A1: If you are targeting a multi-phase fluid system, it is important to fully satisfy the mass conservation equation for each fluid phase. Finite element method (FEM) can solve the mass conservation equation as close to zero for the weighted residual throughout the target area, however, it is difficult to guarantee the mass conservation of each and every fluid element in the model locally. Integrated Finite difference Method (IFDM) is adopted in GETFLOWS for each fluid element in the “input (inflow)” and “out (outflow)” boundaries where complete mass balance derived from changes in storage capacities both in locally and globally for gas, liquid and NAPL phases. It can solve the mass balance numerically close to zero as the residual of the whole system. It has been used historically in the field of petroleum engineering with high numerical accuracy, and it has a number of achievements in terms of robustness and stability of the numerical solution
A2: The general differential grid-block is regularly distributed pattern of grid points. “Corner Points” means the 8 vertices of a 3D hexahedron grid-block. When the 3D coordinates in X, Y, Z directions are considered the vertices of each 3D hexahedron grid-block can be set up freely. More than that deformed shaped hexahedron grid-blocks are possible to create. This is called the corner-pointed finite difference scheme.
A3: If you simulate a saturated flow or a general saturated-unsaturated groundwater flow problem, gas phase is not very much important to consider. Since GETFLOWS tries to reproduce the geospheric fluid behavior as natural as possible, we use a methodology to reproduce the descriptive simultaneous movement of air and water in the subsurface.In the case of simulating the underground injected gas flow and pore pressure variation due to atmospheric pressure changes you have to consider the gas phase in your model.
A4: GETFLOWS uses Preconditioned Conjugate Residual method (PCR) as the solver. The application of preprocessing techniques named as Nested Factorization and Successive Locking Process makes it possible to achieve high-performance computation capability with practical level of calculation speed and numerical stability.
A1: The modeling area should include watersheds, rivers, impermeable rocks, and so on. The model area should be large enough not have entry/exit points of geospheric fluids. Therefore usually vertical boundaries are set as no-flow boundaries at considerable distance from the dynamic area within the model. If you have reliable field data, you can apply constant pressure and constant flow as boundary conditions appropriately. Most of the commercial codes require explicit boundary conditions for whole system. However, in GETFLOWS the approach for boundary conditions is significant as we the no flow boundary condition at a distance far away from the calculation region.
A2: Topography, land use, and hydrological data are often freely available to receive from several databases which are managed by several institutions. Underground geological data is more difficult to collect usually. Although it is hard to estimate the detailed geological structure at your model area without research boring, you can estimate roughly the geological structure based on sedimentology and pale environmental study. Another method to estimate the hydro-geological structure is checking the difference between simulation results and observation data of river flow rates and groundwater levels with model tuning. The simulation model can be continuously developed by such feedback on the comparison between simulation results and real world observations.
A3: Firstly, create a 2D grid model considering the topographical relief, rivers and surface geology. Land use information can also be included to the 2D grid model. Afterwards, the 2D grid is extended to the depth direction (Z direction) and it is divided into several levels by considering the geological layer boundaries as well as the position of underground structures.
A4: It depends on the hydrologic properties of faults but generally reproduced as stair like grid long the depth. Fault fracture zone and internal clay materials can also be considered in the model. If you want to model the flow through a fault (internal flow within the fault) then you should specify permeability and porosity into the individual grids along the fault separately.
A5: There are no limitations on the temporal/spatial resolution and scale of the model. The spatial scale of a few mm for indoor test to a few thousand km for large circulation system and the temporal scale of a few second to a few thousand years can be modeled in GETFLOWS without any complication.
A1: We prepare an input data preparation/make up/development assistance tool for Windows. In order to prepare input data needed for running GETFLOWS efficiently, the user can use the special interface. Commercial GIS and 3D visualization software are used as postprocessors. A data format conversion tool can convert GETFLOWS’s standard output data into ArcGIS, MapInfo, MicroAVS, and TECPLOT data formats.
A2: The analysis of meteorological fluid behavior such as high-speed fluid movement in the atmosphere is not included in GETFLOWS present version. However, through a linkage to a meso-scale meteorological numerical model, such as MM5, any other scale numerical model, GETFLOWS will be able to carry out atmosphere-geosphere integrated analysis in the future.
A3: Yes. The user can embed your own-developed programs into GETFLOWS and compile it. The individual models which are related to evapotranspiration, snowmelt, adsorption of chemical species, and irrigation/fertilization have already been embedded to GETFLOWS.
A1: The most important point is to make sure that the constructed numerical model has right input data and the calculation is properly conducted by GETFLOWS. The unnatural model and incorrect input data will result a more and more numerically unstable conditions during the simulation and lead to a dilemma when you are going to interpret the results. Especially the “unnatural modeling” is often noticed after starting calculation. It is not easy to check the whole composed input data when the 3D model is large and complicated but it’s important to examine the model by visualizing the results using graphic software, etc. at the initial stage to detect errors.
A2: The unnatural model shows physically strange or incompatible fluid and material flow phenomena. A typical example is in the case of pumping a large quantity of groundwater from an aquifer where the permeability is very low. Such a modeling can be easily distinguished because the time step does not increase gradually and the solver and Newton loop interactive calculations are hard to convergent.
A1: The calculation time varies depending on the size of target model, such as grid number and Degrees of Freedom (DOF), and the External Forces. The followings are rough standards for the calculation times if you use a commercial personal computer (Memory 1-2GB / CPU2GHz).
|Grid Number||DOF||Calculation Time|
|Less than 10,000||Less than 20,000||several mins. to hrs.|
|100,000 to 300,000||200,000 to 600,000||6hrs. to 24hrs.|
|300,000 to 600,000||600,000 to 1,200,000||24hrs. to 48hrs.|
|More than 1,000,000||More than 2,000,000||More than 60hrs.|
By using our parallel computing platform (Linux cluster), the calculation speed can be drastically improved. Please contact us for more details.
Copyright© 2000-2016, Geosphere Environmental Technology Corp. All Rights Reserved.