then turned off afterwards The bottommost layer is kept at constant zero head and saturated conditions to simulate the aquifer The lateral boundary is kept at no flow conditions Input file file usi sam1 in sample problem usip name of flow model title pt21 run title tstop 23 1h stopping time is 23 1 hours stepmax 1000 Maximum no of time steps time 0 0 initial time dtmax 1 e30 Maximum time step dt 1m initial time step is 1 minute set output formats output outputs all liquid saturation into file pt21 S1 field format list range variables S1 file ext S1 outtimes O 39m 70m 102m 222m 287m 342m 23h output liquid phase flux from source into file pt21 influx flux history phase flux liquid index con 1 1 1 1 1 2 file ext influx force model to take small time steps when source shuts off forcetimes outtimes 200m 2
14. Transport model is a suite of multi phase multicomponent models for numerical solution of non isothermal flow and transport in porous media with application to subsurface contaminant transport problems These distinct models are imbedded in a single code in order to utilize a common set of utility routines and input file format Currently the code runs on the Unix and DOS operating systems Versions have been success fully compiled and tested for IBM PC compatibles Cray Unicos and the following workstations Sun Hewlett Packward IBM Risc 6000 Silicon Graphics Each set of related models is called a module and has its own user s manual which documents any particular features and input data specific to that module The NUFT Reference Manual Nitao 1993 documents the general numerical algorithms used and gives the documentation of the input to the model common to all or most modules including options not described in the user s manual for each module Available modules are e UCSAT unconfined and confined saturated flow model e USIP single phase unsaturated flow Richard s equation e US1C single component contaminant transport e USNT NP phase NC component with thermal option An integrated finite difference spatial discretization is used to solve the balance equations The resulting non linear equation is solved at each time by the Newton Raphson method Options for solution of the linear equations at each iteratio
15. ch that ambient pressure is H equal to zero The pressure head H is not initialized for the elements that are unsaturated Set H equal to zero for these elements The saturation 1 is set the desired initial saturation for those elements For elements that are saturated set H to the pressure head and S1 to any value greater than or equal to the maximum saturation The general rules for setting the initial values of H and S1 are as follows If for an element H is set to greater than zero then the liquid saturation value is overridden and set to the maximum saturation for that element Otherwise if S1 is set greater than or equal to the maximum saturation or if H is set to a number less than zero then S1 is set to the maximum saturation and H is set to zero Example state S1 by ztable height 33 above 3 WT S1 0 0 1 0 0 2 0 9 0 8 0 5 1 0 0 4 10 0 0 3 S1 by key bc 1 0 H by key 0 0 H by key bc 0 5 First S1 is initialized by interpolating from a table specifying values of saturation vs height above the water table Then saturation is for elements with names starting with the characters bc is set to unit The head is first initialized to all zeros and then the elements with names starting with bc are initialized to 0 5 meters so that they are saturated elements US1 NUFT User s Manual March 20 1996 17 5 Input Data for ustp Flow Model restart Read and write restart information
16. d lt bc name gt word name of boundary condition used in betab lt src name gt word name of source term used in srctab cumulative optional If present the cumulative flux is outputted If not present the instantaneous flux is outputted Note that cumulative fluxes are reset to zero at the beginning of a restart Notes 1 Instead of file ext lt out ext gt one can use file lt file name gt in order to explicitly specify the output file 2 Couttimes lt i gt lt ti gt lt t2 gt can be replaced by either of a Couttimes which means all times b triggers wake lt state var gt range lt el name gt lt el name gt lt op gt lt val gt cond lt state var gt range lt el name gt lt el name gt lt op gt lt val gt wake lt state var gt range lt el name gt lt el name gt lt op gt lt val gt cond lt state var gt range lt el name gt lt el name gt lt op gt lt val gt which checks to see at every time step if any of the triggers goes off A trigger goes off iff its wake condition lt value gt lt op gt lt val gt is true where lt value gt is the value of the state variable lt state var gt at the elements with names in the range of any of the pattern strings lt el name gt If the wake condition is true then the condition in the cond field is checked if true then the trigger goes off and output occurs Trig
17. eSe wi Vet Tie Jp 8 usl tex US1 NUFT User s Manual March 20 1996 39 APPENDIX A THE MATHEMATICAL MODEL A 1 Variable Switching When for an element the liquid pressure is negative py lt 0 the flow equation uses as the primary variable S If the pressure reaches zero the primary variable is switched to pe The primary variable is switched back to saturation if the two conditions are satisfied 1 pressure becomes negative and 2 there is a neighboring connected element that satisfied these two conditions in the previous time step A 2 Numerical Discretization An implicit in time scheme Euler s method is used for the flow equation 7 Voes Voes ALY ae 9 JENG where the phase mass flux is given by eji j Apek n i j e i ped Lig pijgbijl 10 Lij L LY 11 k u E hij 2 12 Lig k p ji Lig k m pe 2 Li pi Lij Pj Pij 1 Les 13 J At t 0 14 Superscripts refer to the time level and subscripts to element index Note that 9 is a non linear equation in the saturation at time level n 1 An implicit in time scheme is used for the transport equation n n n n n i n l opiso t Upesi A YY wi Ga P as JENG where the diffusive flux is given by Api D ahy 427 rw 16 ij Li Di J 17 LS Di LY D An option for an explicit in time scheme is also available usl tex US1 NUFT User s Manual
18. ers The element names will be of the form lt el name prefit gt lt i gt lt j gt lt k gt where lt i gt lt j gt lt k gt denote the i j and k indices word material type Each element in the specified 7 j k range will have this ma terial type see the explanation of index ranges below for examples Each material type must have a corresponding entry in the material type table in rocktab An exception are material types that are pre defined in the model At this time the only pre defined material type is NULL which completely removes the respective element or elements from the model An equivalent method of removing elements is the null blocks option describedin the NUFT Reference Manual Material types that are of the following form denote specific auxillary elements used for boundary conditions BC 1 xxx element for b c of first type BC 2 xxx element for b c of second type BC 3 xxx element for b c of third type An element for first and third type boundary conditions will have zero flow connection from the element centroid to the boundary of any neighboring non auxillary elements An element for a second type boundary condition will have flow connection set to unity so that the reciprocal of the per meability or hydraulic conductivity will be the resistance and the volume of the element is set to the average volume of neighboring non auxillary elements lt ill gt lt il gt lt j0 gt lt jl gt lt k0 gt l
19. gers can go off only once 3 One can use range lt element range gt instead of index range to specify elements by the names where lt element range gt is a pattern string 4 Similarly one can use crange lt element range 0 gt lt element range 2 gt in stead of index crange lt 10 gt lt j0 gt lt k0 gt lt il gt lt jl gt lt k gt tospecify ranges of elements Here lt element range 0 gt lt element range 1 gt specifies all possible connections between elements whose names match the pattern string lt element range 0 gt and the elements whose names match lt element range 1 gt usl tex US1 NUFT User s Manual March 20 1996 28 5 Input Data for ustp Flow Model output 5 One can use connection lt element name 0 gt lt element name 1 gt instead of index con lt i gt lt j0 gt lt kO gt lt il gt lt jl gt lt kI gt to specify a connection usl tex US1 NUFT User s Manual March 20 1996 29 5 Input Data for usip Flow Model Other options Other options Placing the statement sat Lower bound on into the input data unit for the flow model will tell the model to restrict fluxes coming out of an element if necessary in order that the saturation will stay above a certain value The minimum value of saturation lt slow lig gt is set by placing SlowbLiq lt slow lig gt into each material type record in rocktab Le rocktab lt rock type name gt
20. he liquid saturation is greater than the last entry SN the value is pcN lt S gt table entries for liquid saturation They must be in strictly increasing order lt kri gt table entries for capillary pressure head me ters The matrix potential function can be set to a constant function equal to the value lt real gt by replacing the above form pe lt pc func gt lt pe parameters gt by pe lt real gt where lt real gt is in head m word list of data units name of the relative conductivity function vs saturation function and its as sociated parameters Valid options for kr lt kr func gt lt kr parameters gt are kr krVanGen kr krTable lt krN gt m lt m gt Sr lt Sr gt for the van Genuchten formulation lt m gt parameter unitless lt Sr gt residual saturation table lt S0 gt Sr lt kr0 gt lt 51 gt Sr lt kri gt lt SN gt table look up option using linear interpolation If liquid sat uration is less than first entry SO then a value equal to kro is used If the liquid saturation is greater than the last entry SN the value is krN lt Si gt US1 NUFT User s Manual March 20 1996 21 5 Input Data for ustp Flow Model rocktab usl tex table entries for liquid saturation They must be in strictly increasing order lt kri gt table entries for relative permeability The relative conductivity function can be set to a constant function
21. if the parameter isot dir is present in data unit genmsh then the value of lt hycond 0 gt will be used for flow in the first coordinate direction for example z axis direction for rectangular coordinates and r axis for cylin drical lt Aycond 1 gt will be used for the second coordinate direction y or 0 and lt hycond 2 gt for the third coordinate direction z lt pe func gt word lt pc parameters gt list of data units name of the matric potential head vs saturation function and its asso ciated parameters the matrix potential is the approximately the same as capillary pressure except for saturations near residual Valid options for pe lt pe func gt lt pc parameters gt are pc pcVanGen m lt m gt alpha lt alpha gt Sr lt Sr gt Sj lt Sj gt for the van Genuchten formulation lt m gt parameter unitless lt alpha gt usl tex US1 NUFT User s Manual March 20 1996 20 5 Input Data for ustp Flow Model rocktab lt kr func gt lt kr parameters gt usl tex lt Sr gt lt Sy gt pc pcTable lt pcN gt parameter 1 meters residual saturation optional transition saturation below which a linear curve is used Default is Sj Sr 0 05 1 Sr table lt S0 gt Sr lt pce0 gt lt Si gt Sr lt pel gt lt SN gt table look up option using linear interpolation If liquid sat uration is less than first entry S0 then a value equal to pco is used If t
22. ipt will create directories bin and NUFT below your home directory unless they exist already The installation script creates a directory with the name xxx below the NUFT directory derived from the name of the compressed tar file xxx tar Z It then copies the compressed tar file and extracts its contents into this directory It then symbolically links the file nuft to the file nuft dist in the installed directory This is the file that your system executes when you type the nuft command which runs the model NUFT sample input files are also placed into the installed directory The script does not remove the directories containing older NUFT versions The user may delete older versions if they wish The installation script will give a message indicating whether the installation was successful After successful installation the model is ready to run See the next section in order to run the model Installation for IBM PC Compatibles under DOS The NUFT distribution for IBM PC Compatibles under DOS comes on a floppy disk with an installation script called install bat 1 Type a install bat with the floppy in the a drive 2 The directory nuft must be placed as part of your execution path by editing the PATH variable that is set in the autoexec bat file in the top directory if it is not already there For example PATH C C C DOS nuft The installation script will create directories nuft below your home directory unless it exists already
23. j 2 k 2 into 3 well element aquif silt 14 13 5 5 call lowest layer by different prefix The symbols nx ny and nz can be used anywhere in place of a number where an index is required The model interprets these to mean the number of subdivisions in the x y and z directions respectively anisotropic this parameter is optional The older form of the parameter isot dir can also be used instead Affects the choice of the permeability or hydraulic conductivity parameter See the documentation of the parameters Kx Ky and Kz below This parameter should not be present for models where isotropic permeability or hydraulic conductivity is desired unless the three permeability parameters for each material type are equal wrap around usl tex if present model wrap the grid around in the angular direction Le for each element at j 1 connect to the corresponding adjacent element at j ny where ny is the number of elements in the j direction Default is no wrap around US1 NUFT User s Manual March 20 1996 16 5 Input Data for ustp Flow Model state Initial Conditions state lt state var name gt lt method gt lt data gt lt state var name gt lt method gt lt data gt initialize state variables lt state var name gt usl tex The valid primary or state variable names are H pressure head above ambient meters Sl liquid saturation fraction Note that the datum of H is chosen su
24. l The prefix of the output file name will be made up of the prefix set by the output prefix parameter The suffix should be some mnemonic e g the name of the variable being outputted For example if output prefix is set to runA and file ext to C then output will go to file runA C If the file ext parameter is not present then the output will be placed into the main output file Output for different output options can share the same output file lt options gt word format of output Can be any of following list or by list output values at elements along with the element names contsac format by set list element values as a vector of the form n vl v2 vn usl tex US1 NUFT User s Manual March 20 1996 26 5 Input Data for ustp Flow Model output by x list x coord and value by y list y coord and value by z list z coord and value by ijk list i j k index and value by xtable format compatible with state using by xtable method user needs to comment out output header by ytable format compatible with state using by ytable method user needs to comment out output header by ztable format compatible with state using by ztable method user needs to comment out output header tabular multi column format contour format readable by the nview program for MS DOS lt index range gt list of integers range of elements specified by the 7 7 k indices lt comp gt word component name lt el var gt word lt
25. l compprop phaseprop Component and phase properties compprop comp liquid halfLife lt Thalf gt freeDiffusivity lt coef gt set component dependent properties for transport model lt Thalf gt lt coef gt real half life of contaminant units must be in seconds in the liquid phase if the statement decay on is present see subsection on decay If in addition the statement decay on solid on is present the half life also refers to decay of contaminant that is adsorbed the solid otherwise there is no decay on the solid real free diffusion coefficient in liquid phase m s phaseprop liquid tortuosity lt tort fun gt sets phase dependent properties for transport model lt tort fun gt usl tex word name of the correlation for the tortuosity factor r which multiplies the free liquid diffusion coefficient Currently the only legal value is Millington which specifies the Millington correlation T pS where is porosity and S is liquid saturation The diffusion flux is given by q Ser DVC where D is the free diffusion coefficient If a real number is placed instead of a word then that number is used for the tortuosity factor US1 NUFT User s Manual March 20 1996 36 6 Input Data for usic Transport Model decay Contaminant decay decay on optional if present then contaminant decays in the liquid phase with half life equal to the value set by halfLife i
26. ll optional the rest are required There are other optional data items not shown above but are described in the NUFT Reference Manual Nitao 1995 Flow and transport is solved sequentially That is at each major time cycle the code first solves for the flow of the phase and then solves for the transport of the contaminant in an alternating fashion The model has the capability to allow the transport calculation to take different time step sizes than the flow so that several time steps may be taken by the transport model in a single major time cycle if needed Form of Input Data for Flow and Contaminant Transport common data common to flow and transport title output prefix output ext genmsh time tstop usip flow input data dt dtmax stepmax state rocktab usl tex US1 NUFT User s Manual March 20 1996 8 4 HOW TO RUN output srctab bctab usic transport input data dt dtmax stepmax state rocktab compprop output srctab bctab The first principal data unit common holds input data that is common to both the the flow and transport models including the mesh generation data The second unit usip holds the input fo the flow model and except for data now in the the common unit is the same as for the flow only mode The third unit usic holds the input parameters fo
27. luxes at specified points in time Linear interpolation is used for the flux at time intervals between the table values The compf1lux option specifies the total mass flux of the component while phaseflux specifies the total mass flux of the phase Positive flux is flux into an element negative flux is out of an element Units of component fluxes set in comp flux table are in kg s while phase fluxes set in phase flux table are in units of cu m s and kg s respectively for usip and usnt providing the transport veloc ities The setcomp internal command specifies that the component fluxes in phaseflux are equal to the concentration of the respective element times the phase flux Alternatively a table of values can also be specified by the command setcomp lt comp name gt table lt t0 gt lt C0 gt lt tl gt lt C gt usl tex US1 NUFT User s Manual March 20 1996 34 6 Input Data for usic Transport Model betab Boundary conditions bctab lt be name gt basephase liquid range lt elem range gt lt elem range gt _ tables C lt var table gt factor comp lt factor table gt lt be name gt basephase liquid range lt elem range gt lt elem range gt _ tables C lt var table gt factor comp lt factor table gt specifies boundary condition information for the transport model usl tex US1 NUFT User s Manual March 20 1996 35 6 Input Data for usic Transport Mode
28. mesh in the direction of the first coordinate z For a rectangular mesh the subdivisions are in the increasing direction of the rectangular coordinate system of the mesh For a cylindrical mesh the subdivisions are in the increasing r direction Numbers that are repeated can be abbreviated for example 3 5 0 would stand for three repeats of the numeral 5 that is 5 0 5 0 5 0 lt dy 1 gt lt dy 2 gt reals usl tex US1 NUFT User s Manual March 20 1996 14 5 Input Data for usip Flow Model genmsh are the subdivisions of the mesh in the direction of the first coordinate x For a rectangular mesh the subdivisions are in the increasing y direc tion For a cylindrical mesh these parameters set the subdivisions in the increasing direction Angular units are in degrees The cylindrical mesh wraps completely around the 0 direction only for the case of a single angular subdivision whose value is equal to dy 1 360 lt dz l gt lt dz 2 gt reals are the subdivisions of the mesh in the increasing z direction for either a rectangular or cylindrical mesh system lt el name prefir gt word lt mat type gt prefix of element name Each element in the specified t j k index range will a name comprised of this prefix and a number associated with it as the suffix see the explanation of index ranges below for examples This prefix can be used to specify a range of elements in the state output and other input paramet
29. n are direct banded solution and preconditioned conjugate gradient method with various preconditioning schemes The model can solve one two or three dimensional problems Future plans include incor poration of capillary hysteresis non orthogonal mesh discretization finite elements non linear solid sorption isotherms and chemical reactions The first stage of code verification with one dimensional problems has been completed Lee et al 1993 and further verification efforts are planned The distinct models in the code utilize a common set of utility routines and input file format The various models are essentially isolated from each other and hence future models can be added without affecting existing models This also allows for easy maintenance and future enhancements Global variables in the code are virtually non existent The code is written principally in the C language Input data is in the form of that used by the lisp language An internal lisp interpreter for the Scheme dialect of lisp is part of the simulator whose purpose is usl tex US1 NUFT User s Manual March 20 1996 3 1 INTRODUCTION to read the input data file and the internal data files containing default input data values It also performs data checking This manual is self contained and describes a minimal set of the most commonly used input parameters necessary for the use of this module The NUFT reference manual Nitao 1995 contains generic inpu
30. n by POPE _ Tp VAII w 1 where Fickian laws for dispersive and diffusive fluxes are given by Te DV w 2 IP DP Ve 3 and Darcy s law gives 516Vi HCD op peg V2 4 The capillary pressure relationship when the flow is unsaturated is given by pe pe Se 5 where we take the pressure datum is chosen such that the ambient gas pressure is zero pressure The various variables are defined as w mass fraction of y component in phase Se liquid saturation porosity pe liquid density V liquid phase velocity J hydrodynamic dispersive flux I molecular diffusive flux D dispersion tensor Dy diffusion coefficient ko permeability function Le liquid phase viscosity pe liquid phase pressure De capillary pressure function In addition we have the constraint we tu 1 6 Summing the two balance equations gives the balance equation for the liquid phase Jo peSe ot We now make the assumption that the contaminant is dilute i e w lt w so that all of the terms in the balance equation are not functions of w This equation can be solved for Se or pe as the primary variable All constitutive relationships are functions of the primary variable V peSeV 7 The transport of the contaminant obeys the balance equation ObpeSews ot Numerically we first solve at each time step Eq 7 Then solve Eq 8 using the flow velocity V and saturation S calculated from Eq 7 V op
31. n compprop If not present of if this statement is replaced by decay off then there is no decay of contaminant decay on solid on optional if present then contaminant adsorbed onto the solid phase will decay with half life equal to the value set by halfLife in compprop The option decay for decay in the liquid phase must be also be on in order to have decay on the solid If the decay on solid is not present of if this statement is replaced by decay on solid off then there is no decay of contaminant adsorbed on the solid usl tex US1 NUFT User s Manual March 20 1996 37 7 REFERENCES 7 References Lee K A Kulshrestha and J Nitao Interim Report on Verification and Benchmark Testing of the NUFT Computer Code Lawrence Livermore National Laboratory Report UCRL ID 113521 1993 Nitao J J Reference manual for the NUFT flow and transport code Lawrence Livermore National Laboratory Report UCRL ID 113520 1995 usl tex US1 NUFT User s Manual March 20 1996 38 APPENDIX A THE MATHEMATICAL MODEL A The Mathematical Model We derive the balance equations There are two components water w and contaminant c in a single liquid phase The gas phase is assumed to be at a constant spatially uniform pressure conditions The contaminant in the water phase is non volatile The solid phase is assumed nondeformable The balance equations for water and contaminant components in the liquid phase are give
32. nal lt option gt word if set to on then the model will periodically write backup restarts If set to off the model will not do backups Default is on backup period lt backup lime gt optional usl tex US1 NUFT User s Manual March 20 1996 18 5 Input Data for ustp Flow Model restart lt backup time gt t real wall clock time period for model to perform backup restarts Default value is 10m i e 10 minutes usl tex US1 NUFT User s Manual March 20 1996 19 5 Input Data for ustp Flow Model rocktab Material properties rocktab lt rock type name gt porosity lt phil gt Kx lt hycond 0 gt Ky lt hycond I gt Kz lt hycond 2 gt pe lt pe func gt lt pce parameters gt kr lt kr func gt lt kr parameters gt lt rock type name gt porosity lt phil gt Kx lt hycond 0 gt Ky lt hycond I gt Kz lt hycond 2 gt pe lt pe func gt lt pce parameters gt kr lt kr func gt lt kr parameters gt set material properties for flow model lt rock type name gt word the name of the material lt phi gt real the fractional porosity of the material Must greater than zero lt hycond 0 gt real lt hycond 1 gt real lt hycond 2 gt real specifies the hydraulic conductivity m s of material property The model normally will use only the value of lt hycond 0 gt The other two values are ignored although they must be present and can be set to zero However
33. ng of transport input data usic stepmax 10000 dtmax 1 30 dt 3 output output concentrations C into file pt21 C field format list file ext C range variables C outtimes O 39m 70m 102m 222m 287m 342m 23h state C by key RN 0 0 initial concentrations are zero rocktab HI porosity 0 25 Kd 0 conservative tracer MOD porosity 0 3 Kd 0 LO porosity 0 3 Kd 0 SRC porosity 0 99 Kd 0 compprop comp liquid freeDiffusivity 1 e 9 phaseprop liquid tortuosity Millington bctab source basephase liquid factor comp 0 1 0 3 shut off solute source 200m 1 0 33 after 200 m inc 201m 0 0 33 turn off diffusion flux 1 e30 0 0 from source element us tex US1 NUFT User s Manual March 20 1996 45 APPENDIX B SAMPLE PROBLEM usl tex range S tables c o0 1 1 e30 1 0 bottom basephase liquid range W tables C 0 0 0 1 630 0 0 keep WT at zero concen US1 NUFT User s Manual March 20 1996 46
34. ow Model output Output options Valid output variables for the uslp module are H liquid head m S liquid saturation mc volumetric moisture content Kx Ky Kz hydraulic conductivities m s see rocktab for definitions logiOKx logiOKy logi0Kz logarithm base 10 of hydraulic conductivities log10 ms s porosity porosity output field file ext lt out ert gt format lt options gt index range lt i0 gt lt il gt lt j0 gt lt jIl gt lt k0 gt lt kI gt lt i0 gt lt il gt lt j0 gt lt jl gt lt k0 gt lt kI gt variables lt el var0 gt lt el vari gt outtimes lt t0 gt lt ti gt flux field file ext lt out ert gt format lt options gt index crange lt i0 gt lt j0 gt lt kO gt lt il gt lt j1l gt lt kI gt lt i0 gt lt j0 gt lt k0 gt lt tl gt lt jl gt lt kI gt variables lt con var0 gt lt con vari gt outtimes lt t0 gt lt ti gt history file ext lt out ert gt variable lt el var gt element lt element name gt flux history file ext lt out ert gt variable lt con var gt index con lt i0 gt lt j0 gt lt k0 gt lt il gt lt jl gt lt kI gt srcflux comp lt comp gt file ext lt out ert gt Couttimes lt t0 gt lt li gt cumulative optional usl tex US1 NUFT User s Manual March 20 1996 25 5 Input Data for ustp Flow Model output bcflux comp lt comp
35. pe name gt porosity lt phi gt Kd lt Kd gt lt rock type name gt porosity lt phi gt Kd lt Kd gt set material properties for transport model lt rock type name gt word the name of the material lt phi gt real the fractional porosity of the material Should be same value as set for flow model lt Kd gt real dimensionless solid sorption coefficient defined as the quantity ppka where pp is bulk dry density g cu cm kg is the standard solid sorption coefficient ml g and is porosity usl tex US1 NUFT User s Manual March 20 1996 33 6 Input Data for usic Transport Model srctab Sources and sinks srctab compflux compflux phaseflux phaseflux comp lt comp name gt name lt fluz name gt range lt elem range gt lt elem range gt table lt comp flux table gt comp lt comp name gt name lt fluz name gt range lt elem range gt lt elem range gt table lt comp flux table gt phase lt phase name gt name lt fluz name gt range lt elem range gt lt elem range gt setcomp internal table lt phase flur table gt phase lt phase name gt name lt fluz name gt range lt elem range gt lt elem range gt setcomp internal table lt phase flur table gt specifies the flux for component lt comp name gt into an element or range of elements through a table of source f
36. pt21 run title tstop 23 1h stopping time is 23 1 hours time 0 0 initial time genmsh coord cylind note that increasing z coordinate will be downward down 0 0 1 33 set subdivisions in radial r direction dx 0 05 0 05 radius source 0 1 r 0 15 0 2 r 0 35 0 4 33 r 0 75 0 5 33 r 1 25 0 5 0 5 0 5 note nr 8 set angle subdivision dy 360 33 set subdivisions in z direction dz 0 01 set first row of elements which are inactive except for i 1 for surface flux 1 0 sandy material layer 1 0 moderate material layer 1 0 sandy material layer 1 0 Clay material layer 1 0 sandy material layer 0 01 water table k 7 set material type and element name prefix mat S SRC 11 11 11 source element I NULL 28 11 11 inactive elements H HI 18 11 2 2 sandy layer M MOD 18 11 33 moderate perm H HI 18 11 4 4 sandy usl tex US1 NUFT User s Manual March 20 1996 44 APPENDIX B SAMPLE PROBLEM 33 33 be usip 33 33 333 L LO 18 11 55 33 clay H HI 18 11 66 sandy W HI 18 11 77 water table elements 33 end genmsh end of common data ginning of flow model input 3 name of flow model stepmax 1000 maximum no of time steps dtmax 1 e30 Maximum time step dt 1m initial time step is 1 minute the rest of the data for the flow model is same as the above flow only problem end of flow model input 33 beginni
37. r the transport model The two data units phaseprop and compprop are will be described in a later section of this manual usl tex US1 NUFT User s Manual March 20 1996 9 5 Input Data for ustp Flow Model 5 Input Data for ustp Flow Model In this section we describe the input data needed to run the uslp model Some of the data requirements are shared by other models Not all of the options are given here for such types of data but are documented in the NUFT Reference Manual usl tex US1 NUFT User s Manual March 20 1996 10 5 Input Data for ustp Flow Model Time step control time lt start time gt lt start tume gt t real initial starting time of run For restarts this time is not needed if restart is set in which case the initial time will be set to the time read in through the restart file If time appears this overwrites the time read in tstop lt stop time gt lt stop time gt t real run will stop when it reaches this time dt lt initial time step gt lt initial time step gt t real initial time step size dtmax lt mazr step size gt lt max step size gt t real maximum time step size stepmax lt maz steps gt lt maz steps gt anteger maximum number of time steps run will stop if this will be exceeded usl tex US1 NUFT User s Manual March 20 1996 11 5 Input Data for ustp Flow Model tolderdt reltolerdt Time step control tolerdt lt tolerdt S g
38. t lt tolerdt H gt Set absolute tolerances for time step control lt tolerdt S gt lt tolerdt H gt reltolerdt real real The model will control time step such that the magnitude of changes in the primary variables S fraction and head H meters from one time step to the next so they do not exceed the specified values As a rule of thumb the tolerances should be about one to two orders of magnitude larger than the desired accuracy of the primary variable Default tolerance for pressure is 10 m pressure and 0 35 for liquid saturation lt reltolerdt S gt lt reltolerdt H gt Set relative tolerances for time step control The model calculates a time step based on the absolute tolerance set in tolerdt and another time step based on relative tolerances set in reltolderdt and uses the larger of the two lt reltolerdt S gt real lt reltolerdt H gt real usl tex The model will also control time step such that the relative magnitude of changes in the primary variables liquid saturation S and head H from to the next does not exceed the specified values For each primary variable the model calculates the largest magnitude over all the elements in the entire problem domain and then multiplies by the specified relative tolerance to get a maximum change Default tolerance for head is 0 1 and 0 0 for saturation US1 NUFT User s Manual March 20 1996 12 5 Input Data for ustp Flow Model tolerconv reltolerconv
39. t ki gt integers usl tex specifies a range of elements using the 7 7 indices of the element in the direction of the x y z directions Used to specify the range of elements for setting element name prefixes lt el name prefiz gt and the material types lt mat type gt A specification record of the form lt el name prefir gt lt mat type gt lt i0 gt lt tl gt lt j0 gt lt jl gt lt k0 gt lt ki gt will overwrite the effect of any previous records see example below In this way one can specify non rectangular regions to have the same element prefix or material type The z j k indices are consistent with the dx dy and dz parameters That is US1 NUFT User s Manual March 20 1996 15 5 Input Data for usip Flow Model genmsh the dx parameter specifies the z dimension of the elements for i 1 2 in that order Similarly dy specifies the y dimensions of the elements for j 1 2 and dz specifies the z dimensions for k 1 2 Example e silt 2 4 3 10 1 8 This example specifies that the elements in the index range i 2 to 4 j 3 to 10 and k 1 to 8 will have element names e1 e2 and will be of material type silt Example consider a mesh with 4 elements in the direction 3 elements in the j direction and 5 elements in the amp direction mat s silt 14 13 15 3 first set all elements 33 to silt c clay 14 13 2 2 make layer k 2 into clay w well 22 33 22 make i 2
40. t options common to all of most of the modules It also contains further input options not given in this manual acknowledegments The author wishes to thank the management of the Environmental Restoration Division at the Lawrence Livermore National Laboratory LLNL for supporting the documentation and verification of the NUFT code The basic concepts of the code was developed under the funding of the LLNL Institutional Research and Development program usl tex US1 NUFT User s Manual March 20 1996 4 2 HOW TO INSTALL 2 How to Install the Model Installation onto a Unix System The distribution of NUFT for the Unix operating system Cray Unicos and unix workstations comes with a C shell installation script called INSTALL SCRIPT and a compressed tar file The name of the compressed tar file is of the form xxx tar Z where xxx are characters referring to the NUFT version number Copy these two files to any directory for example your home directory 1 Type csh INSTALL SCRIPT tar file 2 The directory bin just below your home directory must be placed as part of your execution path by editing the PATH variable that is usually set in the login file in your home directory if it is not already there For example PATH HOME bin bin usr bin ucb bin 3 The installation script and the tar file can be deleted after successful installation See below for the location of the tar file after installation The installation scr
41. verwrite these values by any set there end state generate mesh genmsh coord cylind note that increasing z coordinate will be downward usl tex US1 NUFT User s Manual March 20 1996 42 APPENDIX B SAMPLE PROBLEM down 0 0 1 33 set subdivisions in radial r direction dx 05 0 05 radius source 0 15 0 35 0 75 1 25 HHHHO u oo o 0 000 0 nonna eNe note nr 8 set angle subdivision dy 360 set subdivisions in z direction dz 0 01 set first row of elements which are inactive except for i 1 for surface flux 1 0 sandy material layer 1 0 moderate material layer 1 0 sandy material layer 1 0 Clay material layer 1 0 sandy material layer 0 01 water table k 7 set material type and element name prefix mat S SRC 11 11 11 source element I NULL 28 11 11 inactive elements H HI 18 11 2 2 sandy layer M MOD 18 11 33 moderate perm H HI 18 11 4 4 sandy L LO 18 11 55 33 clay H HI 18 11 66 sandy W HI 18 11 77 water table elements 33 end genmsh 333 end of input usl tex US1 NUFT User s Manual March 20 1996 APPENDIX B SAMPLE PROBLEM B 2 Transport Problem Description same two dimensional infiltration problem as above but with transport of a conservative tracer in the source water Input file 3 file ustsam2 in sample problem beginning of common data common title
Download Pdf Manuals
Related Search