One of the tasks of Non-Intrusive Load Monitoring (NILM) is load identification, which aims to extract and classify altered electrical signals after switching events are detected. In this subtask, representative and distinguishable load signatures are essential. At present, the literature approach to characterize electrical appliances is mainly based on manual feature engineering. However, the performance of signatures obtained by this way is limited. In this paper, we propose a novel load signature construction method utilizing deep learning techniques. Specifically, three learnable load signatures are presented such as Learnable Recurrent Graph (LRG), Learnable Gramian Matrix (LGM) and Generative Graph (GG). Furthermore, we test different frameworks for learning these signatures and conclude that Temporal Convolutional Networks (TCN) based on residual learning are more suitable for this work than the other schemes mentioned. The results of experiment on the PLAID datasets with submetered and aggregated, WHITED dataset and LILAC dataset confirm that our method outperforms the voltage-current trajectory, Recursive Graph and Gramian Angular Field methods in multiple evaluation metrics.
- numpy==1.19.5
- scikit_learn==1.2.2
- torch==1.6.0+cu101
- tqdm==4.63.1
The format of the data folder is as follows:
data
|-lilac
|-aggregated
|-current.npy
|-labels.npy
|-voltage.npy
|-plaid2018
|-aggregated
|-current.npy
|-labels.npy
|-voltage.npy
|-sub
|-plaid2018_sub.pickle
|-whited
|-whited.pickle
The .pickle file contains the current, voltage, label, _ in the PLAID2018 and WHITED data set
You can train a new model in plaid2018 dataset with the following command:
python main.py --dataset_name plaid2018
You can train a new model in submetered plaid2018 dataset with the following command:
python main.py --dataset_name plaid2018 --sub
If you find this project useful in your research, please consider cite:
@ARTICLE{10025855,
author={Zhang, Yusen and Wu, Hao and Ma, Qing and Yang, Qingrong and Wang, Yiwen},
journal={IEEE Transactions on Smart Grid},
title={A Learnable Image-Based Load Signature Construction Approach in NILM for Appliances Identification},
year={2023},
volume={},
number={},
pages={1-1},
doi={10.1109/TSG.2023.3239598}}