Rule Based Control

This example displays how to use rule-based control (RBC) to control a simple microgrid.

In rule-based control, modules are deployed in a preset order. You can either define this order by passing a priority list or the order will be defined automatically from the module with the lowest marginal cost to the highest.

Setting up the algorithm

Setting up a rule-based control algorithm in straightforward. Simply define your microgrid and pass it to the pymgrid.algos.RuleBasedControl class.

import pandas as pd

from matplotlib import pyplot as plt

from pymgrid import Microgrid
from pymgrid.algos import RuleBasedControl

microgrid = Microgrid.from_scenario(microgrid_number=0)
rbc = RuleBasedControl(microgrid)

Running the algorithm is straightforward:

rbc.reset()
rbc_result = rbc.run()

Investigating the results

At this point, all the results are stored in the DataFrame rbc_result. We can investigate the costs of running the microgrid, the usages of the various modules, and so on.

Most of the cost of running the microgrid is from grid usage:

rbc_result.loc[:, pd.IndexSlice[:, :, 'reward']].cumsum().plot()
plt.show()

As we would hope, there are no excess costs due to overgeneration or loss load:

print(f"Total overgeneration or loss load costs over the course of the year:\n\
    {rbc_result.loc[:, pd.IndexSlice['unbalanced_energy', :, 'reward']].sum().item()}")

Total overgeneration or loss load costs over the course of the year:
    -3.0811264650765224e-10

days_in_month = [
    ('January', 31),
    ('February', 28),
    ('March', 31),
    ('April', 30),
    ('May', 31),
    ('June', 30),
    ('July', 31),
    ('August', 31),
    ('September', 30),
    ('October', 31),
    ('November', 30),
    ('December', 31)
]

month_start_end_dates = {days_in_month[0][0]: [0, 24 * days_in_month[0][1]]}

for month_n, (month, days_in) in enumerate(days_in_month[1:], start=1):
    last_end = month_start_end_dates[days_in_month[month_n-1][0]][-1]
    month_start_end_dates[month] = [last_end, 24 * days_in + last_end]

load_less_renewable_available = rbc_result[('load', 0, 'load_met')] - rbc_result[('pv', 0, 'renewable_current')]
grid_import = rbc_result[('grid', 0, 'grid_import')]

for month, (start_hour, end_hour) in month_start_end_dates.items():

    pd.concat([load_less_renewable_available, grid_import],
              keys=['Net load', 'Grid import'],
              axis=1).iloc[start_hour:end_hour].plot(alpha=0.5, title=f'Net load vs grid import in {month}')

    plt.show()