Nutrient concentrations

[val-ismaili]

This PR makes an addition for calculating nutrient concentrations of discharge flows. This first pass makes some simplistic assumptions that:

• All enduses have the same nutrient types (though different values)
• Nutrient concentrations per enduse in ww_nutrients.toml are given in g/l

Firstly, a small addition is made that categorises the wastewater of each enduse into blackwater or greywater discharge types. This allows for summarising across discharge types such as discharge.sum(["enduse","user"]).sel(dischargetypes="greywater").plot()

Nutrient concentrations

The main section of this PR is an addition to calculate nutrient concentrations from the discharge flows of each enduse. You can think of this as post-processing analysis applied after the simulation stage. As this occurs after the simulation, I thought it would be better to not edit the discharge object directly. Instead with process_discharge_nutrients I've made use of the xarray.Dataset which allows you to store a xarray.DataArray (the discharge object) in it as well as a number of other variables. dataset_to_df converts these Dataset variables back to a pd.DataFrame

I've included a notebook pysim-nutrients.ipynb showing how to run this for you to check more easily. will tidy before merging.

Note that values in ww_nutrients.toml are just dummy values for now.

[BramHillebrand]

I have added Mirjam Blokker as a reviewer since she has more experience with the nutrients part of simdeum

[Mirjam-blokker]

for the Wc, I can imagine you would have different values for a full flush (faeces) and half flush (urine). For the dishwasher and washing machine you may also want to discern between first discharge (dirt and grease), second discharge (soap residu added) and 3rd + 4th discharge (mainly water). This may be too much detail for now, but would be nice to at least have a simple solution in a later version. We may want to discuss this next week

[val-ismaili]

This makes sense. Wc we think we've got the data to do that quite easily. For WashingMachine and Dishwasher we're looking around. I've set up an issue about this which we'll address next week so we can get this PR closed. #76

[Mirjam-blokker]

I like the greywater and blackwater type. In the future this may be extended to light grey and dark grey, blue, green and yellow water :-)

[sophieoneill]

The hard coded 'NL' will need an update with the roll out of PR 73, as well as a new config file

[val-ismaili]

good point, we'll address that in your PR when pulling from master #77

[sophieoneill]

This is a useful addition!