Simon Waldman PhD Researcher

[quote author=Jonas Althage link=topic=1954.msg4434#msg4434 date=1392904975]
Hi,
If we assume that you have some additional software at disposal, I've got a few tips.

If your scatter data set is still too big for your model to handle, you can use either MatLab or e.g. ArcGIS (if you have access to it) to add elevation units to your mesh.

[/quote]

Sorry to be slow, I didn't notice the reply - but a great answer, so thank you.

Hi Will,

I've been here - it's possible but non-obvious! I assume that you want to export a vertical plane - a 2D vertical transect through the water column. Note that these instructions refer to MIKE 2012 - I don't know what may have changed in the 2014 edition.

First you must get your model to output a 3D .dfsu file, a "Volume" file. There is no way to get directly to a vertical plane.
Having done this, create a two-line .xyz file with the start and end coordinates of the transect that you want. It doesn't matter what the z-values are, they will be ignored, but they need to be there. I usually just set them to zero.

Open your .dfsu in Data Manager, and go to Tools -> Extract. Under "Dimension" select "Profile". Under "Profile" select "Load from .xyz file".

The rest is obvious, except for "Nearest segment list". If you leave this unchecked, then MIKE will interpolate to estimate the correct values on your line. If you check it, then as I understand it instead of doing this it will simply use the values for each mesh element that the line passes through, even though the centres of those elements may not lie on the line.. This will run much faster, but may be less accurate.

Hope that helps, and that it's still relevant four months later (and anyway, it may apply to others).

I'm trying to import some high-resolution bathymetry in MIKE 2012 and am getting "Out of memory" errors, whether I use xyz or dfs2 files. My PC is certainly not actually out of memory, so I'm guessing that MIKE is hitting a maximum number of scatter data points or some such...

I realise that the 64-bit builds in MIKE 2014 may solve this problem, but I don't have access to that.

So... to save me a lot of trial and error, does anybody know what the maximum number of scatter data points that the 2012 Mesh Generator can handle is? Is this number per file, or total, or...?

Thanks.

Never mind, I just found the right-click option "inside polygon". In typical MIKE fashion it's there, just well hidden ;-)

I have an existing flexible mesh which I have spent some time refining by hand beyond what the Mesh Generator originally produced.

Due to a new area of interest I would like to retriangulate one part of this mesh (at a higher resolution), but because of the time spent on refinement I would prefer not to retriangulate *all* of it.

I've noticed the "re-triangulate region" tool in the mesh generator, but this seems to only apply to a very small area (I think the elements that touch a specific node). Is there any way to re-triangulate a larger area, but not the whole domain?

Thanks.

I'm putting this here because it applies to all, or at least a number of, modules.

I'm doing some lengthy simulations in MIKE3 HD FM, and we sometimes suffer from power cuts (and don't currently have a UPS available). These power cuts sometimes cost me days of runtime, as I have to start the simulation again.

One way to deal with this might be to save a full volume output of the entire domain, as it should be possible to use this to provide initial conditions to continue the run. However, this would be hundreds of gigabytes.

I am told that some other modelling software has the facility to create "resume" files that contain the full state of the model at just the last timestep - and so are overwritten on each time step so that only one time step is stored. This could make it simple to resume after an interruption.

Does MIKE have this facility in some hidden corner? If not, perhaps it should be a feature request!