locie / energy_plus_wrapper

Python lib running energy plus across platform and version
MIT License
16 stars 6 forks source link

energy+ wrapper

This little library has been written in order to run energy+ simulation in linux and windows in a thread-safe way.

The main goal is to ensure a stable behavior across platform and version, and to make the link between the e+ building model tools written in python and the different analysis and optimization tools.

Install

For now, the package is available on PyPI, and via the github repo.

pip install git+https://github.com/locie/energy_plus_wrapper

for the requirements.

Usage

There is two main classes in the library: the EPlusRunner and the Simulation.

The first one is linked with the EnergyPlus local installation and allow to run one or many simulation. It also ensure the compatibility between the executable and the IDF file.

The later is generated with the runner.run_one and runner.run_many. It contains the EnergyPlus simulation reports and results.

A small utility routine ensure_eplus_root allow to automatically download, extract and install EnergyPlus (but only for linux for now).

from energyplus_wrapper import EPlusRunner, ensure_eplus_root

eplus_root = ensure_eplus_root("http://url/to/eplus/install.sh")
# or
eplus_root = "path/to/energyplus/install/dir"

runner = EPlusRunner(eplus_root)
simulation = runner.run("my/idf/file.idf", "my/weather/file.epw")
print(simulation.reports) # be aware that the result is a complex nested dict
print(simulation.time_series)

The runner.run is multi-process safe, allowing to run multiple EnergyPlus simulation at the same time.

The method runner.run_many allow to run multiple simulation at the same time. It uses joblib.parallel under the hood : how the simulation are ran and how many CPU are used can be controlled by joblib.parallel_backend

from energyplus_wrapper import EPlusRunner, ensure_eplus_root

eplus_root = ensure_eplus_root("http://url/to/eplus/install.sh")
# or
eplus_root = "path/to/energyplus/install/dir"

runner = EPlusRunner(eplus_root)

samples = {"sim01": ("idf01.idf", "weather01.epw"),
           "sim02": ("idf02.idf", "weather01.epw"),
           "sim03": ("idf03.idf", "weather02.epw"),
           "sim04": ("idf04.idf", "weather02.epw")}

# run on 4 CPU
with joblib.parallel_backend("loky", n_jobs=4):
    sims = runner.run_many(samples)

print(sims.keys())

run_one, run_many common mecanism

eppy compatibility

Every time you need to provide an idf file, you can use a eppy.modeleditor.IDF object instead (as obtained using eppy.openidf). This allow dynamic change of the model without the need to rewrite one file per simulation.

backup

According to the backup_strategy, the runner can save the files generated by EnergyPlus in the backup folder. It can save the files 'always', 'on_error', or never (with None).

version check

According to version_mismatch_action, a mismatch between the e+ executable and the idf file version will raise an error, warn the user or be ignored.

custom post-process

You can provide a custom simulation post process. By default, the file eplus-table.htm is parsed and put in the simulation.reports attribute, and all the csv files generated by the run are parsed and put in simulation.time_series.

For example, you could save all the files generated by the simulation with

from path import Path
import zipfile

archive_folder = Path("./archive/").abspath()

def save_and_compress(simulation):
    with zipfile.ZipFile(archive_folder / f"{simulation.name}.gz", "w", zipfile.ZIP_DEFLATED) as zip_handler:
        for file in simulation.working_dir.files("*"):
            zip_handler.write(file)

runner.run_one(idf, wf, custom_process=save_and_compress)