Publications

Gaming strategies in European imbalance settlement mechanisms

S.S. Karimi Madahi, K. Bruninx, B. Claessens and C. Develder

Transmission System Operators (TSOs) rely on balancing energy provided by Balancing Service Providers (BSPs) to maintain the supply-demand balance in real time. Balance Responsible Parties (BRPs) can deviate from their day-ahead schedules in response to imbalance prices, e.g., by controlling flexible assets such as batteries. According to the European Electricity Balancing Guideline (EBGL), these imbalance prices should incentivize BRPs performing such implicit or passive balancing to aid the TSO in restoring the energy balance. Here we will demonstrate that BRPs are unintentionally offered the opportunity to exploit gaming strategies in European imbalance settlement mechanisms, because of a disconnect between sub-quarter-hourly dynamics that determine the imbalance prices and the financial settlement on a quarter-hourly basis. We illustrate this behavior in a case study in Belgium and the Netherlands. Our results show that, in both countries, BRPs can, in theory, exploit the imbalance mechanism by increasing the instantaneous system imbalance for some minutes within the quarter-hour that determine the imbalance price, while still contributing to restoring the system balance for the rest of the quarter-hour.

Gaming strategies in European imbalance settlement mechanisms

S.S. Karimi Madahi, K. Bruninx, B. Claessens and C. Develder

Generalization of graph neural network models for distribution grid fault detection

B. Karabulut, C. Manna and C. Develder

Fault detection in power distribution grids is critical for ensuring system reliability and preventing costly outages. Moreover, fault detection methodologies should remain robust to evolving grid topologies caused by factors such as reconfigurations, equipment failures, and Distributed Energy Resource (DER) integration. Current data-driven state-of-the-art methods use Recurrent Neural Networks (RNNs) for temporal modeling and Graph Neural Networks (GNNs) for spatial learning, in an RNN+GNN pipeline setting (RGNN in short). Specifically, for power system fault diagnosis, Graph Convolutional Networks (GCNs) have been adopted. Yet, various more advanced GNN architectures have been proposed and adopted in domains outside of power systems. In this paper, we set out to systematically and consistently benchmark various GNN architectures in an RNN+GNN pipeline model. Specifically, to the best of our knowledge, we are the first to (i) propose to use GraphSAGE and Graph Attention (GAT, GATv2) in an RGNN for fault diagnosis, and (ii) provide a comprehensive benchmark against earlier proposed RGNN solutions (RGCN) as well as pure RNN models (especially Gated Recurrent Unit (GRU)), particularly (iii) exploring their generalization potential for deployment in different settings than those used for training them. Our experimental results on the IEEE 123-node distribution network show that RGATv2 has superior generalization capabilities, maintaining high performance with an F1-score reduction of  12% across different topology settings. In contrast, pure RNN models largely fail, experiencing an F1-score reduction of up to  60%, while other RGNN variants also exhibit significant performance degradation, i.e., up to  25% lower F1-scores.

Generalization of graph neural network models for distribution grid fault detection

B. Karabulut, C. Manna and C. Develder

Scalable attention-based reinforcement learning method for multi-asset control

S.S. Karimi Madahi, G. Gabriele, B. Claessens and C. Develder

In this paper, we propose a scalable centralized reinforcement learning method to jointly optimize a fleet of flexible assets. The attention layer in our proposed architecture enables the agent to make decisions for each asset based on both its local and aggregated asset-specific information. As a proof-of-concept, we investigate the performance of the proposed method on an electric vehicle (EV) charging problem. The results show that the trained agent can effectively control multiple EVs to achieve a common objective (load flattening in our case).

Scalable attention-based reinforcement learning method for multi-asset control

S.S. Karimi Madahi, G. Gabriele, B. Claessens and C. Develder

Interpretable reinforcement learning for heat pump control through asymmetric differentiable decision trees

T. Van Puyvelde, M. Zareh and C. Develder

In recent years, deep reinforcement learning (DRL) algorithms have gained traction in home energy management systems. However, their adoption by energy management companies remains limited due to the black-box nature of DRL, which fails to provide transparent decision-making feedback. To address this, explainable reinforcement learning (XRL) techniques have emerged, aiming to make DRL decisions more transparent. Among these, soft differential decision tree (DDT) distillation provides a promising approach due to the clear decision rules they are based on, which can be efficiently computed. However, achieving high performance often requires deep, and completely full, trees, which reduces interpretability. To overcome this, we propose a novel asymmetric soft DDT construction method. Unlike traditional soft DDTs, our approach adaptively constructs trees by expanding nodes only when necessary. This improves the efficient use of decision nodes, which require a predetermined depth to construct full symmetric trees, enhancing both interpretability and performance. We demonstrate the potential of asymmetric DDTs to provide transparent, efficient, and high-performing decision-making in home energy management systems.

Interpretable reinforcement learning for heat pump control through asymmetric differentiable decision trees

T. Van Puyvelde, M. Zareh and C. Develder

Risk-Sensitive reinforcement learning-based strategies for Dutch implicit balancing

S.S. Karimi Madahi, F. Pavirani, B. Claessens and C. Develder

Adopting renewable energy sources (RES) can pave the way toward reaching net-zero carbon emissions. However, the intermittent nature of RES can pose significant challenges to balance responsible parties (BRPs) and transmission system operators (TSOs) in maintaining the balance of the electricity grid. BRPs can assist TSOs in balancing the grid by occasionally deviating from their nomination to help reduce the system imbalance, which is called implicit balancing. In this paper, we propose data-driven implicit balancing strategies for BRPs in the Dutch imbalance settlement mechanism. Dutch implicit balancing is challenging due to the Dutch imbalance pricing calculation, which is a combination of single and dual pricing methods. To cope with this challenge, a risk management perspective is incorporated into the proposed method through distributional reinforcement learning. Distributional reinforcement learning agents are trained to manage a BRP's battery in the presence of wind farm generation stochasticity to reduce its imbalance cost. Dutch imbalance data of 2024 are used to assess the performance of the learned implicit strategies. The proposed method is benchmarked against deterministic model predictive control and a rule-based controller. The results show that both risk-neutral and risk-averse agents improve daily profit by 29.3% and 20.7% respectively, compared to the rule-based controller. Moreover, the risk-averse agent decreases the average portfolio deviation during dual pricing situations by 19.2% compared to the risk-neutral agent, resulting in a lower imbalance cost for the BRP in these situations.

Risk-Sensitive reinforcement learning-based strategies for Dutch implicit balancing

S.S. Karimi Madahi, F. Pavirani, B. Claessens and C. Develder

System-Aware reinforcement learning for optimized implicit imbalance participation in Belgium

F. Pavirani, S.S. Karimi Madahi, B. Claessens and C. Develder

The increasing integration of renewable energy sources into electrical grids has disrupted the balance between production and consumption. To address this challenges, transmission system operators such as the Belgian one have introduced imbalance tariffs that penalize harmful energy deviations. Although the imbalance settlement mechanism allows balance responsible parties to dynamically adjust their energy positions, it also exposes them to significant risks. In fact, Belgian imbalance prices are determined retrospectively at the end of each settlement block, meaning that energy deviations occur under uncertain pricing conditions. Reinforcement learning (RL) offers a promising solution for navigating this uncertainty thanks to its ability to manage stochastic environments and deliver long-term rewards. However, achieving profitable participation in imbalance settlement requires more than just handling price volatility; it also demands a deep understanding of the grid dynamics. This paper examines how enriching an RL agent's observation space with grid-related data can enhance its awareness of system dynamics and improve decision-making. We specifically focus on the agent's performance during unstable periods -- i.e., quarters where last-minute deviations in system imbalance heavily influence prices -- by introducing a related metric. Using the soft actor-critic algorithm, we control a simulated battery energy storage system participating in the Belgian imbalance settlement, leveraging historical data spanning three years. Our findings indicate that, compared to a system-agnostic RL agent (i.e., an agent that does not have grid-related values in the observation space), the system-aware agents develop more effective policies, particularly during unstable quarters.

System-Aware reinforcement learning for optimized implicit imbalance participation in Belgium

F. Pavirani, S.S. Karimi Madahi, B. Claessens and C. Develder

Distill2Explain: Differentiable decision trees for explainable reinforcement learning in energy application controllers

G. Gokhale, S.S. Karimi Madahi, B. Claessens and C. Develder

Demand-side flexibility is gaining importance as a crucial element in the energy transition process. Accounting for about 25% of final energy consumption globally, the residential sector is an important (potential) source of energy flexibility. However, unlocking this flexibility requires developing a control framework that (1) easily scales across different houses, (2) is easy to maintain, and (3) is simple to understand for end-users. A potential control framework for such a task is data-driven control, specifically model-free reinforcement learning (RL). Such RL-based controllers learn a good control policy by interacting with their environment, learning purely based on data and with minimal human intervention. Yet, they lack explainability, which hampers user acceptance. Moreover, limited hardware capabilities of residential assets forms a hurdle (e.g., using deep neural networks). To overcome both those challenges, we propose a novel method to obtain explainable RL policies by using differentiable decision trees. Using a policy distillation approach, we train these differentiable decision trees to mimic standard RL-based controllers, leading to a decision tree-based control policy that is data-driven and easy to explain. As a proof-of-concept, we examine the performance and explainability of our proposed approach in a battery-based home energy management system to reduce energy costs. For this use case, we show that our proposed approach can outperform baseline rule-based policies by about 20-25%, while providing simple, explainable control policies. We further compare these explainable policies with standard RL policies and examine the performance trade-offs associated with this increased explainability.

Distill2Explain: Differentiable decision trees for explainable reinforcement learning in energy application controllers

G. Gokhale, S.S. Karimi Madahi, B. Claessens and C. Develder

Sample efficient reinforcement learning for building control: Leveraging physics-informed latent representations

G. Gokhale, B. Claessens and C. Develder

Given that the residential sector accounts for over 40% of final energy consumption, unlocking its flexibility will be a key step in energy transition. However, accessing this flexibility requires a control framework that (1) easily scales across different, diverse households, and (2) is easy to design, deploy and maintain. While data-driven reinforcement learning based control has emerged as a potential solution for such problems, its widespread commercialization is still limited. A major challenge being the large amount of data required for training such RL-based controllers. To address this problem, our preliminary work investigates the application of physics-informed neural networks for improving the (training) sample efficiency of RL-based controllers. Specifically, we employ physics-informed neural networks to learn low-dimensional, physically relevant representations that can be used with any standard RL algorithm to learn high quality control policies. Using a two-state building simulator, we show that our proposed physics-informed framework can learn high quality control policies (5-10% improvement over business-as-usual controller) using fewer training samples.

Sample efficient reinforcement learning for building control: Leveraging physics-informed latent representations

G. Gokhale, B. Claessens and C. Develder

Explainable reinforcement learning-based home energy management systems using differentiable decision trees

G. Gokhale, B. Claessens and C. Develder

With the ongoing energy transition, demand-side flexibility has become an important aspect of the modern power grid for providing grid support and allowing further integration of sustainable energy sources. Besides traditional sources, the residential sector is another major and largely untapped source of flexibility, driven by the increased adoption of solar PV, home batteries, and EVs. However, unlocking this residential flexibility is challenging as it requires a control framework that can effectively manage household energy consumption, and maintain user comfort while being readily scalable across different, diverse houses. We aim to address this challenging problem and introduce a reinforcement learning-based approach using differentiable decision trees. This approach integrates the scalability of data-driven reinforcement learning with the explainability of (differentiable) decision trees. This leads to a controller that can be easily adapted across different houses and provides a simple control policy that can be explained to end-users, further improving user acceptance. As a proof-of-concept, we analyze our method using a home energy management problem, comparing its performance with commercially available rule-based baseline and standard neural network-based RL controllers. Through this preliminary study, we show that the performance of our proposed method is comparable to standard RL-based controllers, outperforming baseline controllers by  20% in terms of daily cost savings while being straightforward to explain.

Explainable reinforcement learning-based home energy management systems using differentiable decision trees

G. Gokhale, B. Claessens and C. Develder

Control policy correction framework for reinforcement learning-based energy arbitrage strategies

S.S. Karimi Madahi, G. Gokhale, M.-S. Verwee, B. Claessens and C. Develder

A continuous rise in the penetration of renewable energy sources, along with the use of the single imbalance pricing, provides a new opportunity for balance responsible parties to reduce their cost through energy arbitrage in the imbalance settlement mechanism. Model-free reinforcement learning (RL) methods are an appropriate choice for solving the energy arbitrage problem due to their outstanding performance in solving complex stochastic sequential problems. However, RL is rarely deployed in real-world applications since its learned policy does not necessarily guarantee safety during the execution phase. In this paper, we propose a new RL-based control framework for batteries to obtain a safe energy arbitrage strategy in the imbalance settlement mechanism. In our proposed control framework, the agent initially aims to optimize the arbitrage revenue. Subsequently, in the post-processing step, we correct (constrain) the learned policy following a knowledge distillation process based on properties that follow human intuition. Our post-processing step is a generic method and is not restricted to the energy arbitrage domain. We use the Belgian imbalance price of 2023 to evaluate the performance of our proposed framework. Furthermore, we deploy our proposed control framework on a real battery to show its capability in the real world.

Control policy correction framework for reinforcement learning-based energy arbitrage strategies

S.S. Karimi Madahi, G. Gokhale, M.-S. Verwee, B. Claessens and C. Develder

Frequency containment reserve and imbalance participation: A battery-integrated reinforcement learning strategy

F. Pavirani, S.S. Karimi Madahi, B. Claessens and C. Develder

With the increasing integration of renewable energy sources (RES), the electrical grid is facing an amplified uncertainty in the energy supply. Transmission System Operators (TSOs) are offering remuneration in exchange for energy exchanges that reduce system imbalances. Helping stabilize the grid frequency is hence an economically viable endeavor, but it requires strategies that can properly manage stochasticities. To tackle this, we analyze the participation of grid-scale batteries in Frequency Containment Reserve (FCR) using a Reinforcement Learning (RL) control strategy. Acting in a multi-market scenario, the RL agent learns to effectively leverage imbalance prices for a high-quality energy recovery strategy. We trained the agent to maximize the imbalance settlement profit while ensuring conforming participation in the FCR service. The agent is also trained to keep the battery yearly cycles below a planned value. In our simulations, we demonstrated the efficacy of RL when dealing with different FCR participation magnitudes, obtaining an average improvement of +9% in profit compared to a rule-based controller baseline.

Frequency containment reserve and imbalance participation: A battery-integrated reinforcement learning strategy

F. Pavirani, S.S. Karimi Madahi, B. Claessens and C. Develder

HomeLabGym: A real-world testbed for home energy management systems

T. Van Puyvelde, M.-S. Verwee, G. Gokhale, M. Zareh Eshdoust and C. Develder

Amid growing environmental concerns and resulting energy costs, there is a rising need for efficient Home Energy Management Systems (HEMS). Evaluating such innovative HEMS solutions typically relies on simulations that may not model the full complexity of a real-world scenario. On the other hand, real-world testing, while more accurate, is labor-intensive, particularly when dealing with diverse assets, each using a distinct communication protocol or API. Centralizing and synchronizing the control of such a heterogeneous pool of assets thus poses a significant challenge. In this paper, we introduce HomeLabGym, a real-world testbed to ease such real-world evaluations of HEMS and flexible assets control in general, by adhering to the well-known OpenAI Gym paradigm. HomeLabGym allows researchers to prototype, deploy, and analyze HEMS controllers within the controlled test environment of a real-world house (the IDLab HomeLab), providing access to all its available sensors and smart appliances. The easy-to-use Python interface eliminates concerns about intricate communication protocols associated with sensors and appliances, streamlining the evaluation of various control strategies. We present an overview of HomeLabGym, and demonstrate its usefulness to researchers in a comparison between real-world and simulated environments in controlling a residential battery in response to real-time prices.

HomeLabGym: A real-world testbed for home energy management systems

T. Van Puyvelde, M.-S. Verwee, G. Gokhale, M. Zareh Eshdoust and C. Develder

Transfer learning in transformer-based demand forecasting for home energy management system

G. Gokhale, J. Van Gompel, B. Claessens and C. Develder

Increasingly, 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 (i) forecasting accuracy ( 15% MAE reduction) and (ii) control performance ( 2% energy cost reduction), using real-world household data.

Transfer learning in transformer-based demand forecasting for home energy management system

G. Gokhale, J. Van Gompel, B. Claessens and C. Develder

Real-world implementation of reinforcement learning based energy coordination for a cluster of households

G. Gokhale, N. Tiben, M.-S. Verwee, M. Lahariya, B. Claessens and C. Develder

Given its substantial contribution of 40% to global power consumption, the built environment has received increasing attention to serve as a source of flexibility to assist the modern power grid. In that respect, previous research mainly focused on energy management of individual buildings. In contrast, in this paper, we focus on aggregated control of a set of residential buildings, to provide grid supporting services, that eventually should include ancillary services. In particular, we present a real-life pilot study that studies the effectiveness of reinforcement-learning (RL) in coordinating the power consumption of 8 residential buildings to jointly track a target power signal. Our RL approach relies solely on observed data from individual households and does not require any explicit building models or simulators, making it practical to implement and easy to scale. We show the feasibility of our proposed RL-based coordination strategy in a real-world setting. In a 4-week case study, we demonstrate a hierarchical control system, relying on an RL-based ranking system to select which households to activate flex assets from, and a real-time PI control-based power dispatch mechanism to control the selected assets. Our results demonstrate satisfactory power tracking, and the effectiveness of the RL-based ranks which are learnt in a purely data-driven manner.

Real-world implementation of reinforcement learning based energy coordination for a cluster of households

G. Gokhale, N. Tiben, M.-S. Verwee, M. Lahariya, B. Claessens and C. Develder

Graph neural networks for fault diagnosis of geographically nearby photovoltaic systems

J. Van Gompel, D. Spina and C. Develder

Faults in photovoltaic (PV) systems significantly reduce their efficiency and can pose safety risks. Nevertheless, most residential PV systems are not actively monitored, because existing methods often require expensive sensors, which are only cost-effective for large PV systems. Therefore, we propose a graph neural network (GNN) to monitor a group of nearby PV systems without relying on dedicated sensors. Instead, the GNN compares 24 h of current and voltage measurements obtained from the inverters. Four GNN variants are experimentally compared using simulated data of six different PV systems in Colorado. Results show that all GNN variants outperform a state-of-the-art PV fault diagnosis method based on gradient boosted trees. Moreover, some GNN variants can even generalize to PV systems which were not in the training data, enabling monitoring of new PV systems without retraining.

Graph neural networks for fault diagnosis of geographically nearby photovoltaic systems

J. Van Gompel, D. Spina and C. Develder

Temporal convolutional networks for fault diagnosis of photovoltaic systems using satellite and inverter measurements

J. Van Gompel, D. Spina and C. Develder

Over time, photovoltaic (PV) systems become increasingly susceptible to faults. Early fault detection and identification not only limits power losses and increases the systems lifetime, but also prevents more serious consequences, such as risk of fire or electrical shock. Although several accurate fault diagnosis methods have been proposed in literature, most PV systems remain unmonitored as the installations are not equipped with the required sensors. In this work, we propose a fault diagnosis technique that does not require on-site sensors. Rather, weather satellite and inverter measurements are used as inputs for the proposed machine learning model. As no dedicated sensors are needed, our method is widely applicable and cost-effective. A temporal convolutional neural network is developed to accurately identify 6 common types of faults, based on the past 24 h of measurements. The proposed approach is tested extensively on a simulated PV system, taking into account multiple severities of each fault type, and reaches an accuracy of over 86%.

Temporal convolutional networks for fault diagnosis of photovoltaic systems using satellite and inverter measurements

J. Van Gompel, D. Spina and C. Develder

Physics-informed recurrent neural networks for the identification of a generic energy buffer system

M. Lahariya, F. Karami, C. Develder and G. Crevecoeur

Energy storage is ubiquitous in industrial processes and comes in many forms such as material, chemical, electromechanical buffers. System identification of such energy buffers demands proper estimation/prediction of their physical quantities and unknown parameters. Once these parameters are determined, the identified model can be employed to predict the industrial process dynamics, which finally assist to build efficient control for these processes. This paper proposes physics-informed neural networks-based grey-box modeling methods for the identification of energy buffers. The underlying system dynamics are enforced on the neural network structure to ensure that the identified grey-box model follows the approximate physics. We define two novel grey-box models based on simple and recurrent neural network architectures and test these models for a generic energy buffer. Performance and training time for the proposed grey-box models are compared against a black-box baseline model. Results confirm that imposing the dynamic system’s physics on the network improves the performance, and utilizing a recurrent architecture leads to a further improvement.

Physics-informed recurrent neural networks for the identification of a generic energy buffer system

M. Lahariya, F. Karami, C. Develder and G. Crevecoeur

Defining a synthetic data generator for realistic electric vehicle charging sessions

M. Lahariya, D. Benoit and C. Develder

Electric vehicle (EV) charging stations have become prominent in electricity grids in the past years. Analysis of EV charging sessions is useful for flexibility analysis, load balancing, offering incentives to customers, etc. Yet, limited availability of such EV sessions’ data hinders further development in these fields. Addressing this need for publicly available and realistic data, we develop a synthetic data generator (SDG) for EV charging sessions. Our SDG assumes the EV inter-arrival time to follow an exponential distribution. Departure times are modeled by defining a conditional probability density function (pdf) for connection times. This pdf for connection time and required energy is fitted by Gaussian mixture models. Since we train our SDG using a large real-world dataset, its output is realistic.

Defining a synthetic data generator for realistic electric vehicle charging sessions

M. Lahariya, D. Benoit and C. Develder

Reduced state space and cost function in reinforcement learning for demand response control of multiple EV charging stations

M. Lahariya, N. Sadeghianpourhamami and C. Develder

Electric vehicle (EV) charging stations represent a substantial load with signi cant exibility. Balancing such load with model-free demand response (DR) based on reinforcement learning (RL) is an attractive approach. We build on previous RL research using a Markov decision process (MDP) to simultaneously coordinate multiple charging stations. The previously proposed approach is computationally expensive in terms of large training times, limiting its feasibility and practicality. We propose to a priori force the control policy to always ful ll any charging demand that does not o er any exibility at a given point, and thus use an updated cost function. We compare the policy of the newly proposed approach with the original (costly) one, for the case of load attening, in terms of (i) processing time to learn the RL-based charging policy, and (ii) overall performance of the policy decisions in terms of meeting the target load for unseen test data.

Reduced state space and cost function in reinforcement learning for demand response control of multiple EV charging stations

M. Lahariya, N. Sadeghianpourhamami and C. Develder

Techno-economic study of hydrogen production using PV, wind power and battery storage

V. Papadopoulos, J. Knockaert, C. Develder and J. Desmet

Hydrogen is regarded by many scientists as the energy fuel of the future, provided that it is produced by non-polluting renewable energy systems (RES) such as photovoltaics and wind turbines. The majority of studies focusing on hydrogen production with RES have shown that such installations are not yet feasible, at least from an economic perspective, primarily due to significant capital expenditures. Conversely, in this paper, we show that the hydrogen technology can already deliver some interesting investment opportunities. In the present paper, we address a system comprising a 15 MW PV park, wind power, battery storage and a 1 MW PEM electrolyzer. Our study comprises two parts. (i) First, we present the technical analysis of the system. We show how the rate of hydrogen production, here using the term utilization, is affected by the design of the renewable energy system (i.e. wind power capacity, battery size and technology). (ii) Second, we present the economic analysis. In particular, we provide assessments regarding the payback period of the installation. The results of the study let us conclude that (a) to maximize utilization, it is necessary to combine both source and storage components in a hybrid topology, and that (b) in the best scenario, PV co-exists with wind power achieving 65.5% utilization and delivering a payback period in less than 6 years, when hydrogen is sold at at least 6 euro/kg.

Techno-economic study of hydrogen production using PV, wind power and battery storage

V. Papadopoulos, J. Knockaert, C. Develder and J. Desmet

VI-based appliance classification using aggregated power consumption Data

L. De Baets, T. Dhaene, D. Deschrijver, M. Berges and C. Develder

Non-intrusive load monitoring detects active appliances in a household (and their power consumption) from measuring the aggregated power at just one point in that household. Our previous works focused on classifying a single appliance, assuming that the voltage and current trace could be isolated from an aggregated signal by considering the difference in current before and after the event. In this paper, we show that this assumption holds and that it is a viable approach in practice. We experimentally validate this for two classification methods we proposed earlier: (1) random forests using elliptical Fourier descriptors of the appliances’ VI trajectories and (2) convolutional neural networks using the appliances’ VI images. We benchmark these approaches on the aggregated data from the 2018 version of PLAID. We obtain, respectively for each of these classifiers, a maximal Fmacro-measure of 85.31% and 87.95%. We also show that using submetered data for training does not improve the performance.

VI-based appliance classification using aggregated power consumption Data

L. De Baets, T. Dhaene, D. Deschrijver, M. Berges and C. Develder

Modeling real-world flexibility of residential power consumption: Exploring the cylindrical WeiSSVM distribution

N. Sadeghianpourhamami, D.F. Benoit, D. Deschrijver and C. Develder

Flexibility in residential power consumption is a cheap and widely available resource for demand-supply balancing in smart grid. A user’s power consumption flexibility is defined in terms of amount, time and duration of availability. We note that the timing of flexibility is circular in nature: configuration times of real-world observations form clusters which cross over from one day to the next (across the midnight boundary). Therefore, it seems only natural to adopt circular distributions to model this data. In this paper we look into the key research question whether that leads to better generative models than using conventional linear distributions. In particular, we fit Gaussian mixture models (GMMs) and a very flexible recent cylindrical (WeiSSVM) distribution mixture to real-world field trial data. Using a predictive accuracy performance measure,
we find that the latter does not provide substantially better fits. We point out shortcomings of both models and conclude that research for appropriate statistical models for the observed data is still open.

Modeling real-world flexibility of residential power consumption: Exploring the cylindrical WeiSSVM distribution

N. Sadeghianpourhamami, D.F. Benoit, D. Deschrijver and C. Develder

Achieving scalable model-free demand response in charging an electric vehicle fleet with reinforcement learning

N. Sadeghianpourhamami, J. Deleu and C. Develder

To achieve coordinated electric vehicle (EV) charging with demand response (DR), a model-free approach using reinforcement learning (RL) is an attractive proposition. Indeed, it eliminates the need for expert knowledge, and the methodology should generalize from one scenario to the other (e.g., with different EV arrival/departure patterns). Using RL, the DR algorithm is defined as a Markov decision process (MDP). Initial work in this area comprises algorithms to control just one EV at a time, because of scalability challenges when taking coupling between EVs into account. In this paper, we outline our idea on tackling these issues. As a first step, we propose a novel MDP definition for charging an EV fleet. More specifically, we propose (1) a relatively compact aggregate state and action space representation, and (2) a batch RL algorithm (i.e., an instance of fitted Q-iteration, FQI) to learn the optimal EV charging policy. We illustrate the proposed approach in an exemplary scenario.

Achieving scalable model-free demand response in charging an electric vehicle fleet with reinforcement learning

N. Sadeghianpourhamami, J. Deleu and C. Develder

Design and experimental validation of an LTE-based synchrophasor network in a medium voltage distribution grid

A. Derviškadić, P. Romano, C. Ge, W.K. Chai, C. Develder, L. Zanni, M. Pignati and M. Paolone

We present and experimentally validate in a real-scale medium voltage (MV) grid a synchrophasor network that exploits the availability of a public 4G LTE communication infrastructure. An 18 buses, 10kV feeder located in Huissen, The Netherlands, has been equipped with 10 Phasor Measurement Units (PMUs) connected to the MV grid by means of dedicated voltage and current sensors. The PMUs stream synchrophasor data through a public 4G LTE network via an information-centric networking-based middleware, named C-DAX. The measurements are received and time-aligned at a phasor data concentrator and fed to a real-time state estimation application. The paper presents the various field-trial components and validates the feasibility of exploiting the 4G LTE technology for PMU-based applications. Specifically we assess the performance of the adopted wireless telecommunication infrastructure with and without the C-DAX middleware, as well as the accuracy of the real-time state estimation process.

Design and experimental validation of an LTE-based synchrophasor network in a medium voltage distribution grid

A. Derviškadić, P. Romano, C. Ge, W.K. Chai, C. Develder, L. Zanni, M. Pignati and M. Paolone

Handling imbalance in an extended PLAID

L. De Baets, C. Develder, T. Dhaene, D. Deschrijver, J. Gao and M. Berges

The objective of this paper is twofold: (1) introduce an extension of the the Plug-Level Appliance Identification Dataset (PLAID) [1], and (2) discuss how to deal with appliance class imbalance when using this data for NILM performance evaluation and training. Like the original, PLAID 2 is a public dataset for load identification research consisting of short voltage and current measurements (a few seconds) for different appliances found in residential buildings in the United States. Besides increasing the number of appliance instances, this extension includes measurements made during different operating modes for many appliance types. Because of the imbalanced nature of these datasets, alternative metrics to adequately evaluate the performance of classification algorithms are suggested. In particular, the Receiver Operating Characteristics (ROC) curve and its summary statistics lead to more consistent and meaningful results than accuracy or F1-measure. Additionally, the class imbalance can cause the classifier to only focus on the majority classes. The paper addresses methods handling this. However, it is found that these are unnecessary for the PLAID dataset.

Handling imbalance in an extended PLAID

L. De Baets, C. Develder, T. Dhaene, D. Deschrijver, J. Gao and M. Berges

Automated classification of appliances using elliptical fourier descriptors

L. De Baets, C. Develder, D. Deschrijver and T. Dhaene

Non-intrusive load monitoring methods aim to disaggregate the total power consumption of a household into individual appliances by analysing changes in the voltage and current measured at the grid connection point of the household. The goal is to identify the active appliances, based on their unique fingerprint. This information can be communicated to the electricity provider and the end-user enabling the potential of smart grids. An informative characteristic to attain the appliance classification is the voltage-current trajectory. In this paper, this trajectory is represented as a binary image from which the contours are extracted. From these contours, the elliptic Fourier descriptors are calculated and used as input for several classification algorithms outputting the appliance name. Benchmarking this method on the PLAID dataset shows that the descriptors can yield a prediction accuracy up to 79%, comparable to the state-of-the-art, based on only a very compact representation (12 numbers).

Automated classification of appliances using elliptical fourier descriptors

L. De Baets, C. Develder, D. Deschrijver and T. Dhaene

Optimized statistical test for event detection in non-intrusive load monitoring

L. De Baets, J. Ruyssinck, C. Develder, T. Dhaene and D. Deschrijver

Event detection plays an important role in non-intrusive load monitoring to accurately detect when appliances are switched on or off in a residential environment. Besides being accurate, it is important that these methods are robust on real-life power traces. This paper shows that some state-of-the-art event detection methods may miss events when there is a substantial base load caused by active power consuming devices. In order to address this problem, this paper extends the existing chi-squared goodness-of-fit test with a a voting scheme. Furthermore, a workflow is proposed using surrogate-based optimisation for tuning the parameters of these tests in an efficient way. Results on the BLUED dataset indicate that the novel voting chi-squared GOF method outperforms the standard chi-squared GOF test when applied to traces with a higher base load.

Optimized statistical test for event detection in non-intrusive load monitoring

L. De Baets, J. Ruyssinck, C. Develder, T. Dhaene and D. Deschrijver

Quantifying flexibility in EV charging as DR potential: Analysis of two real-world data sets

C. Develder, N. Sadeghianpourhamami, M. Strobbe and N. Refa

The increasing adoption of electric vehicles (EVs) presents both challenges and opportunities for the power grid, especially for distribution system operators (DSOs). The demand represented by EVs can be significant, but on the other hand, sojourn times of EVs could be longer than the time required to charge their batteries to the desired level (e.g., to cover the next trip). The latter observation means that the electrical load from EVs is characterized by a certain level of flexibility, which could be exploited for example in demand response (DR) approaches (e.g., to balance generation from renewable energy sources).
This paper analyzes two data sets, one from a charging-at- home field trial in Flanders (about 8.5k charging sessions) and another from a large-scale EV public charging pole deployment in The Netherlands (more than 1M sessions). We rigorously analyze the collected data and quantify aforementioned flexibility: (1) we characterize the EV charging behavior by clustering the arrival and departure time combinations, identifying three behaviors (charging near home, charging near work, and park to charge), (2) we fit statistical models for the sojourn time, and flexibility (i.e., non-charging idle time) for each type of observed behavior, and (3) quantify the the potential of DR exploitation as the maximal load that could be achieved by coordinating EV charging for a given time of day t, continuously until t + Delta.

Quantifying flexibility in EV charging as DR potential: Analysis of two real-world data sets

C. Develder, N. Sadeghianpourhamami, M. Strobbe and N. Refa

Energy flexibility assessment of an industrial coldstore process

J. van der Herten, F. Depuydt, D. Deschrijver, M. Strobbe, C. Develder, T. Dhaene, R. Bruneliere and J.-W. Rombouts

Power-intensive industry plays a key role in balancing supply and demand in the energy grid: by offering flexible power, industry provides competitive alternatives to gas fired power plants and reduces the CO2 impact of grid balancing. Furthermore, operating costs can be reduced and grid operators can avoid technical failures. Recently, research has started to try and address the challenging question of determining the amount of power curtailment (i.e., how much power can be reduced for how long) without violating any process constraints. We consider several machine learning methods to assess the curtailment potential in a coldstore, based on historical data.

Energy flexibility assessment of an industrial coldstore process

J. van der Herten, F. Depuydt, D. Deschrijver, M. Strobbe, C. Develder, T. Dhaene, R. Bruneliere and J.-W. Rombouts

Real-world user flexibility of energy consumption: Two-stage generative model construction

N. Sadeghianpourhamami, M. Strobbe and C. Develder

Since the inception of smart grids, a substantial amount of research has focused on the development of scalable Demand Response (DR) approaches. For example, to flatten peak load, or to balance renewable energy production. A crucial assumption in DR is that at least some portion of the load is flexible, i.e., can be shifted in time. While the flexibility potential of smart devices has been analyzed extensively based on the device characteristics, little effort has been devoted to establishing potential factors in their owner’s behavior. In this paper, we focus on sharpening the analysis of flexibility in residential user load and contribute with: (1) a quantitative specification of such flexibility, (2) a systematic methodology to derive a generative model for user flexibility behavior from data, (3) application of the methodology on a real world data set from a field trial with smart appliances, and (4) analysis of factors determining that flexibility.

Real-world user flexibility of energy consumption: Two-stage generative model construction

N. Sadeghianpourhamami, M. Strobbe and C. Develder

Enabling Smart Grid Applications with ICN

W.K. Chai, K.V. Katsaros, M. Strobbe, P. Romano, C. Ge, C. Develder, G. Pavlou and N. Wang

We have harnessed the salient features of information-centric networking (ICN) and implemented a communication infrastructure, called C-DAX, for supporting smart grid applications. We will demonstrate the operations of C-DAX both in a laboratory setup and a real field trial that involves the deployment of C-DAX in a live electricity grid in the Netherlands. This demo will showcase the capabilities of C-DAX, highlighting how ICN satisfies stringent smart grid application requirements.

Enabling Smart Grid Applications with ICN

W.K. Chai, K.V. Katsaros, M. Strobbe, P. Romano, C. Ge, C. Develder, G. Pavlou and N. Wang

Evaluation of Belgian energy market models with demand response

G. Casier, J. Van Ooteghem, M. Strobbe, S. Verbrugge and C. Develder

Energy markets worldwide are evolving towards smart grids. The value network of the current energy market will be subject to changing actors and roles in the evolution towards smart grids. There are several ways in which the market could respond to these changes, depending on who will take up newly emerging responsibilities. We consider the architecture of a large-scale pilot that was conducted within the Flemish Linear project and model the current energy market. Drawing upon the insights derived from the pilot and the current models, we propose several possible future market models that could take place when smart meters will be implemented and opportunities for demand response will emerge. Subsequently, the different models are evaluated by means of a PEST analysis.

Evaluation of Belgian energy market models with demand response

G. Casier, J. Van Ooteghem, M. Strobbe, S. Verbrugge and C. Develder

Deploying the ICT architecture of a residential demand response pilot

M. Strobbe, K. Vanthournout, T. Verschueren, W. Cardinaels and C. Develder

The Flemish project Linear was a large scale res- idential demand response pilot that aims to validate innovative smart grid technology building on the rollout of information and communication technologies in the power grid.
For this pilot a scalable, reliable and interoperable ICT infrastructure was set up, interconnecting 245 residential power grid customers with the backend systems of energy service providers (ESPs), flexibility aggregators, distribution system operators (DSOs) and balancing responsible parties (BRPs).
On top of this architecture several business cases were rolled out, which require the sharing of metering data and flexibility information, and demand response algorithms for the balancing of renewable energy and the mitigation of voltage and power issues in distribution grids.
The goal of the pilot is the assessment of the technical and economical feasibility of residential demand response in real life, and of the interaction with the end-consumer.
In this paper we focus on the practical experiences and lessons learnt during the deployment of the ICT technology for the pilot. This includes the real-time gathering of measurement data and real-time control of a wide range of smart appliances in the homes of the participants. We identified a number of critical issues that need to be addressed for a future full-scalle roll- out: (i) reliable in-house communication, (ii) interoperability of appliances, measurement equipment, backend systems, and business cases, and (iii) sufficient backend processing power for real-time analysis and control.

Deploying the ICT architecture of a residential demand response pilot

M. Strobbe, K. Vanthournout, T. Verschueren, W. Cardinaels and C. Develder

Large-scale residential demand response pilot ICT architecture

M. Strobbe, K. Vanthournout, T. Verschueren, W. Cardinaels and C. Develder

The Flemish project Linear is an example of an ongoing large scale residential demand response pilot that aims to validate innovative smart grid applications that exploit the rollout of information and communication technologies (ICT) in the power grid. In this paper, we discuss the design of such a scalable, reliable and interoperable ICT infrastructure that interconnects 245 residential power grid customers with the backend systems of various actors: e.g., energy service providers (ESPs), flexibility aggregators, distribution system operators (DSOs), balancing re- sponsible parties (BRPs). The use cases rolled out in Linear, built on top of our proposed ICT architecture, involve sharing both metering data and flexibility information (esp. for time shifting) of the households, and demand response (DR) algorithms for the balancing of renewable energy and the mitigation of voltage and power issues in distribution grids.

Large-scale residential demand response pilot ICT architecture

M. Strobbe, K. Vanthournout, T. Verschueren, W. Cardinaels and C. Develder

Economic evaluation of active network management alternatives for congestion avoidance - the DSO perspective

L. Van Halewyck, J. Verstraeten, M. Strobbe and C. Develder

The introduction of distributed renewable energy generators in the electricity grid implies a number of challenges for energy suppliers and utilities. One of these challenges is the increased risk for grid congestion issues due to the installation of large amounts of e.g. solar panels and wind turbines to the existing grid. As alternative to costly grid reinforcements several Active Network Management (ANM) techniques to mitigate this risk are researched, including (1) dynamic line rating and (2) demand side management. In this paper, we make an economic evaluation of those options from a Distribution System Operator (DSO) perspective. We discuss the aforementioned techniques and elaborate their business case for a specific use case in the MV-grid. Our calculations show that dynamic line rating can be considered as an alternative to network reinforcement, depending on the regulatory framework.

Economic evaluation of active network management alternatives for congestion avoidance - the DSO perspective

L. Van Halewyck, J. Verstraeten, M. Strobbe and C. Develder

Exploiting road traffic data for very short term load forecasting in smart grids

J. Aparicio, J. Rosca, M. Mediger, A. Essl, K. Arzig and C. Develder

If accurate short term prediction of electricity consumption is available, the Smart Grid infrastructure can rapidly and reliably react to changing conditions. The economic importance of accurate predictions justifies research for more complex forecasting algorithms. This paper proposes road traffic data as a new input dimension that can help improve very short term load forecasting. We explore the dependencies between power demand and road traffic data and evaluate the predictive power of the added dimension compared with other common features, such as historical load and temperature profiles.

Exploiting road traffic data for very short term load forecasting in smart grids

J. Aparicio, J. Rosca, M. Mediger, A. Essl, K. Arzig and C. Develder

Design of a management infrastructure for smart grid pilot data processing and analysis

M. Strobbe, T. Verschueren, S. Melis, D. Verslype, K. Mets, F. De Turck and C. Develder

Future smart grids will combine power grid tech- nologies with information and communication technologies to enable a more efficient, reliable and sustainable energy produc- tion and distribution. To realize such a smart grid, large scale pilot projects are currently implemented and evaluated. Such pilot projects generate an excessive amount of data that needs to be processed: energy measurements, information on available flexibility from smart devices that can be shifted in time, control signals, dynamic prices, environmental data, etc. To validate and analyze the gathered data and adjust the running experiments in real-time, an optimized data management infrastructure is needed as well as comprehensive visualization tools.
In this paper we present a data management infrastructure optimized for the follow up of the large scale smart grid project called Linear. In this project a pilot in over 200 households is implemented to evaluate several business cases including the balancing of renewable energy supply and the mitigation of voltage and power issues in distribution grids. By decoupling the gathering of the incoming data, the processing and storage of the data, and the data visualization and analysis on different servers, each with their own, optimized database, we obtain an efficient system for validation of the generated data in the pilot, management of the deployed set-ups and follow-up of the ongoing experiments in real-time.

Design of a management infrastructure for smart grid pilot data processing and analysis

M. Strobbe, T. Verschueren, S. Melis, D. Verslype, K. Mets, F. De Turck and C. Develder

Distributed smart charging of electric vehicles for balancing wind energy

K. Mets, F. De Turck and C. Develder

To meet worldwide goals of reducing CO2 footprint, electricity production increasingly is stemming from so-called renewable sources. To cater for their volatile behavior, so-called demand response algorithms are required. In this paper, we focus particularly on how charging electrical vehicles (EV) can be coordinated to maximize green energy consumption. We present a distributed algorithm that minimizes imbalance costs, and the disutility experienced by consumers. Our approach is very much practical, as it respects privacy, while still obtaining near-optimal solutions, by limiting the information exchanged: i.e. consumers do not share their preferences, deadlines, etc. Coordination is achieved through the exchange of virtual prices associated with energy consumption at certain times.
We evaluate our approach in a case study comprising 100 electric vehicles over the course of 4 weeks, where renewable energy is supplied by a small scale wind turbine. Simulation results show that 68% of energy demand can be supplied by wind energy using our distributed algorithm, compared to 73% in a
theoretical optimum scenario, and only 40% in an uncoordinated business-as-usual (BAU) scenario. Also, the increased usage of renewable energy sources, i.e. wind power, results in a 45% reduction of CO2 emissions, using our distributed algorithm.

Distributed smart charging of electric vehicles for balancing wind energy

K. Mets, F. De Turck and C. Develder

Infrastructure for collaborating data-researchers in a smart grid pilot

W. Labeeuw, S. Claessens, K. Mets, C. Develder and G. Deconinck

A large amount of stakeholders are often involved in Smart Grid projects. Each partner has its own way of storing, representing and accessing its data. An integrated data storage and a joint online analytical mining infrastructure is needed to limit the amount of duplicated work and to raise the overall security of the system. The proposed infrastructure is composed of standard application software and an in-house developed data analysis tool that allows researchers to add and share their own functionality without compromising security.

Infrastructure for collaborating data-researchers in a smart grid pilot

W. Labeeuw, S. Claessens, K. Mets, C. Develder and G. Deconinck

Design and evaluation of an architecture for future smart grid service provisioning

M. Strobbe, T. Verschueren, K. Mets, S. Melis, C. Develder, F. De Turck, T. Pollet and S. Van de Veire

The current Internet evolves to a true Internet of Things where a plethora of heterogeneous devices are connected, communicate with each other and allow the deployment of new intelligent services, creating new ecosystems in diverse domains. One such a domain is the power grid. The increase of distributed renewable electricity generation, e.g. solar cells and wind turbines, requires new energy management systems where real-time measurements and communication between end users, suppliers and utilities are vital. To address this need, we propose a common service architecture that allows houses with renewable energy generation and smart energy devices to plug into a distributed energy management system, integrated with the public power grid. The presented architecture facilitates end-users to optimize their energy consumption, enables power network operators to better balance supply and demand, and creates a platform where new market players (e.g. ESCOs) can easily provide new services. This service architecture has been implemented and is currently evaluated in a field trial with 21 users, of which we present the initial results.

Design and evaluation of an architecture for future smart grid service provisioning

M. Strobbe, T. Verschueren, K. Mets, S. Melis, C. Develder, F. De Turck, T. Pollet and S. Van de Veire

Distributed multi-agent algorithm for residential energy management in smart grids

K. Mets, M. Strobbe, T. Verschueren, T. Roelens, C. Develder and F. De Turck

Distributed renewable power generators, such as solar cells and wind turbines are difficult to predict, making the demand-supply problem more complex than in the traditional energy production scenario. They also introduce bidirectional energy flows in the low-voltage power grid, possibly causing voltage violations and grid instabilities. In this article we describe a distributed algorithm for residential energy management in smart power grids. This algorithm consists of a market-oriented multi-agent system using virtual energy prices, levels of renewable energy in the real-time production mix, and historical price information, to achieve a shifting of loads to periods with a high production of renewable energy. Evaluations in our smart grid simulator for three scenarios show that the designed algorithm is capable of improving the self consumption of renewable energy in a residential area and reducing the average and peak loads for externally supplied power.

Distributed multi-agent algorithm for residential energy management in smart grids

K. Mets, M. Strobbe, T. Verschueren, T. Roelens, C. Develder and F. De Turck

Smart and secure charging of electric vehicles in public parking spaces

M. Strobbe, K. Mets, M. Tahon, M. Tilman, F. Spiessens, J. Gheerardyn, K. De Craemer, S. Vandael, K. Geebelen, B. Lagaisse, B. Claessens and C. Develder

Governments worldwide are starting to give incentives to promote the use of (hybrid) electrical vehicles to achieve cleaner and more energy-efficient road transport with a low carbon footprint. Through tax/VAT reductions and free additional services — such as free parking, and/or battery charging or lower traffic congestion taxes — private users, public organizations and car fleet operators are stimulated to adopt the plug-in (hybrid) electrical vehicle (PHEV). This upcoming breakthrough of PHEVs will impose various challenges to the power grid, such as a significant increase of the load in residential areas and parking spaces as the limited range requires frequent charging. Another challenge is the seamless driver and car identification and fraud-sensitive measuring of the charged energy anywhere a car is charged. On the other hand new opportunities are created as these vehicles allow the storage of power on a wide scale, and typically offer some flexibility during the charging process which can be exploited by smart charging services to balance demand and supply or maximize the local consumption of renewable energy. In this paper we discuss the main actors involved in the charging of leased cars in public parking spaces. We present a novel service architecture and detail the security aspects. Furthermore we discuss how a market-based smart charging algorithm that is plugged into the architecture can exploit the flexibility of the parked cars to maximize the consumption of local generated wind energy. We conclude this paper with the presentation of a new business model for the smart charging of leased cars in public parking spaces, detailing the different actors and value flows.

Smart and secure charging of electric vehicles in public parking spaces

M. Strobbe, K. Mets, M. Tahon, M. Tilman, F. Spiessens, J. Gheerardyn, K. De Craemer, S. Vandael, K. Geebelen, B. Lagaisse, B. Claessens and C. Develder

Exploiting V2G to optimize residential energy consumption with electrical vehicle (dis)charging

K. Mets, T. Verschueren, F. De Turck and C. Develder

The potential breakthrough of pluggable (hybrid) electrical vehicles (PHEVs) will impose various challenges to the power grid, and esp. implies a significant increase of its load. Adequately dealing with such PHEVs is one of the challenges and opportunities for smart grids. In particular, intelligent control strategies for the charging process can significantly alleviate peak load increases that are to be expected from e.g. residential vehicle charging at home. In addition, the car batteries connected to the grid can also be exploited to deliver grid services, and in particular give stored energy back to the grid to help coping with peak demands stemming from e.g. household appliances. In this paper, we will address such so-called vehicle-to-grid (V2G) scenarios while considering the optimization of PHEV charging in a residential scenario. In particular, we will assess the optimal car battery (dis)charging scheduling to achieve peak shaving and reduction of the variability (over time) of the load of households connected to a local distribution grid. We compare (i) a business-as-usual (BAU) scenario, without any intelligent charging, (ii) intelligent local charging optimization without V2G, and (iii) charging optimization with V2G. To evaluate these scenarios, we make use of our simulation tool, based on OMNeT++, which combines ICT and power network models and incorporates a Matlab model that allows e.g. assessing voltage violations. In a case study on a three- feeder distribution network spanning 63 households, we observe that non-V2G optimized charging can reduce the peak demand compared to BAU with 64 If we apply V2G to the intelligent charging, we can further cut the non-V2G peak demand with 17% (i.e., achieve a peak load which is only 30% of BAU).

Exploiting V2G to optimize residential energy consumption with electrical vehicle (dis)charging

K. Mets, T. Verschueren, F. De Turck and C. Develder

Assessment and mitigation of voltage violations by solar panels in a residential distribution grid

T. Verschueren, K. Mets, B. Meersman, M. Strobbe, C. Develder and L. Vandevelde

Distributed renewable electricity generators, such as solar cells and wind turbines introduce bidirectional energy flows in the low-voltage power grid, possibly causing voltage violations and grid instabilities. The current solution to this problem comprises automatically switching off some of the local generators resulting in a loss of green energy. In this paper we study the impact of different solar panel penetration levels in an residential area and the corresponding effects on the distribution feeder line. To mitigate these problems, we assess how effective it is to locally store excess energy in batteries. A case study on a residential feeder serving 63 houses shows that if 80% of them have photo-voltaic (PV) panels, 45% of them will be switched off, resulting in 482 kWh of PV-generated energy being lost. We show that providing a 9 kWh battery at each house can mitigate some voltage violations, and therefor allowing for more renewable energy to be used.

Assessment and mitigation of voltage violations by solar panels in a residential distribution grid

T. Verschueren, K. Mets, B. Meersman, M. Strobbe, C. Develder and L. Vandevelde

Integrated simulation of power and communication networks for smart grid applications

K. Mets, T. Verschueren, C. Develder, T. Vandoorn and L. Vandevelde

Innovative architectures, control mechanisms and network technologies are being proposed to realize the future smart grid. To assess their impact and effectiveness, simulation is key. Simulation in both areas of communication networks as well as power systems has been widely adopted. However, the coupling of those two worlds calls for tools able to address both. In this paper, we propose an innovative integrated framework that models and simulates both the communication network and power networks. We discuss the design and operation of the simulation environment, and illustrate this by means of a case study that employs it.

Integrated simulation of power and communication networks for smart grid applications

K. Mets, T. Verschueren, C. Develder, T. Vandoorn and L. Vandevelde

Evaluation of multiple design options for smart charging algorithms

K. Mets, T. Verschueren, F. De Turck and C. Develder

We evaluate the impact of limitations that may exist in actual implementations of plugin electrical vehicle chargers (e.g. no controllable charging current) on smart charging algorithms. We use quadratic programming to control and coordinate the charging of multiple vehicles in order to reduce the peak load and load profile variability observed by a distribution grid transformer. Simulation results are presented for a section of a residential distribution grid comprising 150 households.

Evaluation of multiple design options for smart charging algorithms

K. Mets, T. Verschueren, F. De Turck and C. Develder

Smart Grids and the role of ICT

C. Develder, W. Haerick, K. Mets and F. De Turck

Smart Grids and the role of ICT

C. Develder, W. Haerick, K. Mets and F. De Turck

A simulator for the control network of smart grid architectures

K. Mets, W. Haerick and C. Develder

We present an extensible simulation framework that has been developed to facilitate research on smart data exchange in power grids. Smart grids integrate traditional power grid technologies and information and communication technologies to generate, transport, distribute and consume energy in a more efficient manner. The framework not only incorporates the main power grid characteristics, but also explicitly models the communication network and control entities. Thus we enable the study of realistic implementation approaches to coordination mechanisms for smart grid applications. These coordination mechanisms are required for example to balance demand and supply, especially considering energy originating from renewable energy sources. We illustrate the design of the simulator with a use case in which the integration of electric vehicles in the distribution grid is managed.

A simulator for the control network of smart grid architectures

K. Mets, W. Haerick and C. Develder

Optimizing smart energy control strategies for plug-in hybrid electric vehicle charging

K. Mets, T. Verschueren, W. Haerick, C. Develder and F. De Turck

The electrification of the vehicle fleet will result in an additional load on the power grid. Adequately dealing with such pluggable (hybrid) electrical vehicles (PHEV) forms part of the challenges and opportunities in the evolution towards Smart Grids. In this paper, we investigate the potential benefits of using control mechanisms, that could be offered by a Home Energy control box, in optimizing energy consumption stemming from PHEV charging in a residential use case. We present smart energy control strategies based on quadratic programming for charging PHEVs, aiming to minimize the peak load and flatten the overall load profile. We compare two strategies, and benchmark them against a business-as-usual scenario assuming full charging starting upon plugging in the PHEV. The first, local strategy only uses information at the home where the PHEV is charged: as a result the charging is optimized for local loads. The local strategy is compared to a global iterative strategy which controls the charging of multiple vehicles based on global load information over a residential area. Both strategies control the duration and rate of charging and result in charging schedules for each vehicle. We present quantitative simulation results over a set of 150 homes, and discuss the strategies in terms of complexity and performance (esp. resulting energy consumption), as well as their requirements concerning infrastructure and communication.

Optimizing smart energy control strategies for plug-in hybrid electric vehicle charging

K. Mets, T. Verschueren, W. Haerick, C. Develder and F. De Turck

Architectures for smart end-user services in the power grid

T. Verschueren, K. Mets, W. Haerick, C. Develder, F. De Turck and T. Pollet

The increase of distributed renewable electricity generators, such as solar cells and wind turbines, requires a new energy management system. These distributed generators introduce bidirectional energy flows in the low-voltage power grid, requiring novel coordination mechanisms to balance local supply and demand. Closed solutions exist for energy management on the level of individual homes. However, no service architectures have been defined that allow the growing number of end-users to interact with the other power consumers and generators and to get involved in more rational energy consumption patterns using intuitive applications. We therefore present a common service architecture that allows houses with renewable energy generation and smart energy devices to plug into a distributed energy management system, integrated with the public power grid. Next to the technical details, we focus on the usability aspects of the end-user applications in order to contribute to high service adoption and optimal user involvement. The presented architecture facilitates end-users to reduce net energy consumption, enables power grid providers to better balance supply and demand, and allows new actors to join with new services. We present a novel simulator that allows to evaluate both the power grid and data communication aspects, and illustrate a 22% reduction of the peak load by deploying a central coordinator inside the home gateway of an end-user.

Architectures for smart end-user services in the power grid

T. Verschueren, K. Mets, W. Haerick, C. Develder, F. De Turck and T. Pollet

LINEAR: towards a breakthrough of smart grids in Flanders

E. Peeters, C. Develder, J. Das, J. Driesen and R. Belmans

Smart grids refer to electricity networks that enable a more efficient (both economical and energetic), reliable and sustainable energy production and distribution. Such networks integrate innovative tools, technologies, products and services throughout the value chain — starting from production through transmission, distribution and supply completely to the devices and installations of the consumers — by the use of monitoring, communication and control technologies. Smart grids support bidirectional real-time exchange of both energy and information. As a result the end users have access to more accurate and timely information regarding their energy consumption and to several options regarding different tariff structures, which enables demand side integration and an improvement of the energy-efficiency. To implement the new structure for sustainable energy supply on a large scale in Flanders by 2020 (and beyond), a transition is necessary with short term action points, that are however based on a mid and long term strategy. The “breakthrough” project Linear (Local Intelligent Networks and Energy Active Regions) is a first crucial step in this transition towards Smart Grids. The project focuses on the realization of a technological and implementation breakthrough by innovative technological research and a large scale pilot in a residential area. All this in close collaboration with industrial partners and associated Flemish innovation platforms. The project has a budget of 9,5 M€ for the research institutes (ESAT/Electa-KULeuven, IBBT, VITO and IMEC) during a period of 5 years starting from may 2009. Besides this, the industrial partners invest a budget of more than 32 M€. This paper presents an overview of the project, with a focus on its specific short and long term goals and objectives.

LINEAR: towards a breakthrough of smart grids in Flanders

E. Peeters, C. Develder, J. Das, J. Driesen and R. Belmans