Evaluating drawdown at array of coordinates

[robinkeegan]

I have a model that is surrounded by a leaky line string that acts as a semi impermeable model extent. I want to compute drawdown within this extent. Using headgrid is inefficient as due to the geometry more than 50 % of the grid points are outside of the extent, meaning we are computing head at a lot of unnecessary points. Is there a way to pass a mask or a set of coordinates to headgrid? Currently, I am evaluating each point using head, which works fine but is quite slow.

[mbakker7]

We are currently working on this, see #139. This function will allow you to compute heads on an arbitrary array, using all the cores on your machine. One of the applications we had in mind is to evaluate the solution on an irregular grid for more efficient contouring (more detail near elements, for example). Your use case will work in the same fashion. We are also working on (hopefully massively) improving evaluation of the analytic element solution on a grid for contouring. See #140. Feel free to check out those PRs and if you have any suggestions on those, let us know!

[robinkeegan]

Thanks I will check those out

[dbrakenhoff]

The upcoming release (probably tomorrow) should allow you to use ml.head_array(x, y, parallel=True) for unstructured and much faster head array computations!

[mbakker7]

Just to clarify, ml.head_array takes a 1D array as input for x and y, so you have to convert your grid to a 1D array (or apply your mask) yourself (that should be relatively) easy, and then you can use the new function.