Presenting at PowerTech 2021
Published:
This week, I had the pleasure of presenting our recent research at PowerTech 2021, the flagship IEEE conference for power and energy engineering in Europe. The event, organized by IEEE, PES, and Comillas Pontifical University, is being held online from Madrid between June 28 and July 2.
The paper I presented, titled “Non-Intrusive Load Monitoring using Multi-Output CNNs”, focuses on a classic challenge in the smart energy domain: how to estimate the power usage of individual household appliances when all you have is the total electricity consumption of the home. This approach, known as Non-Intrusive Load Monitoring (NILM), could play a key role in improving energy efficiency, personalized feedback, and load forecasting at scale.
Our model uses a multi-output convolutional neural network that can infer the on/off status and consumption levels of multiple devices simultaneously, making it both scalable and highly interpretable. The architecture was trained and evaluated on open benchmark datasets and performs well even in noisy, real-world settings.
It was an honor to share this work with the PowerTech community and to connect with fellow researchers tackling some of the biggest challenges in sustainable energy systems.
All the best,
Daniel Precioso
