pysimdeum 1.0.1

A stochastic water demand end-use model in Python

PYSIMDEUM

pysimdeum is a Python package for modelling and simulating residential stochastic water demand at the end-use level.

Main functionalities:

• Build and populate houses with users and water end-use devices according to country specific statistics
• Simulate water usage stochastically based on the statistics
• The results are stored as xarray.DataArray, so all the simulation information can be accessed and aggregated afterwards (e.g., specific end-uses, sums over water usage of users, rolling means over time, ...)
• Serialization: pysimdeum supports different output formats (e.g., csv, excel, netcdf, ...)
• Plotting results using matplotlib

A detailed documentation will be soon available under https://pysimdeum.readthedocs.io.

Output is based on statistics about household sizes and water use of the Netherlands. These can be changed either within the code or by creating the correct toml files. An overview of worldwide differences is available in:

Mazzoni, F., Alvisi, S., Blokker, E. J. M., Buchberger, S. G., Castelletti, A., Cominola, A., Gross, M. P., Jacobs, H. E., Mayer, P., Steffelbauer, D. B., Stewart, R. A., Stillwell, A. S., Tzatchkov, V., Yamanaka, V. H. A. and Franchini, M. (2022). "Investigating the characteristics of residential end uses of water: a worldwide review." Water Research, art. no. 119500, doi:https://doi.org/10.1016/j.watres.2022.119500.https://www.sciencedirect.com/science/article/pii/S0043135422014452 https://livelink.kwrwater.nl/livelink/livelink.exe/Open/69638292

---

Warning!

Be warned, that pysimdeum is still changing a lot. Until it's marked as 1.0.0, you should assume that it is unstable and act accordingly. We are trying to avoid breaking changes but they can and will occur!

---

Installation

pysimdeum uses features only available in a newer Python version, which is why Python >= 3.8 is needed along with several Python package dependencies.

pysimdeum is available on PyPI and can be easily installed together with its dependencies using pip:

pip install pysimdeum

Alternatively, you can install pysimdeum from its repository:

pip install git+https://github.com/KWR-Water/pysimdeum.git

Dependencies

pysimdeum requires the following Python packages:

• matplotlib
• numpy
• pandas
• toml
• xarray
• scipy

Basic Usage

To use pysimdeum , you first have to import it in your script:

import pysimdeum

In pysimdeum , everything is about the House. If you want to start with a new, empty House, type the following:

house = pySIMDEUM.built_house(house_type='one_person')

If you want to build a specific House, e.g., a one-person household, you can use the house_type keyword:

# Built a house (one-person household)
house = pySIMDEUM.built_house(house_type='one_person')

The house is automatically populated by people, which follow certain statistics, and "furnished" with water end-use devices or appliances (e.g., toilet, bathtub, ...). You can check, which appliances are available by using the appliances or users property of the House:

# Show users and water end-use devices present in the house
print(house.users)
print(house.appliances)

To simulate the water consumption of a house, you can use the House`s simulate method:

# Simulate water consumption for house (xarray.DataArray)
consumption = house.simulate(num_patterns=100)

The simulation result is an xarray.DataArray --- basically a labelled numpy.ndarray with four dimensions / axes (i.e., time, user, enduse, patterns).

You can easily create statistics over the consumption object, for example, to compute the average total consumption (sum of consumption of all users and enduses as an average over the patterns), you can build the sum over the user and enduse axes (the total consumption), and then build the mean over the patterns axes. There are two flowtypes defined. totalflow and hotflow. totalflow reflects the total water use while hotflow reflects the water that has been heated up.

# Build statistics from consumption
tot_cons = consumption.sum(['enduse', 'user']).sel(flowtypes='totalflow').mean([ 'patterns'])

If you want to plot the results pand additionally depict some rolling averages (e.g., hourly means = 3600 seconds), you can this in the following way

# Plot total consumption
tot_cons.plot()
tot_cons.rolling(time=3600, center=True).mean().plot()
plt.show()

License

pysimdeum is available under a EUPL-1.2 license.

Contributing

If you want to contribute, please check out our Code of Conduct and our Contribution Guide. Looking forward to your pull request or issue!

Citing

If you publish work based on pysimdeum , we appreciate a citation of the following reference:

• Steffelbauer, D.B., Hillebrand B., Blokker, E.J.M., 2022. pySIMDEUM: An open-source stochastic water demand end-use model in Python. Proceedings of the 2nd joint Water Distribution System Analysis and Computing and Control in the Water Industry (WDSA/CCWI2022) conference, Valencia (Spain), 18-22 July 2022.