- This event has passed.
Energy-Open 2018
Nov 29, 2018 - Nov 30, 2018
Sponsored by
Summary
Global warming, climate agreements, EU policies to reduce CO2 emissions and many national energy programs reflect the need to make our energy systems more sustainable. This process, called energy transition, entails a fundamental change of the energy system from a centralized system with electricity production mainly based on fossil fuels to a decentralized local system based on sustainable energy production based on e.g. solar, wind, tidal, hydro or geothermal sources. Other changes include the ongoing growth of the electricity consumption resulting from an electrification of transport (electric vehicles) and heating (heat pumps) resulting in an increasing integration between the different energy systems of electricity, transportation and heat.
The envisioned direction of the energy transition leads to drastic changes within the current energy system; e.g. larger parts of the energy generation will be based on fluctuating, non-controllable and hard to predict sources like wind and sun and the ongoing electrification will introduce large peaks in the consumption. To be able to ensure also in the future the stability of this energy system and the balancing of energy supply and demand without extremely extending the capacities of the underlying grids, the setup and the organization of our energy system has to be changed drastically. This includes different roles of stakeholders (e.g. consumers get prosumers), changing or new market structures and players (e.g. local (independent) micro-grids at community level and smart buildings/homes), changing end-user involvement and new technologies (e.g. storage and conversion between energy carriers).
The workshop aims to bring together researchers working on new concepts and solutions
supporting the energy transition. Topics of interest include, but are not limited to:
- Decentralized energy management
- Control of Storage in Smart Grids
- Optimization algorithms for Smart Grids
- Power quality in distribution grids
- Prediction algorithms of renewable generation
- Legislation for energy transition
- User behavior
Organizers
Madeleine Gibescu (TU Eindhoven)
Phuong Nguyen (TU Eindhoven)
Bert Zwart (TU Eindhoven and CWI)
Keynote Speakers
Bert Claessens (REStore)
Florian Dörfler (ETH Zurich)
Damien Ernst (Université Liège)
Johann Hurink (University Twente)
Zita Vale (Polytechnic Institute of Porto)
Abstract Submission
There will be a poster session, and a limited number of slots for contributed talks. If you want to present your research, please indicate this on the registration form, and provide an abstract of at most 500 words.
Abstracts
Bert Claessens
The virtual power plant
Using a virtual power plant for Demand response applications is a non-trivial challenge from a, among others, control context. This is direct resulting from the fact that the problem is a sequential decision making problem under uncertainty with a large state- and action-space. To meet these challenges REstore uses and explores a range of modelling and optimization techniques, both model-based and model-free in a centralized and decentralized setting. In this presentation, REstore will explain how it controls one of the most advanced VPPs in Europe, comprising a 18 MW battery, providing Frequency Containment Reserves (FCR). A second example, shows how to combine FCR and self-consumption by solving a chance-constrained optimization problem using robust optimization, resulting in a significant value increase for batteries. Finally, a third set of results obtained through a collaboration with UC Berkeley demonstrates how reinforcement learning can be both a practical and efficient control paradigm when combined with model based control in the context of residential demand response.
Through this presentation, we want to convey the message that real-life demand response benefits from being able to handle and combine a range of optimization and modelling techniques.
BIO: Bert holds a PhD in applied physics. After developing control algorithms for photolithography applications for four years, he switched to the field of smart grid research. He was active in this field for six years before taking up a research position at REstore. His main interests are directed towards residential demand response and applying state of the art in artificial intelligence for energy applications.
Florian Dörfler
Real-Time Feedback Optimization of Power Systems
I will focus on online optimization of AC power systems in closed loop. In contrast to the conventional approach where an optimal power flow solution is computed offline and online controllers enforce these set-points, our objective is to design an adaptive feedback controller that steers the system robustly and in real time to the optimal operating point. Our methodological approach is based on online algorithms for manifold optimization that can be applied in feedback with real-time measurements and actuation. We treat the power flow equations as implicit constraints that are naturally enforced by the physics and hence give rise to the power flow manifold. Based on our theoretical results for this type of optimization problems, we propose a projected gradient descent scheme on the power flow manifold. In detailed simulation case studies we validate the performance of our algorithm and show that it reliably tracks the time-varying optimum of the underlying AC optimal power flow problem.
BIO: Florian Dörfler is an Assistant Professor at the Automatic Control Laboratory at ETH Zürich. He received his Ph.D. degree in Mechanical Engineering from the University of California at Santa Barbara in 2013, and a Diplom degree in Engineering Cybernetics from the University of Stuttgart in 2008. From 2013 to 2014 he was an Assistant Professor at the University of California Los Angeles. His primary research interests are centered around distributed control, complex networks, and cyber–physical systems currently with applications in energy systems and smart grids. His students were winners or finalists for Best Student Paper awards at the 2013 European Control Conference, the 2016 American Control Conference, and the 2017 PES PowerTech Conference. His articles received the 2010 ACC Student Best Paper Award, the 2011 O. Hugo Schuck Best Paper Award, the 2012-2014 Automatica Best Paper Award, and the 2016 IEEE Circuits and Systems Guillemin-Cauer Best Paper Award. He is a recipient of the 2009 Regents Special International Fellowship, the 2011 Peter J. Frenkel Foundation Fellowship, and the 2015 UCSB ME Best PhD award.
Svetlana Dubinkina
Uncertainty quantification and random energy systems
Predicting the amount of gas or oil extracted from a subsurface reservoir depends on the soil properties such as porosity and permeability. These properties, however, are highly uncertain due to the lack of measurements. Therefore decreasing these uncertainties is of a great importance.
Mathematically speaking, permeability can be represented by a random process, which in turn leads to a random partial differential equation.
The solution of such a partial differential equation, for example pressure, is only partially observed and, moreover, contaminated with measurement errors. Therefore, instead of a well-posed forward problem of finding pressure from certain permeability, we are faced with an ill-posed inverse problem of finding uncertain random process from a few pressure measurements. The golden method to deal with this type of problems is Markov Chain Monte Carlo, which is however, computationally prohibitive for large systems. Another popular method is Ensemble Kalman Filter, which is computationally affordable but assumes Gaussian processes. We develop an inverse method that is computationally affordable and does not make any assumptions.
In the future electricity markets, energy suppliers may act on the market using day-ahead planning that specifies the amount of energy to be consumed or produced for set of future time intervals. This asks for forecasting of production and consumption for each of these time intervals. A further trend is towards local energy markets in order to integrate renewable energy sources and controllable devices into the energy system. As a result, the market size but also the size of the participating players is getting smaller and making an accurate day-ahead planning as base for acting on these markets is getting more important. Especially the small size of the participants makes that deviations from the offered planning very likely to occur.
In order to stay as close as possible to the offered planning, the participants may use available flexibilities of some of their devices to correct the deviations by themselves. In this talk we present a real-time control method which aim is to deal with such deviations on device level to realize a planned profile as good as possible. Based on the reason of deviation from the planning, we use different approaches to use the flexibilities of the devices in order to follow the offered planning as close as possible.
Damien Ernst
Reinforcement learning, energy systems and deep neural nets
Reinforcement leaning is a highly-successful subfield of artificial intelligence where an agent is ought to interact with its environment to maximize a sum of rewards. In this talk, Damien Ernst will argue that this learning paradigm can be very powerful to solve many decision-making problems in the energy sector, as for example investment problems, the design of bidding strategies for playing with the intraday electricity market or problems related to the control of microgrids. He will also describe some very recent progresses in the field on deep reinforcement learning that could be used to foster the performances of reinforcement learning agents when confronted with environments that can exhibit sudden changes in their dynamics, as it is often the case with energy systems.
Maryam Hajighasemi
Planning based real-time control
In the future electricity markets, energy suppliers may act on the market using day-ahead planning that specifies the amount of energy to be consumed or produced for set of future time intervals. This asks for forecasting of production and consumption for each of these time intervals. A further trend is towards local energy markets in order to integrate renewable energy sources and controllable devices into the energy system. As a result, the market size but also the size of the participating players is getting smaller and making an accurate day-ahead planning as base for acting on these markets is getting more important. Especially the small size of the participants makes that deviations from the offered planning very likely to occur.
In order to stay as close as possible to the offered planning, the participants may use available flexibilities of some of their devices to correct the deviations by themselves. In this talk we present a real-time control method which aim is to deal with such deviations on device level to realize a planned profile as good as possible. Based on the reason of deviation from the planning, we use different approaches to use the flexibilities of the devices in order to follow the offered planning as close as possible.
Bart Homan / Victor Reijnders
From simulation to field test
The ‘GridFlex Heeten’ project started in 2017 with following goal: to use the local flexibility of batteries and households to reduce the overall stress on the network, thereby postponing or even entirely avoiding reinforcing the grid as well as reducing losses in the grid. In the project, this is done by realizing a local energy market to test innovative price mechanisms. As an exemption on the Dutch energy law for experimenting was obtained, we can validate these concepts in a field test.
However, before we can start the field test, we first need the tools, equipment and households that participate in the test. 47 households in Heeten (a village in the Netherlands) were willing to participate in the field-test, all of which are located behind a single transformer. Some of the houses are already equipped with PV-panels, and all households have home energy management systems. In addition to that we will equip 22 households with a Seasalt battery, and a Battery Management System (BMS) with which the behaviour of the battery can be monitored and influenced.
An important part of the BMS is the DEMKit software, developed at the University of Twente (UT). However, management of the battery is only part of what the DEMKit software is used for in the 47 houses. On the basis of data on the weather, houses, and residents, and models of houses and the available devices the software can predict energy usage and production for the 47 houses. Based on these predictions desired behaviour of each battery can be planned; i.e. when is the most feasible time to charge and discharge each battery. Lastly the battery behaviour can be influenced in real time. The principles of how to match local energy demands and production have been thoroughly investigated in our “16 houses case” papers.
The software was also made compatible with an off-the-shelf inverter system that interacts physically with the battery. The Seasalt battery is a new battery, for which no off-the-shelf BMS exists yet, so an appropriate model was developed at the UT to predict the behaviour of the Seasalt battery; the DiBu-model.
With all assets in place, the testing can finally start. A whole new area of research now emerges. By reducing the stress on the network, savings are earned as a group, but how should we divide this? How do we take into account the community feeling, fairness, and user participation? Should we give someone who reduce the stress further more money? How do we communicate different price mechanisms to participants and give them insight in these? More research needs to be done, but one thing is clear. You need a field test to validate the research, but you also need research to make a field test a success.
Johann Hurink
Decentralized Energy Management
The energy transition asks for a paradigm shift in the organisation of the energy supply chain from centralized control towards decentralized energy management. In this talk a general concept for such a bottom up energy management concept is presented. Its base consist of three main step: prediction, planning and real time control. The whole process is organized along a tree structure, where the leafs correspond to the loads (controllable and non-controllable) in the system and the nodes in the other layers are aggregating nodes corresponding to e.g. households, transformers, markets, etc.
In a second part some specific device level planning problems are addressed. Within the envisioned decentralized energy management concept, devices that offer flexibility in their load profile play an important role. These devices should schedule their flexible load profile based on steering signals received from centralized controllers. Some of these problem of finding optimal or robust device schedules based on the received steering signals falls into the framework of resource allocation problems and are variations or extensions of more classical problems from literature.
Stella Kapodistria
ML interpretability
In various sectors, for the purpose of decision making under uncertainty (e.g., preventive maintenance), a machine learning (ML) approach is used. However, such approaches frequently suffer form lack of interpretability of the results and are considered to be black boxes. Understanding the structure of the ML decisions is crucial in assessing trust, which is fundamental if one plans to take action based on a prediction, and it facilitates adapting the decision rules when undergoing design changes, plus it enables design improvements.
In this talk, we will demonstrate how to open one of the black boxes of ML in the context of maintenance for the energy sector.
Ioannis Lampropoulos
A system perspective to the deployment of flexibility through aggregator companies in the Netherlands
Recent developments in distribution grids, environmental policy, and the energy market liberalisation process, have resulted in a quest for flexibility in power systems operation, with the focus increasingly placed on the aggregation of distributed resources. A generic method is proposed for the identification of opportunities, barriers and potential solutions in developing flexibility mechanisms through aggregator companies by concentrating on the market integration aspects. The method is applied to the Netherlands as a case study, and the outcome is a state-of-the-art review of the electricity market development concerning all relevant issues for advancing the market integration of aggregator companies within the Dutch system and in line with the new European grid codes. Opportunities were framed among six categories which outline the potential for the provision of market-based products and services in the Dutch system, whereas barriers were decomposed into market, regulatory, technical and social issues. In total, there were thirty-one (31) identified elements of barriers which are categorised among six areas where opportunities for the provision of market based products and services were identified, namely wholesale trade in spot markets, ancillary services markets, over-the-counter trade of flexibility services, retail markets and other miscellaneous issues related to the provision of data services such as the roll-out status of smart metering systems and access to metered data. A set of recommended actions is provided to facilitate the market integration of aggregator companies in the Netherlands, which point out the need for policy interventions and follow-up research activities.
Koos v.d. Linden
Toolbox for optimal trading in day ahead and reserve markets
An aggregator of flexible demand, such as charging a battery, can exploit this flexibility to reduce operational costs by optimally deciding when to buy energy on the day-ahead market, when to buy it on the real-time market, and when and how to make competitive bids in the reserve market.
A number of optimization methods have been proposed for this problem, but a comparison of these methods and their performance in different market configurations has not been made yet.
We have developed a toolbox that allows for an evaluation of these methods in different market configurations.
We present the use of this toolbox by means of a practical use case. In this example, an electrical vehicle aggregator wants to investigate the possibility of selling its flexibility on the reserve market versus trading on the day-ahead market. The aggregator also wants to investigate the added value of using stochastic optimization. We show how to investigate this by using our toolbox.
Anne Markensteijn
A graph-based framework for load flow analysis of multi-carrier energy systems
Multi-carrier energy systems (MES) have become more important over the years as the need for efficient, reliable and low carbon energy systems increases. In MES, different energy carriers, such as electricity and heat, interact with each other leading to one combined energy network. They have higher performance than classical single-carrier energy systems due to increased flexibility, reliability, use of renewables and distributed generation, and reduced carbon emission. MES could also be used to support energy self-sufficient regions, which operate disconnected from a national grid. Because these MES integrate two or more separate energy systems, they are sometimes called integrated energy systems.
An important tool for designing energy systems is a steady-state load flow analysis of the energy
transportation or distribution network. This analysis determines the network state parameters and
the flow of energy in the network. Conventional load flow models for the single-carrier networks have been widely studied. However, they are not able to capture the full extend of the coupling. Recently, different models for multi-carrier networks have been proposed, either using the energy hub concept or using a case specific approach. Although the energy hub concept can be applied to a general MES, it does not state how the graphs of single-carrier networks can be combined into one multi-carrier network. Furthermore, the recent studies do not consider the effect of this combination on the load flow model for a MES. Usually, the coupling models introduce more unknowns than equations to the system. Therefore, additional equations or boundary conditions must be added for the system to be (uniquely) solvable. Depending on how the single-carrier networks are connected to form an integrated network, and depending on the choice of the additional equations, the resulting system of load flow equations might have none, one or infinite solutions.
We present a general steady-state load flow model framework for MES, based on graph and network theory. We propose a graph representation for MES by introducing coupling nodes and dummy links to combine the single-carrier networks into a multi-carrier network. Furthermore, we provide guidelines for combining the single-carrier load flow models, with conversion models, into one multicarrier load flow model. Based on the resulting model framework, we discuss the effect of coupling on the system of non-linear equations.
The framework is tested on a small MES consisting of gas, electricity, and heat. Different coupling models were used for comparison with case studies available in literature. The tests showed
that our framework is applicable to a general network, as opposed to the case specific approach, and that it can be used with different coupling models, as opposed to the energy hub concept. Moreover, it determines both energy flows and network state parameters. Therefore, our proposed graph-based framework for steady-state load flow analysis of MES extends and generalizes the currently available models.
(joint work with Cees Vuik)
Jorrit Nutma
A field-tested architecture for an open energy market to enable cost-efficient congestion management
In recent years, an enormous amount of work has been done to deal with the expected challenges that arise due to the changes in our energy system. We observe that academics are developing smart algorithms to exploit flexibility and optimize the energy system. At the same time, DSO’s are investigating how flexibility can be used to deal with envisioned challenges in their grid. We also see new market parties arising because changes in the energy system enable (and need) new businesses cases. In one of the working packages of the Interflex (H2020) project, those three trends meet each other. The goal is to study the economic value of flexibility, when it is used for congestion management.
In the project, companies and a research institute are cooperating in a Dutch pilot to develop a value chain for congestion management services. This means that, in addition to technical solutions for congestion management, the project also concerns contractual agreements and financial flows between parties. The basis for the value chain is an architecture which features a clear separation of roles and responsibilities. By doing this, it ensures an open market for aggregators to provide congestion management services to DSO’s. As a result, the price for those services is likely to be reduced when new players enter the market, reducing the costs for congestion management through flexibility.
Our architecture consists of three layers. At the bottom we have Energy Service Companies (ESCo’s), who are responsible for maintaining and operating distributed energy resources. The second layer is formed by aggregators, who are using the flexibility of the energy resources by trading on energy wholesale markets and providing congestion management services. Finally, the DSO is in the upper layer and operates a Grid Management System, which forecasts congestion and dispatches congestion management services provided by the aggregators.
The architecture uses as much existing communication standards as possible. The selected communication standards provide a clear separation of concerns and do not rely on a centralized system. In this way, aggregators have the freedom to implement their own business logic, or to change the business logic in the future when their technology evolves. Aggregators have an interface with the DSO via the open specification called USEF (https://www.usef.energy/) and the interface between aggregator and ESCo is the Energy Flexibility Interface (EFI, https://flexible-energy.eu/efi/).
To validate the concepts, the participating parties implemented the architecture to run it in a field-test. It comprises the following types of distributed energy resources: charge points for EV’s, a neighborhood-scale battery, curtailable PV panels, and inflexible loads.
(joint with Wilco Wijbrandi, Bob Ran, Joost Laarakkers)
Eric Persoon
A model for the variable electricity from wind and PV panels and its application
In developing a system for the electricity supply in the Netherlands a model is needed that contains the main parameters of the highly variable supply of electricity during the year. We assume that the electricity generation will be done with zero CO2 emission and this will require a significant amount of installed power of wind turbines and PV panels. Due to the variable nature of this electricity two main technologies will be needed to be able to match this supply with the expected demand. One is the conversion of electricity to hydrogen using electrolyzers and the other is the usage of batteries to reduce the variability over a short duration of time. Using this model we can optimize and minimize the required installed power of wind turbines and PV panels. Moreover it is allows us to estimate the needed capacity of the electricity grid in the future which will be much higher than the current capacity. A model will be presented that allows us to do this analysis and this will be presented. Moreover we also studied how the demand will look like in the future where the several kinds of energy demand will be provided for by either electricity or hydrogen. In a last step we developed a system that matches the supplied electricity with the demand site.
Jeroen Reinders
Open problem: Analytical Probabilistic Load Flow Using Gaussian Mixture Models
Uncertainty in power systems is rising, due to decentralized renewable power production and ongoing electrification of society. Analytical probabilistic load flow methods provide a means to incorporate uncertainty in the load flow equation, retaining acceptable accuracy without requiring significant computational power. A commonly adopted assumption is a Gaussian distribution of input variables, which allows major simplifications to the load flow equation. This assumption however, does not hold for networks with high shares of renewable power production. Could this method still be used by representing more complicated distributions as a combination of several Gaussian distributions (Gaussian Mixture Models)?
– Simplifying the non-linear load flow equation, using first-order Taylor approximation, I can represent power flows in grids with Gaussian Distributions as input parameters.
– Representing non-Gaussian distributions as combinations of several Gaussian distributions (GMM) I want to use this method for non-Gaussian distributions.
– I have been calculating load flows for each combination of GMM using 1st order Taylor approximation, then convoluting (with weights from GMM) to come to one result.
– This does not give me correct results (comparing to numerical probabilistic load flow, using Monte-Carlo style selection of variables). I’m wondering whether I’m making mistakes in applying GMM and/or convolution, or is this approach theoretically impossible?
Arjan vd Schaft
The flow equations of electrical networks: from Kirchhoff to smart grids
The analysis of the flow equations of electrical networks goes back to the classical work of Kirchhoff, but still poses interesting questions, of direct relevance to electrical power networks. In this talk we will present, directly starting from Kirchhoff, some new observations that are of relevance to solving the load flow equations for complex networks. Another open problem that we will address is the identifiability of the conductance parameters from the boundary behavior of the electrical network.
Martijn Schoot Uiterkamp
Fill-level prediction in online valley-filling algorithms for electric vehicle charging
Due to the large increase in electric vehicles (EVs), smart charging strategies are required in order for the distribution grid to accommodate all these EVs. A popular EV scheduling approach in order to flatten the overall power profile is valley-filling. Intuitively, this approach “fills up” the given base load profile with EV charge until the charging requirement is met and an “optimal fill-level” is reached. This optimal level uniquely characterizes the optimal EV schedule in the sense that we can easily reconstruct this optimal schedule given the optimal level and the base load profile.
To compute an optimal EV schedule, perfect knowledge of the base load profile is required. However, in practice, this information is not known beforehand. To compensate for this, many EV scheduling approaches use predictions of the base load profile as input for a (deterministic) EV scheduling algorithm. Unfortunately, accurate prediction of base load profiles is very difficult, especially when local energy production, e.g., photovoltaic (PV) is included. As a consequence, the resulting EV schedules are very sensitive to prediction errors in the base load profile.
Online valley-filling algorithms circumvent the above problem by predicting the optimal fill-level instead of the base load profile. Using this prediction and the aforementioned characterization of optimal EV schedules, one can construct an online EV schedule. The main advantage of this approach is that charging decision for each time slot can be made at the start of that time slot. This is because this decision requires only knowledge of the predicted fill level and the base load of the current time slot. As a consequence, we can adequately react to sudden consumption or production peaks and adjust the EV charging rate accordingly. Therefore, online valley-filling is more robust against variations in both the time and magnitude of consumption and production peaks.
Our contribution is concerned with the question how well the optimal fill-level can be predicted and thus how suitable online valley-filling algorithms are for real-life energy management systems. For this, we develop a simple but accurate prediction approach. In this approach, we model the optimal fill-level as a stochastic variable with an unknown probability distribution. We learn this distribution using historic base load data and optimal EV schedules for previous (fictional) charging sessions. Using basic probability theory, we can compute a prediction such that the probability of over-predicting the optimal fill-level equals a given risk parameter. This parameter models the preference for over- or under-predicting the optimal fill-level, which is relevant, e.g., when multiple EVs are being charged simultaneously.
Using our approach, we study the influence of several factors on the predictability of the fill-level, such as the charging requirement, time of day, and PV prediction. Furthermore, we analyze the influence of the timespan of the historic training data to learn the distribution of the optimal fill-level. All in all, we show that online valley-filling algorithms can realize near-optimal online EV schedules when the fill-level is predicted using our approach.
Shahab Shariat Torbaghan
Flexibility market framework
At the distribution level, due to large scale integration of distributed renewable energy sources (DRES), incidents such as network asset congestion, as well as over/under voltage are becoming a new routine that system operators (e.g., DSOs) have to deal with on daily basis. One solution to tackle the increasing uncertainty and emerging problem(s) is to increase the flexibility of the system. This can be done by enabling larger involvement of pro-active end-users and resolving the network operation limit violation in the low-voltage grids by implementing demand response (DR) programs.
The concept of Decentralized Energy Management or Demand Response is emerging as one of the main approaches to resolve the violations of the network operation limits and to increase the flexibility of the system. This paper introduces an interaction framework for trading flexibility among proactive end-users in an economically efficient way. It proposes new market participants with their roles and functionalities, that will operate alongside the existing ones to ensure market efficiency and to enable secure operation of distribution grids. The proposed framework consists of a main mechanism called ‘ahead-markets scheduling’. The ahead-markets scheduling provides a trading platform that allows market participants to reflect their need(s) for flexibility and to monetize flexibility services in a fair and competitive manner. It enables flexibility trades which will eventually facilitate network management for the system operator.
To ensure success of such a program, the DSO needs to determine the minimum amount of flexibility that it requires to dispatch, as well as an adequate price or command signal to encourage a suitable response from the pro-active end users.
For this panel, we propose an optimization framework that solves the optimal flexibility dispatch (OFD) problem for the DSO. The OFD problem delivers the quantity and price of the flexibility that the DSO needs to procure at the distribution level.
The OFD problem considers the physics of power flows, generation units and consumer devices. Such problems are generally non-convex, nonlinear, complex-valued problems that are difficult to solve. To make them tractable in close to real-time, we use second-order cone relaxed formation of the AC power flow formulation to convexify the OFD problem.
Ni Wang
Formulating coordination mechanisms in the investment optimization models of self-sufficient regional energy systems
Dutch regional municipalities increasingly take an active role in the transition to more sustainable and autonomous energy supply systems, using local energy sources like wind, solar and biomass. In this study, the key research areas in self-sufficient regional energy systems are reviewed. It is found that existing literature show that it is of importance to consider the coordination mechanisms between local actors. However, there lacks a study on the formulation of the coordination mechanisms in the investment optimization models of self-sufficient regional energy systems. We propose to define different information flows, energy flows and monetary flows as the differentiators for three coordination mechanisms, namely one with a central planner, one with a regional market, and a reference one with individual supply. Accordingly, the mathematical investment optimization models for each coordination mechanism are formulated.
Zita Vale
Intelligent management of distributed energy resources in the context of smart grids
Distributed generation, demand response, distributed storage, and electric vehicles are bringing new challenges to the power and energy sector. The keynote addresses the current and envisioned solutions for the management of these distributed energy resources in the context of smart grids.
Artificial intelligence based approaches bring important new possibilities enabling efficient individual and aggregated energy management. Such approaches can provide different players aiming to accomplish individual and common goals in the frame of a market-driven environment with advanced decision-support and automated solutions.
MARTINE (Multi-Agent based Real-Time INfrastruture for Energy), a platform to support real-time energy management and simulation of buildings and smart grids, will be described. The platform will be used as the basis to present different data-driven and cognitive approaches to support efficient energy management in buildings and smart grids.
Bio: Zita Vale is full professor at the Engineering School of the Polytechnic Institute of Porto, Portugal. She received her diploma in Electrical Engineering in 1986 and her PhD in 1993, both from University of Porto. She works in the area of Power and Energy Systems, with special interest in the application of Artificial Intelligence techniques. She has been involved in more than 50 funded projects related to the development and use of Knowledge-Based systems, Multi-Agent systems, Neural networks, Particle Swarm Intelligence, and Data Mining. The main application fields of these projects comprise:
– Smart Grids, accommodating an intensive use of Renewable Energy Sources, Distributed Energy Resources (DER), namely Distributed Generation (DG), storage, electrical vehicles, including the ones with gridable capability (V2G), and demand flexibility. Real-time management and simulation of energy resources, considering electrical networks, buildings, and diverse Internet of Things (IoT) and Machine to Machine (M2M) approaches are relevant aspects of her work in this field;
– Electricity markets, addressing decision-support for market participants, prices and tariffs, ancillary services, energy transactions, service provision, and market simulation in the scope of wholesale and emergent local markets. The integration of DER, demand response, and EVs in electricity markets and Transactive Energy (TE) approaches are important aspects of her work. Her work also focuses on the conception, development and test of new business models for market participants and aggregation models for energy resources and the respective management and operation methods.
She published over 800 works, including more than 100 papers in international scientific journals, and more than 500 papers in international scientific conferences.
See some of her selected publications in http://www.gecad.isep.ipp.pt/GECAD/Pages/Pubs/Publications.aspx
Registration
To register for this event, please use the online registration tool.
Practical Information
Link to Information Page