OpenDC EEMM¶

OpenDC Extension for Energy Modelling & Management

Document: https://opendc-eemm.rtfd.io.

Get Started¶

First, you need to download market data in CSV format from the following offical websites:

Day-ahead market data: ENTSO-E
Imbance market data: TenT

Note that data from ENTSO-E is in CET, whilst data from TenT is in GMT. Please make sure that all market data you selected are of the same period. The ./example/market/ directory contains two such sample datasets.

Next, to run the example, you also need the simulation results produced by the OpenDC datacenter simulator. A sample parque file can be found here.

Installation¶

Please follow the instructions presented here.

Usage¶

Top-level commands¶

usage: opendc-eemm [-v] [-h] -t path [--pue float] {trace,market,decision} ...

CLI of OpenDC Extension for Energy Modelling & Managament.

optional arguments:
  -v, --version         Show version number of the package and exit.
  -h, --help            Show the help messages and exit.
  -t path, --trace path
                        Path to simulation results (expecting a Parque file).
  --pue float           PUE value of the simulatied datacenter.

subcommands:
  Available commands.

  {trace,market,decision}
    trace               Visualize simulation results.
    market              Compare costs in different markets.
    decision            Optimize fine-grained decision-making.

Visualize simulation results¶

usage: opendc-eemm trace [-h] -s ['power', 'oc'] [-f float] [-g value]

optional arguments:
  -s ['power', 'oc'], --show ['power', 'oc']
                        Choose 'power' to show power draw; choose 'oc' to show over-commissioned.
  -f float, --frequency float
                        Frequency of simulated machines.
  -g value, --governor value
                        Governor to visualize.

Analyze energy markets¶

usage: opendc-eemm market [-h] -s ['load', 'strategy'] -o float -d path -i path

optional arguments:
  -s ['load', 'strategy'], --show ['load', 'strategy']
  -o float, --od_price float
                        On-demand energy price.
  -d path, --dayahead_prices path
                        Path to day-ahead energy prices (expecting a CSV file).
  -i path, --imbalance_prices path
                        Path to imbalance energy prices (expecting a CSV file).

Invoke DVFS scheduler¶

usage: opendc-eemm decision [-h] -o ['score', 'schedule'] [-f float] -d path -i path -p path -a ['first', 'last', 'mean'] [-s path]

optional arguments:
  -o ['score', 'schedule'], --option ['score', 'schedule']
                        Choose 'score' to compute the agreement accuracy (AA) sore of the predictions; choose 'schedule' for DVFS
                        scheduling.
  -f float, --factor float
                        Damping factor of the DVFS scheduler.
  -d path, --dayahead_prices path
                        Path to day-ahead energy prices (expecting a CSV file).
  -i path, --imbalance_prices path
                        Path to imbalance energy prices (expecting a CSV file).
  -p path, --predictions path
                        Machine learning predictions (expecting a CSV file).
  -a ['first', 'last', 'mean'], --aggregate ['first', 'last', 'mean']
                        Aggregation method for machine learning predictions.
  -s path, --save_to path
                        Destination path of the DVFS schedule.