A library for optimizing energy systems using mixed integer linear programming.
Currently the library has two models:
- electric battery operating in price arbitrage,
- a combined heat & power plant.
The battery model is well developed and tested - the CHP model is working and requires more development.
The battery model is optimized against a set of prices, and returns a list of dictionaries - one per interval:
import energypylinear as epl model = epl.Battery(power=2, capacity=4, efficiency=1.0) prices = [10, 50, 10, 50, 10] results = model.optimize(prices, freq="60T")
pandas can be used to transform the results into a dataframe:
import pandas as pd pd.DataFrame().from_dict(results) Import [MW] Export [MW] Power [MW] Charge [MWh] 0 2.0 0.0 2.0 0.000000 1 0.0 2.0 -2.0 0.066667 2 2.0 0.0 2.0 0.000000 3 0.0 2.0 -2.0 0.066667 4 NaN NaN NaN 0.000000
The last row is all
NaN except for
Charge indicates the battery position at the start of each interval. The last row is included so we can see the battery level at the end of the optimization run.
It is also possible to send in forecast prices along with actual prices, and to change the initial charge.
If the model receives forecasts it will optimize for them - this allows measurement of forecast quality by comparing actual with forecast costs:
# a forecast that is the inverse of the prices we used above forecasts = [50, 10, 50, 10, 50] results = model.optimize(prices, forecasts=forecasts, timestep='60T', objective='forecasts')
The battery model also accounts for carbon. If no carbon profile is passed in, a constant value of 0.5 tC/MWh is assumed.
We can switch the optimization to focus on carbon:
results = model.optimize(prices, forecasts=forecasts, carbon=carbon, timestep="60T", objective='carbon')
Install as an editable package:
$ make setup
The main dependency of this project is PuLP. For further reading on PuLP:
- the white paper - An Introduction to pulp for Python Programmers,
- the blog post series Introduction to Linear Programming with Python and PuLP - especially Part 6 which covers how to formulate more complex conditional constraints.
$ make test