Sambaiga

selected publications

1. DeepNLMTK

   Unlocking the Full Potential of Neural NILM: On Automation, Hyperparameters and Modular Pipelines

   IEEE Transactions on Industrial Informatics 2022

   Abs PDF Code

   Non-Intrusive Load Monitoring (nilm) techniques are increasingly becoming a key instrument for identifying the power consumption of individual appliances based on a single metering point. Particularly Deep learning models are gaining interest in this regard. However, the challenges brought by NILM datasets and the non-availability of common experimental guidelines tend to compromise comparison, research transparency and replicability. The limited adoption of efficient research instruments and a lack of best practices guidelines contribute in huge part to this problem, where no features, encouraging standardised formats for benchmarking and results sharing, are offered. To address these issues, we first present a brief overview of recent best practices for Deep Learning (DL) and highlight how deep NILM research can benefit from these practices. Furthermore, we suggest a novel open-source toolkit leveraging these practices: Deep-NILMTK. The proposed toolkit offers a common testing bed for NILM algorithms independently of the underlying deep learning framework with a modular NILM pipeline that can easily be customised. Furthermore, Deep-NILMTK introduces the concept of Experiment Templating to offer pre-designed experiments allowing to enhance research efficiency

2. FPSeq2Q

   FPSeq2Q: Fully Parameterized Sequence to Quantile Regression for Net-Load Forecasting with Uncertainty Estimates

   Faustine, Anthony, and Pereira, Lucas

   IEEE Transactions on Smart Grid 2022

   Abs PDF Code

   The increased penetration of Renewable Energy Sources (RES) as part of a decentralized and distributed power system makes net-load forecasting a critical component in the planning and operation of power systems. However, compared to the transmission level, producing accurate short-term netload forecasts at the distribution level is complex due to the small number of consumers. Moreover, owing to the stochastic nature of RES, it is necessary to quantify the uncertainty of the forecasted net-load at any given time, which is critical for the real-world decision process. This work presents parameterized deep quantile regression for short-term probabilistic net-load forecasting at the distribution level. To be precise, we use a Deep Neural Network (DNN) to learn both the quantile fractions and quantile values of the quantile function. Furthermore, we propose a scoring metric that reflects the trade-off between predictive uncertainty performance and forecast accuracy. We evaluate the proposed techniques on historical real-world data from a lowvoltage distribution substation and further assess its robustness when applied in real-time. The experiment’s outcomes show that the resulting forecasts from our approach are well-calibrated and provide a desirable trade-off between forecasting accuracies and predictive uncertainty performance that are very robust even when applied in real-time.

3. AWRGNILM

   Adaptive Weighted Recurrence Graphs for Appliance Recognition in Non-Intrusive Load Monitoring

   Faustine, Anthony, Pereira, Lucas, and Klemenjak, Christoph

   IEEE Transactions on Smart Grid 2021

   Abs PDF Code

   To this day, hyperparameter tuning remains a cumbersome task in Non-Intrusive Load Monitoring (NILM) research, as researchers and practitioners are forced to invest a considerable amount of time in this task. This paper proposes adaptive weighted recurrence graph blocks (AWRG) for appliance feature representation in event-based NILM. An AWRG block can be combined with traditional deep neural network architectures such as Convolutional Neural Networks for appliance recognition. Our approach transforms one cycle per activation current into an weighted recurrence graph and treats the associated hyper-parameters as learn-able parameters. We evaluate our technique on two energy datasets, the industrial dataset LILACD and the residential PLAID dataset. The outcome of our experiments shows that transforming current waveforms into weighted recurrence graphs provides a better feature representation and thus, improved classification results. It is concluded that our approach can guarantee uniqueness of appliance features, leading to enhanced generalisation abilities when compared to the widely researched V-I image features. Furthermore, we show that the initialisation parameters of the AWRG’s have a significant impact on the performance and training convergence.