Pre- & Post- processors - What are their Pros and Cons?
I need to develop reactive-transport models with TOUGHREACT and run simulations as part of a project on "subsurface barrier formation as a CO2 leakage mitigation and remediation technology".
I would like to hear people's opinion about pre- and post- processors used for this type of problems. I am aware about Petrasim, Pytough, Toughvisual, Tecplot RS and Excel but can't find a review of their pros- and cons-.
Thanks in advance.
There are several choices now for post-processors of TOUGHREACT output, and a few options for preprocessing. Preprocessing and post processing choices include:
1) "A workflow for handling heterogeneous 3D models with the TOUGH2 family of codes: Applications to numerical modeling of CO2 geological storage" Audigane et al. (2011) Computers & Geosciences (free from Authors). This uses the free ParaView or the commercial TecPlot software for post processing. Windows (tools work on Mac and Linux, interface maybe). Has tools to extract data/mesh from Petrel.
2) PetraSim (commercial) Windows only.
3) PyTOUGH (free) seems very powerful, but I haven't used it yet.
Post processing mainly:
TecPlot (commercial) (Mac, Windows, Linux?)
VisIt (free) - difficult to use but very powerful for 3-D (Mac, Windows, Linux)
ParaView (free) - similar to VisIt, a little easier to use, although I had more difficulty! (Mac, Windows, Linux)
GMT (free) very good 1-D, 2-D, some 3-D plotting, data manipulation, and several tools which can be used for preprocessing. Requires unix scripting knowledge for best use. Works on Mac, Linux, Windows.
I would also like to hear from other users what they like best, particularly free software which we can be used easily in classes and works on Mac and Windows. New output file types could be introduced into upcoming versions if there is some consensus on what works best on both platforms.
In the past for similar problems I have used GMS (Groundwater Modeling System) from Aquaveo. GMS is not free and has very powerful geologic and conceptual modeling tools for pre- and post-processing. GMS is originally made for simulation using MODFLOW and other commonly used USGS codes for hydrology. I have prepared a "translator" from the MODFLOW grid to a TOUGH2 grid that you can download for free from the TOUGH2 web page. I also have a translator for post processing from the TOUGH2 output to the GMS grid that is useful for interpolating back the data into the MODFLOW grid for visualization and analysis. This code though is still in the process of being completed (I can provide a copy if you like), but it may need some work for the chemical part of TOUGH-REACT.
We have used GMS with TOUGH-REACT for modeling the phosphogypsum dumps in the Venetian Lagoon (a problem similar to yours, I guess) with very good success. Using the same interface for MODFLOW and TOUGH2 allows to get the best from both modeling codes. We use MODFLOW to get the water table with great accuracy, and then import the results into TOUGH2 (you have two options: a. using the bock GENER or - using a "fixed" water table at local constant pressure). This is the reference paper: Borgia A., et al., 2010. Using a MODFLOW grid, generated with GMS, to solve transport problem with TOUGH2 with complex geological environments: the intertidal deposits of the Venetian Lagoon. TOUGH 2009. Computer & Geosciences, doi 10.1016/j.cageo.2010.11.007.
I have also used Leapfrog (very good for modeling intrusions), mView (very powerful but hard to use) and Petrasim (simpler to use but some times not so stable) with success - not the latest versions though...
I hope this helps.
just two short comments: I personally use TecPlot for Postprocessing and it works quite well. Maybe not best, but finally you can modify any parameter or visualisation you want. But it is commercial and rather expensive.
I also used PetraSim and found it generally nice to use, but they still do not include TOUGHREACT V3 simulations, especially concerning CO2 phases and trace gases. Thus I am now only working with the "original" TOUGHREACT executables.
I also use MATLAB for automatic generation of GENER input files, but that is a different story. Just ask if you want to know more about this.
Best regards, Lennard
I noticed that you recommended Tecplot as a post-processor of TOUGHREACT, which can visualize the results quite well. And Recently I also wanna to try to use Tecplot as the Post-processor of TOUGH2, but I found that normally the cell-centered data sets are generated by TOUGH2, which will be automatically considered as the nodal data for the interpolation, causing the wrong contour. So is there any method to generate the result of nodal data in TOUGH2, or should I have to give more information when using Tecplot.
Thanks in advance.
I am not sure if I have really understood your problem, but I can explain how I proceed to produce nice, coloured pictures :-).
Firstly, I assume you are talking about 2D data, right? 1D data can be simply used as it is for xy line plots.
To visualise 2D data, you need to convert the original data, which is interpreted by Tecplot as 1D data (only "I", no "J"), by interpolating it with an empty rectangular grid (as you wrote). So...
- Load your TOUGH output file data.
- Create a Rectangular zone (data menu). Use sufficient I and J values to represent the original resolution of data points of your TOUGH grid.
- If your original TOUGH x (or y) coordinates are not equally distributed, but e.g. exponentially increasing (for instance in a 2D radial grid for CO2 storage simulation), than recompute the x values of the rectangular grid by using the same equation as in TOUGH.
- Now the coordinates of your rectangular grid should exactly match the original TOUGH coordinates. This is important. In this case, you can start the interpolation by using nearest n and only 1 data point. This basically means, that the TOUGH values are simply copied into the rectangular zone.
Now it should work to draw contours using the rectangular zone. I don't know if it is the best way for dealing with the data conversion, but it works.
Best regards, Lennard
SO many thanks for your detailed explanation of using TECPLOT to do post-process, but i am still afraid that i did not make myself clear. And actually my issue is like that the result of elements obtained from TOUGH2 is all cell-centered, which means that i can get the coordinates and attributes of center of each element. And what TECPLOT actually needs for plotting is nodal values, but my solutions of TOUGH+ that are stored at the cell centers of the mesh. I would like to import the solution into Tecplot and have it interpolate/extrapolate the solution to the nodes. What is the best way to go about this?
Thanks in advance!
Well, that should be similar to what I wrote:
- Import the data
- Create a rectangular zone
- recompute the x/y values of the rectangular zone (data/ alter/ specify equations) if necessary
- interpolate, but in case of transferring cell centered data to nodal data you should better use e.g. the octant method
You could send me some of your data if it is not too large, then I could try to understand what you need? Send me an email to JanLennard.Wolf (at) bgr.de if you like.