Research on Smart Grids has recently focused on the energy monitoring issue, with the objective of maximizing the user consumption awareness in building contexts on the one hand, and providing utilities with a detailed description of customer habits on the other. In particular, Non-Intrusive Load Monitoring (NILM), the subject of this book, represents one of the hottest topics in Smart Grid applications. NILM refers to those techniques aimed at decomposing the consumption-aggregated data acquired at a single point of measurement into the diverse consumption profiles of appliances operating in…mehr
Research on Smart Grids has recently focused on the energy monitoring issue, with the objective of maximizing the user consumption awareness in building contexts on the one hand, and providing utilities with a detailed description of customer habits on the other. In particular, Non-Intrusive Load Monitoring (NILM), the subject of this book, represents one of the hottest topics in Smart Grid applications. NILM refers to those techniques aimed at decomposing the consumption-aggregated data acquired at a single point of measurement into the diverse consumption profiles of appliances operating in the electrical system under study.
This book provides a status report on the most promising NILM methods, with an overview of the publically available dataset on which the algorithm and experiments are based. Of the proposed methods, those based on the Hidden Markov Model (HMM) and the Deep Neural Network (DNN) are the best performing and most interesting from the future improvement point of view. One method from each category has been selected and the performance improvements achieved are described. Comparisons are made between the two reference techniques, and pros and cons are considered. In addition, performance improvements can be achieved when the reactive power component is exploited in addition to the active power consumption trace.
Roberto Bonfigli was born in Sant'Elpidio a Mare (FM, Italy) in 1989. He received his Bachelor's and Master's Degree in Electronic Engineering at the Università Politecnica delle Marche, respectively in 2011 and 2014. Subsequently, he received the PhD of "Doctor Europaeus" in Information Engineering, in 2018 at the same university, under the supervision of prof. Stefano Squartini. Within the research doctorate program, he made a period of visiting at the University of Lincoln (UK), under the supervision of prof. Mingjun Zhong. His research interests are the application of Digital Signal Processing and Machine Learning approaches to speech/ambiental audio processing and energy monitoring. He is author of tens peer-reviewed articles in those area, where he is a regular reviewer for several (IEEE, Springer, Elsevier) journal and conference proceedings. Stefano Squartini (IEEE Senior Member, IEEE CIS Member, and ISCA/AES Member) was born in Ancona, Italy, onMarch 1976. He got the Italian Laurea with honors in electronic engineering from University of Ancona (now Polytechnic University of Marche, UnivPM), Italy, in 2002. He obtained his PhD at the same university (November 2005). He worked also as post-doctoral researcher at UnivPM from June 2006 to November 2007, when he joined the DII (Department of Information Engineering) as Assistant Professor in Circuit Theory. He is now Associate Professor at UnivPM since November 2014. He got also the Full Professorship qualification from Italian Ministry MIUR in 2017. His current research interests are in the area of computational intelligence and digital signal processing, with special focus on speech/audio/music processing and energy management. Dr. Squartini is one of the founding members of the research group A3LAB, and has actively participated to various (funded) regional, national and European projects on multimedia Digital Signal Processing and Smart Home Energy Management. He was co-founder and CEO of the UnivPM Spin-off DowSee an engineering company developing environmentally sustainable ICT solutions for the rational use and saving of energy in smart grids, launched in 2012. He is author and coauthor of many international scientific peer-reviewed articles (more than 200), and member of the Cognitive Computation, Big Data Analytics and Artificial Intelligence Reviews Editorial Boards (starting from 2011, 2014, and 2016 respectively). He is also Associate Editor (AE) of the IEEE Transactions on Cybernetics and IEEE Transactions on Emerging Topics in Computational Intelligence (2017-to date), and of the IEEE Transactions on Neural Networks and Learning Systems (2018-to date), for which he served as AE also in the period 2010-2016. He is a regular reviewer for several (IEEE, Springer, Elsevier) Journals, Books and Conference Proceedings and in the recent past he organized several Special Sessions at international conferences with peer-reviewing and Special Issues of ISI journals. He joined the Organizing and the Technical Program Committees of more than 70 International Conferences and Workshops in the recent past. He is the Chair of the IEEE CIS Task Force on Computational Audio Processing, and a member of the IEEE CIS Task Force on Computational Intelligence in the Energy Domain. He was the lead organizer of the 4th International Workshop on Computational Energy Management in Smart Grids (2017). He is a member of the Executive Board of the SIREN (Italian Society of Neural Networks), and responsible for his University's participation to the Texas Instruments European University Program. He is also a member of the Texas Instrument Expert Advisory Panel.
Inhaltsangabe
1 Introduction 2 Non-Intrusive Load Monitoring 2.1 Problem statement 2.2 State of the Art 2.3 Datasets 2.4 Evaluation metrics 2.5 Remarks 3 Background 3.1 Hidden Markov Model (HMM) 3.1.1 Baum-Welch algorithm 3.1.2 Factorial HMM 3.2 Deep Neural Network (DNN) 3.2.1 Stochastic gradient descent (SGD) 3.2.2 Autoencoder 4 HMM based approach 4.1 Additive Factorial Approximate Maximum A-Posteriori (AFAMAP) 4.1.1 Appliance modelling 4.1.2 Rest-of-the-World model 4.2 Algorithm improvements 4.2.1 Experimental setup 4.2.2 Results 4.3 Exploitation of the reactive power 4.3.1 AFAMAP formulation 4.3.2 Experimental setup 4.3.3 Results 4.4 Footprint extraction procedure 4.4.1 Experimental setup 4.4.2 Results 5 DNN based approach 5.1 Neural NILM 5.2 Denoising AutoEncoder approach 5.3 Algorithm improvements 5.3.1 Experimental setup 5.3.2 Results 5.4 Exploitation of the reactive power 5.4.1 Experimental setup 5.4.2 Results 6 Conclusions 6.1 Future Research Topics 7 References
