ABSTRACT. In this paper we propose a droop-free distributed secondary control with admissible voltage profile guarantees for DC microgrids. The proposed distributed control includes an average voltage regulator, a voltage variance regulator, and a relaxed current sharing regulator. The voltage regulator ensures global average voltage of the distributed generators (DGs) to be the rated voltage and the voltage variance regulator regulates the global variance of the DG voltage magnitudes to a predetermined reference. In order to achieve the objectives of voltage regulation, the current sharing from one of the DGs which may be owned by the microgrid community is relaxed. The global dynamic model of the DC microgrid with the proposed control is derived. Besides, steady-state analysis is performed to show that all objectives can be achieved. Finally, the effectiveness of the proposed control is validated through simulations on a 4-DG DC microgrid.

ABSTRACT. Electrical distribution system is transforming due to increased penetration of distributed generation and consumer connected storage units. With the introduction of favorable regulatory conditions, battery owners are expected to participate in electricity markets through aggregators to deliver various grid services. One such service is electricity generation/consumption offered through whole-sale day-ahead electricity markets. When offering this service, an aggregator optimizes the charging and discharging schedule of the batteries based on the market price forecasts to maximize the profits. Any uncertainty or error in the price forecasts will adversely impact the aggregator’s profit. Researchers have proposed two different ways to reduce the impact of market price uncertainty on the profit: (a) Using the Conditional Value-at-Risk, and (b) Using Robust Optimization techniques. However, there is no comparative study between these two approaches in the literature to determine which is better. Further, it remains to be seen if there are other measures, apart from the forecast error, that can impact an aggregator’s profitability. In this paper, we focus on answering these two specific questions using real world logs obtained from the day-ahead market of an European Power exchange.

ABSTRACT. The energy transition confronts utilities with major challenges due to the decentralised and volatile feed-in of renewable energies. As part of the Cloud|E research project, an intelligent cloud-based energy management system (EMS) is being developed in which flexibility options are made available to the utilities by connecting components for power generation, storage and large loads that are found in private households. The potential of this system is analysed in this paper using a simulation scenario. The individual components for generating flexibilities are considered separately. The flexibility potential is presented in the form of additional revenue opportunities on the part of the utilities. It is shown that each flexibly controlled component leads to an additional profit for the utility. By varying the framework conditions in a wide-ranging scenario analysis, the robustness of the system is proven. The results demonstrate the possibility of flexibility generation in the area of electromobility. In total, the potential of a decentralised cloud-based EMS to support the utilities in times of the energy transition can be confirmed.

ABSTRACT. This paper investigates if and how households changing their electricity demand when faced with variable power prices and how this may affect peak demand in the transmission grid. A pricing experiment was set up with artificial variable prices for electricity and real economic consequences for the participating households. The experiment was conducted in three Norwegian cities with different price signals up to a price level of 10 NOK/kWh. In average, the households responded by reducing their electricity consumption with 7.2 % in hours with high prices. In addition, the results showed that the magnitude of the price signal had only a minor effect on the response. These findings suggest that households are price sensitive, but a possible contribution to a lower peak demand is still uncertain due to limited data from the experiment.

ABSTRACT. Renewable Energy Sources (RES) are ever more finding their way into today’s power systems. They have many benefits, but hold challenges when it comes to operating the power system. Their intermittency defies the current prominent approach of keeping system balance; supply-follows-demand. Demand Response (DR) has been proposed numerous times to mitigate the challenges of RES. Additionally, Reinforcement Learning (RL) has been proposed numerous times to mitigate the scalability challenges of DR. Simultaneously, RL has been criticised for its data inefficiency. In this paper, we showcase a transfer learning approach, in a DR setting. The application consists of a household, equipped with inflexible load, a rooftop solar installation, an Electric Water Heater (EWH) and a battery. The household’s electrical energy consumption is billed according to a Time-of-Use-pricing scheme. In this multi-agent DR application, two RL-agents, one for the EWH and one for the battery, have to efficiently operate the system. We show that RL, combined with pre-training of the battery-agent, reduces operation cost by 28.5 %, compared to rule-based control.

ABSTRACT. The flexibility of residential customers can be aggregated and offered in the electricity market. Energy suppliers may exploit this opportunity and participate in the market as aggregators to increase their profit. To accomplish this, the aggregators should consider the customer behaviour in response to incentive-based DR programs to develop their bidding strategy. Moreover, the aggregators should ensure the fulfilment of network operational limits while trading in the real-time market. To address these challenges, this paper presents a decentralised network-constrained bidding strategy model for price-taker aggregators. To model the network constraints, the aggregators use the estimations of the network variables and their sensitivities to the demand variation at different buses to characterize the network variables. The proposed methodology is tested on the IEEE 33-bus distribution test system using real electricity market data. It is observed that the aggregator can simultaneously manage the DR for profit maximization while considering local network constraints; consequently, ensuring secure and reliable network operation without referencing to the network operator.

ABSTRACT. The deployment and operation of advanced monitoring infrastructure in non-residential buildings have accommodated the acquisition of datasets in the form of long multivariate time series pertained to energy use. These datasets may be used to extract information in order to characterize and improve the building’s operation by establishing energy consumption patterns, identifying abnormal operating modes, and detecting and anticipating faults ahead of time. The current research focuses on a two-fold data analysis method applied to academic building energy data. Initially, Matrix Profile (MP) is used for feature extraction - i.e., MP is a time series data mining technique - to extract and build feature datasets of anomaly patterns from academic building energy traces. Second, the extracted feature datasets are fed to two supervised machine learning classification models aiming at discriminating the type of building load pattern. The case study is carried out on a set of public real world energy load consumption patterns of four different types of academic buildings. The method can prove useful for exploiting complementary load patterns in a control infrastructure attaining efficient building energy management and reliable operation.

ABSTRACT. Recent policies have led to the development and deployment of renewable energy sources, introducing new operational challenges. Among those, is the problem of extended downtime of RES, such as wind turbines. In this context, it is important to minimize the downtime of renewable assets by an optimal maintenance strategy via early fault detection. As a Supervisory Control and Data Acquisition (SCADA) system is an integrated part of any production facility and collects large amounts of data, machine learning techniques can be used to detect the underlying failure patterns and notify customers of the abnormal behaviour. Thus, in this work, a novel framework based on machine learning algorithms for fault prediction of wind farm generators is presented for an actual customer. The proposed fault prognosis methodology is tested and validated using historical data from a wind farm in Summerside, Prince Edward Island, Canada, and models are evaluated based on appropriate metrics. The results demonstrate the ability of the proposed methodology to predict wind generator failures, and the viability of the proposed methodology for optimizing predictive maintenance strategies.

ABSTRACT. Optimal power flow (OPF) is the fundamental mathematical model in power system operations. Improving the solution quality of OPF provide huge economic and engineering benefits. The convex reformulation of the original nonconvex alternating current OPF (ACOPF) model gives an efficient way to find the global optimal solution of ACOPF but suffers from the relaxation gaps. The existence of relaxation gaps hinders the practical application of convex OPF due to the AC-infeasibility problem. We evaluate and improve the tightness of the convex ACOPF model in this paper. Various power networks and nodal loads are considered in the evaluation. A unified evaluation framework is implemented in Julia programming language. This evaluation shows the sensitivity of the relaxation gap and helps to benchmark the proposed tightness reinforcement approach (TRA). The proposed TRA is based on the penalty function method which penalizes the power loss relaxation in the objective function of the convex ACOPF model. A heuristic penalty algorithm is proposed to find the proper penalty parameter of the TRA. Numerical results show relaxation gaps exist in test cases especially for large-scale power networks under low nodal power loads. TRA is effective to reduce the relaxation gap of the convex ACOPF model.

ABSTRACT. Various diagnostic systems based on artificial intelligence or machine learning algorithms are already being used today to monitor electrical equipment in power supply systems. The challenge of these data-based diagnostic approaches lies in dealing with the limited fault-condition data available. One possible solution to this problem are data augmentation techniques that generate synthetic data from existing data. In this paper, we develop a model-based data augmentation approach that uses physical models to generate synthetic data. This approach uses statistical information extracted from the available data to draw model parameters and generate synthetic normal- and fault-condition data. It is shown for vibration measurements of a power and distribution transformer that the proposed model-based data augmentation can generate realistic synthetic normal- and fault-condition data.

ABSTRACT. Classically, synthetic data from a data augmentation is used to augment measurement datasets and use them for training of machine learning (ML) algorithms. However, the synthetic data often does not represent the measurements perfectly. In this paper, ML models are trained using only synthetic data. These ML models are then transferred to the available measurement data utilizing transfer learning. This approach is showcased for the detection of a power and distribution transformer fault and benchmarked with state-of-the-art diagnostic systems. The performance of all diagnostic systems is analyzed by limiting the amount of fault-condition measurements available for the training process and by comparison of learning curves. It is shown that the model-based data augmentation combined with fine tuning is capable of improving the accuracy for the analyzed diagnostic task.

ABSTRACT. This paper discusses linearized models of hydropower plants (HPPs). First, it reviews state-of-the-art models and discusses their non-linearities, then it proposes suitable linearization strategies for the plant head, discharge, and turbine torque. It is shown that neglecting the dependency of the hydroacoustic resistance on the discharge leads to a linear formulation of the hydraulic circuits model. For the turbine, a numerical linearization based on a first-order Taylor expansion is proposed. Model performance is evaluated for a medium- and a low-head HPP with a Francis and Kaplan turbine, respectively. Perspective applications of these linear models are in the context of efficient model predictive control of HPPs based on convex optimization.

ABSTRACT. Communication networks are providing enhanced control and management strategies to the smart grids via a set of control center units. However, the interdependency between communication networks and power grids increases the risk of large-scale joint cascading failures. In this paper, the impact of the near-optimal allocation of control centers on the vulnerability of power-communication networks against joint cascading failure attacks is studied. A joint attack strategy is proposed based on the most influential nodes in power grids and communication networks. Then, the number and locations of control centers in such interdependent system are optimized to reduce the system's vulnerability. Results reveal that the allocation of control centers can reduce system vulnerability against joint attacks, however, the damaging impact remains severe.

ABSTRACT. Accurate fault classification in transmission network has been a critical problem with the development of smart grid. Recently, deep learning based methods have shown efficiency to handle this problem. However, the dependency on labeled data is challenging as it is time-consuming and laborious to label all the training data. Moreover, in the real situation, a large number of fault events are recorded without labels. Therefore, this paper represents an approach of diagnosing fault in transmission network with extreme limited labels(under 1% label) using semi-supervised learning method. First, the fault data is simulated through PSS/E to produce four kinds of short circuit faults with random noise, distance and O-U load fluctuation. Then a semi-supervised model based on pseudo label is proposed to detect the fault type. The experimental results show that the proposed method is able to handle large amount of data with extreme limited labels and represents superior performance on fault classification task.

ABSTRACT. Interaction graphs (IG) can capture the hidden interaction among power system components during cascading failures (CF), which is currently applied to explore vulnerability assessment and resilience enhancement in CF. In this paper, we propose an approach targeted at CF propagation pattern transference with respect to system state evolution and setting related mitigation strategy. The IG is constructed by the CF data generated from a CF model considering balance rules and hidden failure. Then two IG based indexes are employed to investigate CF pattern from two perspective, vulnerability and dispersibility. The simulation results have shown CF propagation pattern transferences in two system state evolutions and verified the critical lines could transfer with system states changing. The results also suggest the existence of Braess paradox in some traditional mitigation strategies considering system state evolution, while the proposed mitigation strategy can avoid the Braess paradox.

ABSTRACT. Partial discharge (PD) diagnosis is an effective tool to track the condition of electrical insulation in the medium voltage (MV) power components. The development of Machine Learning Algorithms (MLAs) promote the automated diagnosis solutions for large scale and reliable maintenance strategy. The aim of this paper is to investigate the performance of two MLAs: Support Vector Machine (SVM) and K-Nearest Neighbors (KNN) for the classification of different types of PD sources. Suitable Features are extracted by applying statistical parameters on the coefficients of discrete wavelet transform (DWT) for checking the performance of both MLAs. The performance of the algorithms is evaluated using key performance indicators (KPIs); accuracy, prediction speed and training time. Besides KPIs, confusion matrix is utilized to show the accurately classified and misclassified PD signals for SVM algorithm. Comparative study of both algorithms demonstrates that SVM provides better results for PD classification as compared to KNN algorithm.

ABSTRACT. With increasing use of modular multilevel converters (MMC), diagnosis and localization of submodule (SM) faults are gaining importance. Most of the existing SM open-circuit (OC) fault detection methods perform diagnosis and localization separately leading to longer detection times. In this paper, a support vector classifier (SVC)--based SM double-switch OC fault diagnosis method, combining detection and localization, is proposed. The proposed method does not require any additional hardware. The training set for the SVC is created by simulating normal operation and SM double-switch OC faults in PSCAD/EMTDC software. Simulation results analyze the performance of the proposed method.

ABSTRACT. This paper validates a phasor model of a gridforming controlled Modular Multilevel Converter (MMC) suitable for large AC grid dynamic studies. The adopted MMC model represent with accuracy the energetic exchanges between the AC side, the DC side and the equivalent internal energy inside the MMC (represented by a single state variable). After description of the converter model, the adopted control strategy is presented and discussed. The simplified model of the converter with its control is validated against a detailed electromagnetic transient model in different control modes (active power control and DC voltage control). Finally, the compatibility of the simplified model with large scale AC dynamic simulations under phasor approximations is validated through the simulation of the Nordic 44-bus test system including grid-forming MMCs.

ABSTRACT. This paper studies the impact of the dynamic controls and protection schemes of distributed energy resources (DERs) and dynamic motor loads on the bulk power system (BPS). This will help characterize the distribution system aggregate modelling and parameterization practices that can be recommended for use in BPS planning studies and identify any sensitivity of best practice to the varying degrees of DER penetration. The study uses a detailed unbalanced feeder model with 3-phase and 1-phase DER with dynamic voltage and frequency controls and voltage, rate of change of frequency (RoCoF) and frequency protection alongside 3-phase motor loads with undervoltage protection interfaced to a simple, balanced transmission system model simulated in a coupled RMS domain. The results from this detailed model are then compared with appropriate aggregate models including der_a and a composite load model equivalent.

ABSTRACT. The determination of the grid impedance for stability studies with broadband signals is a popular method. The disturbing influences of harmonic emissions caused by non-linear loads or power electronics can distort these measurements and thus corrupt the calculated grid impedance. This paper presents an in-depth analysis of these distorting influences by a simulative comparison of the estimated grid impedance in a base scenario without harmonics and a scenario with harmonics. It is shown that the harmonics lead to false resonances, corruption of the impedance phase and additional coupling of the impedance elements. In order to prove that this distortion in the measurements and the resulting corruption of the impedance is not only a problem in the simulations but also in reality, the impedance estimation is carried out on an industrial grid with a high proportion of power electronics.

ABSTRACT. The high uptake of distributed energy resources (DER), particularly solar photovoltaics (PV), may cause the technical limits of electricity distribution networks to be breached. It is therefore necessary to determine the network DER hosting capacity, which ensures the technical limits are not violated. Above the network DER hosting capacity, it is then possible to investigate constraint alleviating approaches such as network upgrades or active network management. In this paper, we propose a DER hosting capacity envelope which contains lower (worst-case) and upper (best-case) bounds of the DER hosting capacity of unbalanced three-phase distribution systems. We obtain these bounds using two optimization-based hosting capacity assessment models. These two models are complemented by another hosting capacity model which takes the fairness concept into account. We also show how this latter, fair hosting capacity assessment model, can be calculated analytically using the Jacobian inverse matrix at a given operating point. The proposed hosting capacity assessment models are used to calculate the DER hosting capacity of the IEEE 37-node MV and IEEE 906-node LV unbalanced distribution networks under various load and generation scenarios.

ABSTRACT. The recent policy changes in the energy sector encourage wider investment in microgrids (MG) having a high proportion of renewable energy and their integration into distribution systems (DS). MGs are expected to complement the DS technical and economic operation; however, their power purchase agreement (PPA) can increase the energy bills of the DS customers. The onus of this study is to investigate the impact of multiple MGs on the DS operation where the MG investors are offered different power purchase contract options. A two-phase methodology is developed to solve the operation interactions of the DS operator (DSO) and MG owners. In the first phase, the optimal sizing of the MGs is determined considering a coordinated integration approach for the MG owners. In the second phase, the yearly co-operation of the DSO and MGs is analyzed considering the yearly load profile and deviation in the capacity factor of the respective MGs. Simulations are carried out on a modified IEEE 33-Bus System consisting of a DSO and three MGs. Relevant results and suggestions are provided for the effective inclusion of the MGs in the DS.

ABSTRACT. This paper assesses the ability of a Transient Stability Index (TSI) to evaluate power systems’ transient stability. The assessment is first accomplished by looking into the TSI’s classification accuracy for identifying stable or unstable cases when different transient stability thresholds are used. Although suitable values for these thresholds have been found, there are still some pitfalls in the use of TSI for stability classification. A new approach for the TSI calculation, called Truncated TSI, is proposed, reducing the index’s time window calculation limits. High accuracy of stability classification has been demonstrated using the proposed approach with appropriate settings of the time window parameters.

ABSTRACT. The DG droop gains in a microgrid are mainly determined by load sharing purposes and may not be beneficial to system stability. In this paper, for enhancing the microgrid stability under small disturbance, a BESS droop tuning scheme in an inverter-based microgrid is established. Firstly, instead of treating BESS as adjustable power injections, we introduce droop control scheme into BESS so that BESS can provide both power injection support under steady-state condition and dynamical support under disturbances. Then, by applying the linear power flow model, we propose a small signal stability constrained optimal power flow (SSSC-OPF) problem to tune BESS droop gains for minimizing the system generation cost while ensuring the system stability. Further, the Benders decomposition-based method with sensitivity analysis is applied to solve the optimization model, where the problem for each iteration is a Quadratic Programming. Finally, the proposed approach is adopted in a modified IEEE-14 system, simulation results demonstrate that the stability performance will be significantly improved with rather small impact on generation cost and short CPU time.

ABSTRACT. This paper presents a dynamic model of a virtual air gap variable reactor (VGR). In contrast to traditionally mechanically operated variable reactors, which have been used as arc suppression coils in compensated networks for centuries, VGRs have shown an enhanced response speed. However, the harmonics created by VGRs are considerably higher than those created by mechanically operated reactors. This study’s dynamic model considers the magnetic saturation behavior of VGRs, enabling a detailed examination of VGRs. Furthermore, the model can be utilized to develop an effective controller for a VGR. A VGR prototype has been used to validate the dynamic model. It is shown that the dynamic model replicates the actual behavior of a VGR well.

ABSTRACT. Due to the steadily increasing share of decentralized renewable energies in the German energy system, flexibilities in the low-voltage (LV) grid such as battery storage systems (BSS) and electric vehicles (EV) must be used by controlling EV charging and BSS discharging and charging power. To control such a highly complex system, autonomous algorithms and artificial intelligence (AI) are needed. One possibility for such a system is presented in this paper using reinforcement learning with a Deep Q-Network (DQN) as a multi-agent approach. The DQN agents were trained in a simulated LV grid environment with pseudo-measurement data of household loads, EV charging, and photovoltaics (PV) power generation profiles. The objective of the DQN agents is to increase the proportion of PV power generation used locally to reduce the maximum capacity of the transformer and reduce the bidirectional power flow. The autonomous AI for LV grid control was validated via the generated pseudo-measurement data of load and power generation multiplied by random factors representing random human behavior and fluctuating PV power generation. The results show a 24.4% reduction in maximum transformer capacity, increased use of locally generated PV power, and a 10% reduction of bi-directional power flow to the medium voltage grid.

ABSTRACT. To effectively analyze the impacts of rapidly evolving new and advanced distributed power generation, demand integration in distribution grids requires grid topologies that replicate the existing network topologies. Estimating these can provide datasets with quasi-network topologies maintaining functional electrical characteristics at any location of interest. Herein, these topologies will be used to identify vulnerabilities that affect grid stability, grid reinforcement, and will form the basis of economic analysis. In this study, a new process is developed for generating geo-referenced, synthetic, low-voltage network topologies using openly available data procured from diverse sources for Germany as a case study. The topologies are first synthesized by collecting the geo-referenced building’s data from OpenStreetMap (OSM). Then, a clustering technique is performed to group collected buildings by known information about the transformer count. Next, detailed algorithms are developed to collect street lines, generate graphs combining streets and buildings, transformer placements, and to assign real electrical parameters. Finally, the methodology is applied to Germany’s building data (extracted from OSM), which yielded 500,000 low-voltage network topologies. The estimated topology data covers approximately 29 million buildings as electrical nodes and approximately 1.7 million km or 0.8 million km of electrical lines length with or without service drops.

ABSTRACT. Increasing photovoltaic (PV) systems on a distribution system impact the operation and lifetime of its components. One of the key components to be impacted is the power transformer at the substation. The aim of this paper is to evaluate the impact of power from PV systems on the lifetime of the substation power transformer. At moderate levels of PV penetration, the loading on the substation transformer decreases, and therefore this will help to prolong the lifetime of the transformer. To estimate this expected benefit, a thermal model for the transformer is used to estimate its hot spot temperature as this temperature is the main factor affecting the degradation of the transformer under normal loading conditions. To illustrate the method, a case study is given. A 10-year period is considered for transformer loss of life evaluation, where practical load growth and PV penetration scenarios are considered. The simulation is carried out on a 15 MVA transformer in the IEEE 123 bus system. Simulation results show that PV penetrations below 100% indeed prolong the transformer lifetime. However, the saved transformer lifetime is not considerable compared to the total transformer lifespan.

ABSTRACT. The accurate transmission line parameters are important in various power system analyses, while potential uncertainty in their values affects the accuracy of control center applications and compromises the selectivity of the protection systems. The transmission line parameters are usually calculated based on manufacturers data, ignoring environmental factors (e.g. ambient temperature) which affect the accuracy of the calculated parameters. Thus, the systematic refinement of the line parameters can be very beneficial for the situational awareness of the power system operators. In this paper, the calculation of the positive sequence parameters of a transmission line of the Cyprus power system through real synchronized phasor measurements is presented. The seasonal variation of the transmission line parameters is analyzed by calculating the parameters during different periods of the year, while the impact of the instrument transformers static error is demonstrated in this real case study.

ABSTRACT. Photovoltaic (PV) power curtailment results in less utilization of available PV and it could be reduced significantly if the low voltage (LV) feeder loading could be predicted in advance with the desired accuracy. In this paper, a methodology is proposed based on the nonlinear auto-regressive model with eXogenous Inputs (NARX) to predict the feeder power loading one hour ahead considering high PV injections. The developed model does not require measured historic smart meter or PV generation data. The predicted output would be useful to maximize the PV energy utilization and will support DSO in the decision-making process for short-time congestion management. A case study has been performed to verify the performance of the developed NARX model, considering the measured feeder data from the Strijp-S region, in Eindhoven, the Netherlands. Obtained results from the proposed method are compared with the base case scenario and actual measured feeder power loading values which indicates the satisfactory performance of the proposed methodology.

ABSTRACT. Giving current environmental concerns, the way energy is produced and handled takes on new contours, giving priority to renewables and clean forms of energy. In this regard, modifying one of the largest forms of pollution, transportation, is moving towards the adoption of electric vehicles (EVs). However, it is necessary to understand to what degree a massive adoption of EVs will have an impact on distribution networks, on the management of loads, and on how the population may suffer changes in its behaviour. This work reviews a tool to simulate EVs trips and charging schedules considering user behaviour aspects. The simulator enables the study of price variations on user behavior showing that dynamic tariffs may prove to be compensatory compared to fixed tariff for users.

ABSTRACT. With a fast-growing market, Electric Vehicles (EVs), will require high availability of charging options, including in public and commercial buildings. In such buildings, EVs present a high potential to ensure the much-needed flexibility to improve the matching between local photovoltaic generation and demand through the implementation of Vehicle-to-Building and Building-to-Vehicle strategies. However, EVs and buildings do not belong to the same entity, which means that leveraging EVs' flexibility depends on establishing an economic relationship, but in most countries, the regulation does not yet allow selling and buying electricity between buildings and EV users. This paper aims at solving this problem and proposes a novel and practical framework to ensure transactive interactions in Vehicle-to-Building systems by aligning electricity and parking values. Additionally, to use such flexibility potential, buildings must incentivize the engagement of EV users and therefore, a multi-objective formulation is proposed to simultaneously ensure the minimization of costs from the building and EV user perspectives. Such optimization requirements are implemented with a bi-level model to optimize the tariffs and the charging/discharging schedule. Several scenarios were simulated showcasing the cost reduction and self-consumption benefits of the proposed solution for building owners and EV users.

ABSTRACT. This paper investigates optimal control for a large population of identical plug-in electric vehicles (PEVs). A mean field assumption is formulated to describe the evolution of the PEVs population and its interaction with the central planner, leading to partial differential equations (PDE). The optimization problem is formulated in the setting of discrete time and state spaces and is convex, but not necessarily linear or quadratic. Two constraints on the population of electric vehicles (EVs) are concerned: a minimal state of charge (SoC) per EV at the end of the charging period and an aggregate power demand constraint. The Chambolle-Pock algorithm is applied to obtain a numerical solution. We study, in two numerical examples, different cost functions that avoid charging synchronization among the EV population.

ABSTRACT. The current uptake of electric vehicles (EVs) results in new challenges for the electricity grid. Due to their relatively high power and potentially synchronized demand, existing local grid capacity might become inadequate in the near future. To provide insight into this development, this work proposes an approach to approximate the outage probability in low-voltage (LV) grids due to EV charging. The introduced model can be used as a tool to understand the impact of uncontrolled EV charging in existing and new LV grids. A case study demonstrates this approach with a grid of 40 households: at 35% EV penetration rate the outage probability is approximately once every 50 days.

ABSTRACT. Electric truck charging may lead to new future challenges in the power grid, which are difficult to assess from today's perspective. One way to deal with the lack of measurement data and empirical values is the generation of synthetic charging profiles. This paper presents an approach to consider charging profiles of electric trucks in future grid planning. For the charging profile generation, real journey data of conventional trucks are combined with vehicle parameters of battery-electric trucks and legal regulations on breaks and rest periods. This results in time series of the truck's location and state of charge for an entire year. The time series can be converted into charging profiles by specifying fixed charging locations and charging power. The applicability of the presented methodology is tested using a truck fleet based on German mobility data. This allows not only the determination of the yearly energy and power requirements of electric trucks with different weights and application classes, but also the estimation of changing mobility behavior when switching from conventional to electric drives. The presented approach is therefore of great interest to both grid planners and fleet operators.

ABSTRACT. The rapid growth of the Electric Vehicle (EV) deployment contributes to zero-emission transport but puts great pressure on the power grid. This paper proposed a Vehicle-to-Vehicle (V2V) energy trading platform for EV charging that allows EVs with surplus electricity to provide electric energy for the exhausted batteries of other EVs. The platform applied the Double Auction (DA) mechanism allowing multiple vehicles to conduct energy transactions simultaneously. The price, availability and amount of the traded electricity can be continually adjusted with high flexibility. The final trading price was determined by the k-factor rule so that balances the profits of all participants. A case study was performed by using the EV data collected from a public car park in the UK. The financial benefits of all participants were increased significantly and the energy consumption from the power grid was reduced by using the proposed trading mechanism. These potential reductions in EV running costs encourage the development of the EV market.

ABSTRACT. In heavily inverter-based power systems, the minimal difference between nominal currents and short circuit fault currents makes fault detection difficult. In this paper, a model based approach that utilizes the assumed dynamics of a grid-tied inverter along with a nonlinear observer is proposed to detect system abnormalities and faults. The goal of this observer is to provide insight into unmodelled signatures present in the system. These signatures in turn can be utilized to identify faults in the system creating a detection scheme. Most importantly, this method requires no additional sensors other than those required by the grid-tied inverter which is a key advantage over many proposed solutions in literature. Line-to-ground and line-to-line faults were observed and detected with the proposed observer in the PSCAD simulation environment in this work.

ABSTRACT. Distributed energy resources (DERs), such as photovoltaic units, wind turbines and batteries, are able to increase operational flexibility in Distribution Networks (DNs) that can be used to resolve power quality issues. Furthermore, the advances in metering infrastructure, such as the roll-out of smart meters in many countries worldwide, provides better observability of DNs. However, DERs in DNs usually belong to different stakeholders with alternative objectives that might even be conflicting (e.g., grid constraints are not considered by all). This paper presents a practical approach, based on a data-driven scheme, to safely use the DER operational flexibility and, at the same time, respect the objectives of individual households, network operators, and aggregators. First, a grid health estimator is constructed using Support-Vector Machines (SVM). Then, a real-time modification signal is used to alleviate network constraints violations. The performance of the proposed algorithm is demonstrated using the MV and LV European benchmark grids.

ABSTRACT. Prosumers will play an increasingly important role in the modern smart grid through their ability to modulate consumption as well as supply excess generation back to the grid. Aggregations are a profitable model for prosumers to trade surplus energy, and their design raises important questions on which prosumers and resources would form the most beneficial aggregation. We develop a numerical test which quantifies the degree of complementarity of prosumers and can be used to design optimal aggregations. This result can extend to evaluate resources such as local generation and load shift. We validate our results through simulations in an agent based modeling framework using real building loads and solar generation data.

ABSTRACT. As renewable energy sources penetration is increasing, the energy aggregator entity takes an important role to provide a highly flexible generation and demand as required by the smart grid paradigm. This paper proposes a model for energy resource management problem in intraday operation (hour-ahead). The model is formulated as mixed-integer linear programming and using the CPLEX solver. A distribution network with 180 buses located in Portugal considering high distributed energy resources penetration is used to demonstrate the application of the proposed model. The results show the influence of the forecast errors and the contractual constraints with the energy storage systems and electric vehicles charging stations in the hour-ahead scheduling costs.

ABSTRACT. The minimization of distribution losses is one of the main necessities in the community based DC microgird (MG) with energy sharing between different DERs with surplus energy and/or with load diversity. Energy sharing can take place within a community MG or with other nearby community MGs. Distribution losses are normally calculated based on the distance (or distribution conductor length). However, they can also be dependant on voltage difference between sharing nodes, amount of energy sharing, and conductor size, etc. Therefore, peer-topeer (P2P) energy sharing can’t be optimally managed with only distance-based decision techniques, and must be decided based on appropriate analysis of distribution losses in an optimal power flow framework. In this paper, we proposed a nonlinear P2P energy sharing framework to ensure optimal power flow with minimal distribution losses and compare our proposed framework with nearby energy sharing model. We present the detailed analysis of distribution losses in a DC MG with several distributed generators (DG) and energy storage systems (ESSs) at distributed locations. The simulation results validate the proposed optimal P2P sharing method in reducing distribution losses upto 60%.

ABSTRACT. Due to the increasing shutdown of conventional power plants and the associated decrease in rotating mass as well as the increased use of renewable energy resources in the context of the energy transition, various concepts within the electrical energy system are the focus of research. For example, microgrids, in which the generation and consumption of electrical energy is local and thus a self-sufficient grid operation is targeted. For structured, comparable and reproducible research on microgrids, there is a need for suitable benchmark grid models for simulative analyses. Therefore, this paper presents a methodology to extend the already published ‘SimBench’ benchmark dataset so that benchmark microgrid network models can be created.

ABSTRACT. Local energy markets (LEMs) are a concept for the integration of volatile renewable energy plants and increasing demand from sector coupling (electric vehicles, heat pumps) into the energy system. So far, the focus of research has been on market integration of prosumers, neglecting the effects of LEMs on the distribution grid. This paper therefore presents a market design that ensures validity of the market outcome in terms of line utilization and further reserves capacity for the application of reactive power for voltage control. First, Power Transfer Distribution Factors are used to estimate the line utilization when determining the market outcome. In a second step, a reactive power schedule is calculated by an AC optimal power flow that meets voltage limits in addition to thermal line restrictions. The presented method is evaluated using a real low voltage network in the Allgäu region (Southern Bavaria, Germany). The analysis shows that in the future scenario studied, violations of network restrictions can be avoided by using an LEM in combination with the presented methodology.

ABSTRACT. The abundant solar energy we receive needs to be efficiently utilized for the green environment and benefit mankind. This paper proposes a simple and optimal Power Management System (PMS) for DC-Microgrid (DCMG). The system uses a power switching circuit that interfaces the PV-array, DC-load, and the Energy Storage System (ESS) in each Nano-Grid (NG) of the DCMG. The overall DCMG consists of a cluster of four NGs linked through a common DC bus with a single ESS. The proposed method not only offers a PMS at each NG, but also incorporates the power-sharing capability among the NGs. The proposed method of PMS makes effective use of the power flow by considering the load demands for the power and the resources' availability to the load at each NG, thus forming the merit. The proposed method offers simplicity in the structure and control of the system.

ABSTRACT. Energy communities combine assets of different community members and can create financial value from energy trade and from providing flexibility. Sharing this value among the community members can be done in various ways depending on the main objectives of the energy community and selecting the most suitable value sharing method for a particular energy community is not always straightforward. There are several, partly contradictory, requirements to be fulfilled that include fairness, stability, understandability, computational feasibility and ability to incentivize individual members to act in a way that benefits the whole energy community. In this paper, alternative energy community value sharing methods are compared from different viewpoints. The most suitable value sharing method for a particular energy community can be selected by first defining what are the main requirements for value sharing and then proceeding to a more detailed analysis of the value sharing method options that fulfill those requirements.

ABSTRACT. A predictive real-time Energy Management System (EMS) is proposed which improves PV self-consumption and operating costs using a novel rule-based battery scheduling algorithm. The proposed EMS uses the day-ahead demand and PV generation forecasting to determine the best battery scheduling for the next day. The proposed method optimizes the use of the battery storage and extends battery lifetime by only storing the required energy by considering the forecasted day-ahead energy at peak time. The proposed EMS has been implemented in MATLAB software and using Active Office Building on the Swansea University campus as a case study. Results are compared with published state-of-the-arts algorithms to demonstrate its effectiveness. Results show a saving of 15% and 30% in total energy cost over six months compared to a forecast-based EMS and to a conventional EMS, respectively. Furthermore, a reduction of 54% in the net energy exchanged with the utility by avoiding the unnecessary charge/discharge cycles.

ABSTRACT. The increasing number of Photovoltaic (PV) systems connected to the power grids makes them vulnerable to the projection of shadows from moving clouds. Solar Global Irradiance (GSI) forecasting allows smart grids to optimize energy dispatch preventing cloud coverage shortages. This investigation compares the performances of unsupervised learning algorithms (not requiring labelled images for training) for real-time segmentation of clouds in a ground-base infrared sky-imaging system, which is commonly used to extract cloud features using only the pixels where clouds are detected.

ABSTRACT. Moving clouds affect the global solar irradiance that reaches the surface of the Earth. As a consequence, the amount of resources available to meet the energy demand in a smart grid powered using Photovoltaic (PV) systems depends on the shadows projected by passing clouds. This research introduces an algorithm for tracking clouds to predict Sun occlusion. Using thermal images of clouds, the algorithm is capable of estimating multiple wind velocity fields with different altitudes, velocity magnitudes and directions.

ABSTRACT. This work addresses the problem of reconstructing topology and cable parameters of three-phase low-voltage networks when no a priori information about them is provided and not all the nodes in the grid are equipped with smart meters. This paper presents a methodology to obtain an estimation of the electrical model of each phase of the network, analysing voltage and current time-series measurements provided by the available meters. The proposed methodology involves an iterative algorithm developed to tackle the network reconstruction problem when unmetered nodes are located reasonably far from each other. The algorithm is tested on a 30-node network with different sets of metered nodes, providing relevant solutions in most scenarios having more than 80% of metered nodes.

ABSTRACT. This paper proposes a novel data-driven power system event detection and classification method based on 5TB of actual PMU measurements collected from the US western interconnect. Firstly, a set of comprehensive power quality rules are proposed to pre-filter the raw data and extract the regions of interest (ROI). Six distinct event categories are defined and corresponding patterns are chosen as references. Meanwhile, detailed characteristics of patterns are summarized to enhance our understanding of the actual events. Then, the time-independent feature vectors are generated by extracting the statistical, temporal, and spectral features from the raw time-series data. Furthermore, an ensemble model is proposed to cluster the events by combining multiple K-means clustering models using a voting strategy. Besides, both system-level and PMU-level clustering models are developed. The accuracy and robustness of the event detection method are further improved through interactive evaluation of the two-level clustering results. This paper summarizes the actual characteristics of each event category and provides a reliable basis for accurate label generation. The experiments demonstrate the effectiveness of the proposed event detection and classification method.

ABSTRACT. Energy system models are widely used for operation and expansion planning, scaling from single houses up to supranational energy grids. They can provide essential input for decision makers. These are, however, often non-technical persons and thus unfamiliar with mathematical modeling, which makes them reliant on others to explain the model results to them. In order to make energy system models more directly accessible, we introduce a chatbot that enables also non-expert users to interact with an energy system model through natural language. Built with state-of-the-art natural language processing tools, it allows to manipulate the model inputs and can interactively answer questions about the results, both in free-form text and via structured plots. We present example interactions for a model of the German energy transition.

ABSTRACT. Demand Response relies on customers being capable of curtailing energy consumption during peak hours. While it can be successful in shedding excess load, it may cause customer dissatisfaction and feeder load imbalances. In this paper we address this problem by looking at building-level energy consumption. We propose and test the efficiency of several building selection strategies in reducing the impact on customer comfort and in achieving feeder-level load balancing. Experiments are performed on real-life datasets from a campus microgrid.

ABSTRACT. While sampling in classical signal processing is well-developed and studied, sampling in the Graph Signal Processing framework and its application domains are fairly new. In this paper, sampling of power grids' graph signals are discussed and analyzed in the context of two applications including the recovery of missing measurements during cyber stresses and the optimal PMU placement problem. The effects of various topological and power-dynamical factors in the system are evaluated for the selection of the sampling-set for power grid graph signals and graph signal reconstruction performance. Moreover, a novel sampling-set selection criterion based on the error introduced in the process of band-limiting the graph signal has been proposed.

ABSTRACT. This paper presents a cyber-physical algorithm for grid-interactive DER control to enable two features of Virtual Power Plants (VPPs) dispatch and grid voltage regulation, considering the communication and security impacts. We first formulate the DER dispatch problem as a real-time, iterative, and grid-interactive DER control problem. Thereafter, we consider a probabilistic traffic model to characterize packet delays and loss in a communication network, and study how the delays enter the process of information exchange among the grid measurement units, local DER controllers and the grid control center that coordinately execute this dispatch algorithm. Finally, a strategy combining delay threshold and modified message update rules is proposed to immune the asynchrony resulting from the communications network traffic and it avoids possible numerical instabilities and sensitivities of the tracking and regulation capabilities of this DER control algorithm. By implementing the proposed cyber-physical algorithm on the modified IEEE 37-node system, our preliminary results exhibit that the uncertainties of the underlying communications infrastructure must be considered for the VPP tracking and regulation capabilities of any DER in a generic Cyber-Physical System (CPS), because the delayed voltage measurements in the uplink/bi-link cases result in the off-track in VPP dispatch and jittery in voltage regulation.

ABSTRACT. Estimation of topology of the power grid from electrical markets has been a point of interest of the industry traders. Under a DC model, methods try to recover the grid Laplacian though locational marginal prices (LMPs) exist. However they give incompatible grid Laplacian after the full recovery, leaving positive off diagonal entries. This paper discusses a new model for grid topology estimation, which involves the key property of diagonal dominance to guarantee a compatible full recovery of grid topology. We develop alternating direction method of multipliers (ADMM) algorithm for the new model which is much more scalable than existing algorithms aiming to solve the same problem thanks to the newly introduced sum-log-diagonal term. Numerical experiments indicate that our method is highly scalable to the number of nodes and/or to the number of time steps of the data while giving higher quality solution.

ABSTRACT. The Horizon2020 project FlexPlan (https://flexplan-project.eu/) aims at establishing a new grid planning methodology considering the opportunity to introduce flexibility resources (storage and DSM) in electricity transmission and distribution grids as an alternative to traditional grid expansion. After creating methodology and tools, the project applies them to analyze six regional cases covering nearly whole Europe, aimed at casting a view on grid planning in Europe till 2050. In this way, FlexPlan tries to understand how flexibility can help reducing future grid investments. The project ends up formulating regulatory guidelines for NRAs.

The session will provide first an overview on the FlexPlan methodology and then some details on the innovative tools that are created (planning tool and pre-processor). Thereafter, details are provided on the complex scenario framework shaping the 6 regional cases and on the challenges to build up the planning tool. A 30 minutes Q&A time will close the session.

ABSTRACT. Battery Energy Storage Systems offer new opportunities to address the upcoming challenges of the European Power system but many open questions remains on their value, design and performance. The OSMOSE H2020 project with a large consortium of Transmission System Operators, academics and industrials addresses the following key issues through simulations and demonstrations : - What is the profitable place for storage today? And in long-term decarbonized scenarios? - How to design BESS and their controls to provide efficiently multi-services and maximize their value? - What are the main challenges faced by BESS operators today? How to build on their experiences? After a presentation of the key contributions of the project on those questions, the audience will be invited to provide its feedbacks and own experience.