1,721,002 research outputs found
Physics informed neural networks for control oriented thermal modeling of buildings
Get PDFThis work was supported by the European Union's Horizon 2020 research and innovation programme under the projects BRIGHT (grant agreement no. 957816) , RENergetic (grant agreement no. 957845) and BIGG (grant agreement no. 957047)
Physics informed neural networks for control oriented thermal modeling of buildings
Get PDFBuildings constitute more than 40% of total primary energy consumption worldwide and are bound to play an important role in the energy transition process. To unlock their potential, we need sophisticated controllers that can understand the underlying non-linear thermal dynamics of buildings, consider user comfort constraints and produce optimal control actions. A crucial challenge for developing such controllers is obtaining an accurate control-oriented model of a building. To address this challenge, we present a novel, data-driven modeling approach using physics informed neural networks. With this, we aim to combine the strengths of two prominent modeling frameworks: the interpretability of building physics models and the expressive power of neural networks. Specifically, we use measured data and prior information about building parameters to realize a neural network model that is guided by building physics and can model the temporal evolution of room temperature, power consumption as well as the hidden state, i.e., the temperature of building thermal mass. The main research contributions of this work are: (1) we propose two new variants of physics informed neural network architectures for the task of control-oriented thermal modeling of buildings, (2) we show that training these architectures is data-efficient, requiring less training data compared to conventional, non-physics informed neural networks, and (3) we show that these architectures achieve more accurate predictions than conventional neural networks for longer prediction horizons (as needed for effective control). We test the prediction performance of the proposed architectures using both simulated and real-word data to demonstrate (2) and (3) and argue that the proposed physics informed neural network architectures can be used for control-oriented modeling
Demand response for residential building heating: Effective Monte Carlo Tree Search control based on physics-informed neural networks
No full textThis work was supported in part by the European Union's Horizon 2020 research and innovation programme under the projects BRIGHT (grant agreement no. 957816) and BIGG (grant agreement no. 957047) . We also thank Marie-Sophie Verwee for her technical support in the deployment of our work
Transfer learning in transformer-based demand forecasting for home energy management system
Get PDFIncreasingly, homeowners opt for photovoltaic (PV) systems and/or battery storage to minimize their energy bills and maximize renewable energy usage. This has spurred the development of advanced control algorithms that maximally achieve those goals. However, a common challenge faced while developing such controllers is the unavailability of accurate forecasts of household power consumption, especially for shorter time resolutions (15 minutes) and in a data-efficient manner. In this paper, we analyze how transfer learning can help by exploiting data from multiple households to improve a single house’s load forecasting. Specifically, we train an advanced forecasting model (a temporal fusion transformer) using data from multiple different households, and then finetune this global model on a new household with limited data (i.e., only a few days). The obtained models are used for forecasting power consumption of the household for the next 24 hours (day-ahead) at a time resolution of 15 minutes, with the intention of using these forecasts in advanced controllers such as Model Predictive Control. We show the benefit of this transfer learning setup versus solely using the individual new household’s data, both in terms of using real-world household data
Going Beyond Counting First Authors in Author Co-citation Analysis
Get PDFThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Deployable data-driven control algorithms for residential demand response : a study on specialized, problem-specific neural architectures
No full textDe overgang naar hernieuwbare energiebronnen zoals wind en zon transformeert onze energiesystemen. Deze verschuiving is cruciaal voor duurzaamheid, maar vormt een uitdaging voor betrouwbare netwerking vanwege de onvoorspelbaarheid van deze bronnen. 'Vraagrespons'-programma's winnen aan populariteit, vooral in de residentiële sector. Deze initiatieven maken elektriciteitsverbruik flexibeler en stemmen het af op de beschikbaarheid van hernieuwbare energie, gebruikmakend van thuisbatterijen en elektrische voertuigen. Huidig onderzoek richt zich op autonome systemen die datagestuurde methoden zoals reinforcement learning gebruiken om energiepatronen te optimaliseren.
Er bestaat echter een kloof tussen geavanceerde onderzoeksoplossingen en praktische toepassingen. Deze studie onderzoekt innovatieve methoden om deze kloof te overbruggen, met focus op het beter inzetbaar en gebruiksvriendelijker maken van slimme systemen. We verkennen nieuwe benaderingen zoals op fysica gebaseerde neurale netwerken en differentieerbare beslissingsbomen. Deze innovaties moeten obstakels bij de implementatie van datagestuurde controllers in huizen overwinnen. Het doel is brede acceptatie van deze technologieën te bevorderen, waardoor huishoudens bijdragen aan netstabiliteit en de transitie naar schone energie ondersteunen
Variations on the Author
Get PDF“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Integratie in de elektriciteitsmarkt van flexibiliteit uit batterij-energieopslag
No full textA growing share of intermittent renewable generation in the electric power system is increasing the need for flexibility. At the same time, decreasing battery prices are opening up new opportunities for energy storage. Battery energy storage systems can be used for multiple applications in the power system, such as storing an excess of renewable generated energy for later consumption, wholesale market arbitrage or providing ancillary services to the grid operators. Nevertheless, the return of investment in battery energy storage systems is often still perceived too low and uncertain. Selecting the right application, combining applications, and optimising the control and the size of a battery energy storage system are important steps to reduce uncertainty and increase the return on investment.
This dissertation addresses how to make optimal use of flexibility from battery energy storage in electricity markets and the power system. The thesis provides an overview of the different applications battery storage can be used for and gives a quantitative estimation of the value battery storage can bring when delivering these applications. The results show that providing reserves for frequency control, i.e. supporting the stability of the grid frequency, is one of the applications that has the highest value for a battery storage system. Although arbitraging on short-term wholesale and imbalance markets have theoretically a higher potential value, achieving this value requires a perfect hindsight knowledge of the market prices, so that the practically achievable value lies a lot lower. Finally, there can also be considerable value in battery storage installed behind the meter, providing direct services to the electricity consumer, such as storing locally generated solar energy or reducing peak consumption.
As battery storage systems have a limited energy content, they have to be operated in a different way than traditional power plants. When for instance used to provide symmetric frequency control, a battery energy storage system needs to control its state of charge to ensure the battery is never empty nor full, as this would mean the symmetric frequency control capacity is not available any more. This thesis presents a detailed, holistic framework to optimise such a state of charge controller and determine the optimal size of a battery storage system used for frequency reserves, taking degradation and regulatory requirements into account. As a case study, the optimisation framework is applied to the German frequency containment reserve market, providing some novel, relevant insights into the economics and sizing of a battery energy storage system in this market.
Consecutively, this thesis looks at combining multiple applications simultaneously with battery storage installed behind the meter at residential and industrial consumers. The thesis presents optimised control strategies which allow the use of battery storage for the combination of frequency reserves with increasing self-consumption or with peak shaving. Stochastic optimisation is used together with robust optimisation techniques, giving a safe and tractable approximation to chance constraints, while dynamic programming is adopted to combine the longer-term objective of peak shaving with the daily decision making in the frequency reserve market. Case study results using real data show that there are indeed synergies when combining frequency reserves with increasing self-consumption or with peak shaving and the resulting controllers are able to significantly increase the value of a battery energy storage system compared to the use of the battery energy storage system for one application only.
Finally, as battery storage is often connected to the distribution grid, this dissertation discusses the impact of distribution grid constraints on the aggregated flexibility from battery storage or from other flexible assets connected to the low-voltage distribution grid. The thesis focusses on a regulatory constraint which has been put in place in Belgium to prevent congestion of the distribution grid, limiting the frequency control capacity these assets can provide. A distributed optimisation algorithm is proposed to maximise the total frequency control capacity from low-voltage grid connected flexible assets while respecting these distribution grid constraints.status: Publishe
Monitoring of photovoltaic systems and forecasting of power system imbalance using neural networks
No full textDe transitie naar hernieuwbare energie brengt meer uitdagingen met zich mee dan 'enkel' voldoende productie capaciteit bouwen. Deze thesis onderzoekt hoe we deze uitdagingen kunnen aanpakken met behulp van artificiële intelligentie. Specifiek stellen we methoden voor om fouten in zonnepanelen op een kost-efficiënte manier te detecteren en identificeren. Fouten in zonnepanelen, zoals slijtage aan bedrading en kortsluitingen, leiden tot energieverlies en kunnen zelfs brand veroorzaken. Toch worden de meeste zonnepaneel systemen, zoals die op je dak, niet gemonitord, voornamelijk omdat bestaande monitoring methoden dure sensoren gebruiken. Wij stellen voor om zonnepanelen te monitoren door hun productie te vergelijken met weersschattingen of met de productie van nabijgelegen zonnepaneel systemen. Deze informatie wordt verwerkt met neurale netwerken om accuraat fouten te identificeren zonder dure sensoren. Een tweede uitdaging waarop we ons focussen is het voorspellen van energie tekorten of overschotten in het Belgisch elektriciteitsnet. Door de opkomst van hernieuwbare energie, elektrische voertuigen, warmtepompen, etc. is het steeds moeilijker om elektriciteit productie en consumptie te balanceren. Een grote onbalans kan leiden tot instabiliteiten van het net en zelfs stroomuitval. Om dit aan te pakken, ontwikkelen we een neuraal netwerk om zo nauwkeurig mogelijk de Belgische onbalans te voorspellen, zodat deze tijdig gecompenseerd kan worden
- …
