Non-Intrusive Load Monitoring (NILM) aims to predict the status or consumption of domestic appliances in a household only by knowing the aggregated power load. NILM can be formulated as regression problem or most often as a classification problem. Most datasets gathered by smart meters allow to define naturally a regression problem, but the corresponding classification problem is a derived one, since it requires a conversion from the power signal to the status of each device by a thresholding method. We treat three different thresholding methods to perform this task, discussing their differences on various devices from the UK-DALE dataset. We analyze the performance of deep learning state-of-the-art architectures on both the regression and classification problems, introducing criteria to select the most convenient thresholding method.
Source: see publications
-
Install miniconda
curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh | bashSay yes to everything and accept default locations. Refresh bash shell with
bash -l -
Update conda
conda update -n base -c defaults conda -
Clone this repository and cd into the folder
-
Create and activate conda environment (removing previously existing env of the same name)
conda remove --name better-nilm --all conda env create -f environment.yml --force conda activate better-nilm
UK-DALE dataset is hosted on the following link: https://data.ukedc.rl.ac.uk/browse/edc/efficiency/residential /EnergyConsumption/Domestic/UK-DALE-2017/UK-DALE-FULL-disaggregated
The files needed by this module are ukdale.zip and ukdale.h5.zip. Download both and unzip them in a folder named data inside the root. Once you are done, your local directory should look like this:
better_nilm
|_ better_nilm
|_ [python scripts and subfolders]
|_ data
|_ ukdale
|_ [house_1 to house_5]
|_ ukdale.h5
Credit: Jack Kelly
The folder better_nilm contains an executable script to train the models a plot their results. Run the following line on the root folder (make sure to have the enviroment active and the data downloaded):
python better_nilm/train_test_model.py
This will train and score the CONV model using the default parameters. The script stores several outputs in the outputs folder, including:
- .txt files with the model scores over the test data
- .png files showing samples of the model's prediction.
- .png files with the scores against the classification weight.
The list with all the available parameters and their default values is stored in the configuration file.
If you want to use your own set of parameters, duplicate the aforementioned configuration file and modify the paremeters you want to change (without deleting any parameter). You can then use that config file with the following command:
python better_nilm/train_model.py --path_config <path to your config file>
For more information about the script, run:
python better_nilm/train_model.py --help
Once the models are trained, test them with:
python better_nilm/test_model.py --path_config <path to your config file>
To reproduce the results shown in our paper, activate the environment and then run:
nohup sh train_test_sequential.sh > log.out &
This will first create a folder named configs, where the different configurations of the
models are stored. Then, the script train_test_models.py will be called, using each
configuration each. The outputs will be stored in another created folder named outputs.
There are three threshold methods available. Read our paper to understand how each threshold works.
- 'mp', Middle-Point
- 'vs', Variance-Sensitive
- 'at', Activation Time
NILM as a regression versus classification problem: the importance of thresholding
Author: Daniel Precioso, PhD student at Universidad de Cádiz