NNILM - Neural Non-Intrusive Load Monitoring
A comparison of the rectangles architecture in this repository with the original implementation by Jack Kelly on the UK-DALE dataset.
Predictions for a true positive, a true negative and a false positive example for a dish washer.
Visualization of first convolution layer
The following image shows the 16 learned filters of the first convolution layer.
Activations for a negative sample
The following image shows the activations of the first convolution layer given a negative sample.
Activations for a positive sample
The following image shows the activations of the first convolution layer given a positive sample. The two highlighted filters are thereby interesting. The first filter seems to have learned to detect step changes in the input signal. The second filter seems to have learned to pass the raw input signal to the next layer.
We used Python 2 as programming language, because there are some dependencies that are not compatible with Python 3.
pip install virtualenv
python -m virtualenv env
If you set up virtualenv with Python 3, you have to switch to Python 2:
virtualenv --python=/usr/bin/python2.7 env
Because of a failing dependency
hmmlearn, we use the option
--no-dependencies. The module
not really needed in this implementation.
pip install --no-dependencies -r requirements.txt
pip install -e .
It is possible to use any dataset supported by the NILMToolkit.
You have to convert the dataset to a .h5 file using the converters provided by the NILMToolkit and then place it into the folder
Run training for 30 epochs:
python nnilm/train.py dish_washer_redd -s0 -e30
Resume training from epoch 30:
python nnilm/train.py dish_washer_redd -s30 -e40
Run training on GPU cluster:
nohup python nnilm/experiments/train_hpc.py --gpu=2 &
python predict.py dish_washer_eco
python nnilm/experiments/csv_predictor.py -ddish_washer_redd -idata/aggregated_power.csv -s01-04-2018 -e08-04-2018
nnilm/experiments contains files that have been used to run my experiments.
It shows how the scripts from this repository can be used in another project.