3.2.2. Forward Input File
The forward problem is solved using the executable program h3dtd_v2.exe. The lines of input file are as follows:
Line # |
Description |
Description |
---|---|---|
1 |
path to tensor mesh file |
|
2 |
conductivity model |
|
3 |
susceptibility model |
|
4 |
defines the survey geometry |
|
5 |
defines transmitter waveform and time-stepping |
|
6 |
defines the observation times for predicted data |
|
7 |
choose Pardiso or MUMPS solver |
|
8 |
store factorizations in RAM or write to disk |
3.2.2.1. Line Descriptions
Tensor Mesh: file path to the tensor mesh file
Conductivity Model: Defines the conductivity model for the forward problem. There are two options:
FILE filepath: The user enters the flag FILE followed by the path to the conductivity model file
VALUE val: The user enters the flag VALUE followed by a value representing the conductivity of a wholespace model; e.g. we are modeling a wholespace if this option is used.
Susceptibility Model: Defines the conductivity model for the forward problem. There are two options:
0: A flag of 0 is entered if the susceptibility within the domain is zero
FILE filepath: The user enters the flag FILE followed by the path to the susceptibility model file
VALUE val: The user enters the flag VALUE followed by a value representing the susceptibility of a wholespace model; e.g. the same susceptibility value is assigned to every cell
Locations File: Provide the path to the locations file
Wave File: Provide the path to the wave file. This file defines the transmitter current and time-stepping for the problem.
Time Gates File: Provide the path to the time gates file. This file defines the times at which the TEM data are measured for the forward problem.
Solver: Define the direct solver that will be used to factor and solve linear systems. Enter one of the following flags:
FACTOR_PARDISO: Factor and solve linear systems with Pardiso solver
FACTOR_MUMPS: Factor and solve linear systems with MUMPS solver
Memory Options: Enter one of the following flags:
FACTOR_IC: Store factorization in the computer’s RAM. This options is much faster but can only be used on problems of a reasonable size
FACTOR_OOC: Writes the factorizations of linear systems to disk. Slower but capable of solving much larger problems.