Jarrah Muller Civil Engineer

Hi,
I have an IFM plug-in that looks at heads and pressures and calculates a result when manually activated.

If I load the plug-in before running the model, it is also attached to the DAC and I can activate it to do my my calculation.

However, if I don't attach it to the FEM before running the model, I can't use it. When viewing the DAC file I can't attach new plug-ins.

This plug-in only uses the callback OnActivate, and so shouldn't need to be attached during the model run. I would like to be able to use my plug-in to process old model files without re-running the models. How can I attach my plug-in to existing DAC files?

Can you give an example of how the membership interval operators are supposed to work in an equation?

i.e. E[,] and E[,) etc

I'm sure they'll simplify my life but I don't know the syntax to use them.

I can't see any reference to the parameter recording in the API index. Is it documented?

Can an IFM find out whether a node is in a Node Selection group?

I'm hoping to use similar call backs to the RefDist callbacks to get Node Selection names and then run processes on nodes that are in certain selections.

Sounds awkward.
Is there an easy way to plot *.dar files outside of FEFLOW, or export them in a plottable way?

Is there a way to start the PEST server with Windows?

Maybe a batchfile command that opens FePEST and starts the server?

If not, could you make a PEST server Service that runs with Windows in the background without needing user interaction? I'd like to know that certain modelling computers in my organisation were always ready to accept PEST runs.

You can make the model have a 3D unstructured mesh, so that lower layers have different node layouts to upper layers. You need to use Tetgen to do that, as you mentioned.

There are some caveats:
- you lose the ability to look at results by slice afterwards
- you can't use the free movable surface option for modelling the water table, but need to use unsaturated flow equations which generally are much slower, need more data to define the soil, and make models not run. Unless you have a confined aquifer.

When I've had this sort of problem, I've not gone down the unstructured mesh route, but tried to make the slice model work by compromising on the number of nodes. Going down to 130 m is fine if your layers are thick, for example. Generally models with up to about 300,000 nodes run ok on modern PCs. Mostly the bad effects of too many nodes are limited to long run times.

Or maybe you can run a 2D slice model with the discretisation that you want. They tend to run very quickly.

Attachment shows how to refer to monthly arrays.

You can link the data in the shapefile to a user defined element distribution, but only one time step at a time as user data distributions can't be time varying for some reason. So, you could have one element distribution for January, one for Feb etc. Only practical if you have cyclic yearly data.

Then you need to make the in/out flow on top/bottom equation calculate the recharge for each time step with reference to the user data. For example, you could enter a series of IF equations that check model time, and if in certain ranges read the data from the correct elemental array.

For a model longer than 1 year, I'd set up an intermediate POW file with day in column 1 and the month of the year in column 2. Then in the in/out on top/bottom equation refer to the POW data to look up which monthly array to use as an index, and then use IF statements to pick the actual array to be used.

Having said all that, I do areally distributed recharge differently. I have a POW file with daily rain, a user elemental reference distribution for landuse that contains a multiplication factor, and another user elemental reference distribution for soil that contains a second multiplication factor, and multiply the three in the afore mentioned equation.

Has this functionality been added yet? I need to add wells to an existing FEFLOW 7 model.