Skip to content

Latest commit

 

History

History
115 lines (76 loc) · 3.68 KB

time_series.rst

File metadata and controls

115 lines (76 loc) · 3.68 KB
Raw

Time Series (doc in progress)

The documentation of this section is in progress. It is rather incomplete for the moment, and only expose the most basic features.

If you are interested in collaborating to improve this section, let us know.

Goal

This class aims to make faster (and easier) the computations of the current flows (measured in Amps) at a certain side of a powerline / transformer when the topology is not modified.

It can be used as:

from lightsim2grid import TimeSerie
import grid2op
from lightsim2grid.lightSimBackend import LightSimBackend

env_name = ...
env = grid2op.make(env_name, backend=LightSimBackend())

time_series = TimeSerie(env)
Vs = time_series.compute_V(scenario_id=..., seed=...)
As = time_series.compute_A()  # will contain the flows, in amps at each step (rows) for each powerline (column)

For now this relies on grid2op, but we could imagine a version of this class that can read to / from other data sources (for now please use the more basic lightsim2grid.timeSerie.Computers for such purpose)

Importantly, this method is around 13x faster than simulating "do nothing" (or "one change then nothing") with grid2op (see section timeserie_benchmark )

Note

A more detailed example is given in the examples\time_serie.py file from the lightsim2grid package.

Warning

Topology and injections

The topology is taken from the initial provided grid and cannot be changed when evaluating a given "time serie".

Then, the call to time_series.compute_V(scenario_id=..., seed=...) will only read the injections (productions and loads) from grid2op to compute the voltages.

Note

As this class calls a long c++ function, it is possible to use the python Threading module to achieve high efficient parrallelism. An example is provided in the examples\computers_with_grid2op_multithreading.py file.

Benchmarks

Here are some benchmarks made with:

  • system: Linux 5.11.0-40-generic
  • OS: ubuntu 20.04
  • processor: Intel(R) Core(TM) i7-4790K CPU @ 4.00GHz
  • python version: 3.8.10.final.0 (64 bit)
  • numpy version: 1.18.5
  • pandas version: 1.1.4
  • pandapower version: 2.7.0
  • lightsim2grid version: 0.6.0
  • grid2op version: 1.6.4

Where lightsim2grid has been installed from source with all optimization enabled.

This benchmark is available by running, from the root of the lightsim2grid repository:

cd examples
python3 time_serie.py

For this setting the outputs are:

For environment: l2rpn_neurips_2020_track2
Total time spent in "computer" to solve everything: 0.05s (12277 pf / s), 0.08 ms / pf)
    - time to pre process the injections: 0.00s
    - time to perform powerflows: 0.05s (12697 pf / s, 0.08 ms / pf)
In addition, it took 0.00 s to retrieve the current from the complex voltages (in total 11681.3 pf /s, 0.09 ms / pf)

Comparison with raw grid2op timings
It took grid2op (with lightsim2grid): 0.66s to perform the same computation
    This is a 13.4 speed up from TimeSerie over raw grid2op (lightsim2grid)
It took grid2op (with pandapower): 14.94s to perform the same computation
    This is a 302.9 speed up from TimeSerie over raw grid2op (pandapower)
All results match !

In this case then, the TimeSerie module is more than 15 times faster than raw grid2op.

Detailed usage

lightsim2grid.timeSerie

  • genindex
  • modindex
  • search