Search

Utilizing distribution grid flexibility for ancillary services requires the coordination and dispatch of requested active and reactive power to a large number of distributed energy resources in underlying grid layers. This paper presents an approach to hierarchically dispatch flexibility requests based on Online Feedback Optimization (OFO). We implement a framework of individual controllers coordinating actors, contributing to flexibility provision, to track a requested operating point at the interface between grid layers. The framework is evaluated in terms of performance during coordination and possible interaction between individual controllers, both central and distributed. Results show high reliability and robustness of the OFO controllers as well as an efficient dispatch of active and reactive power. Its computational efficiency and capabilities in set point tracking during online grid operation are making OFO a promising approach to the flexibility dispatch problem. • Hierarchical control of distributed flexibility with Online Feedback Optimization. • Provision of ancillary services across multiple grid layers. • Accurate set point tracking by the proposed control structure. • Robustness against model mismatch and disturbances through closed loop control. • Simulative and experimental case studies. [ABSTRACT FROM AUTHOR]

Active distribution grids can experience voltage fluctuations and violations due to the high penetration of variable distributed energy resources (DERs). These problems might occur because of the uncertain and variable generation natures of these resources, especially solar photovoltaic resources, during panel shadowing scenarios. Volt-VAR control (VVC) is an efficient method that controls the reactive power set-points of the inverters to regulate the voltage of distribution grids. Although several VVC approaches have been proposed recently, the performance of these approaches degrades significantly if behind-the-meter solar generation data are unobservable/missing. Therefore, it is necessary to impute missing/unobservable PV data accurately to be utilized in VVC approaches. This paper proposes a model-free, data-driven, centrally trained, and decentrally executed multi-agent deep reinforcement learning-based VVC architecture to regulate the voltage of distribution networks. A generative adversarial network (GAN) is incorporated to impute the unobservable PV data accurately, which improves the performance of the proposed control architecture. The proposed multi-agent-soft-actor–critic algorithm (MASAC)-based VVC technique utilizes the actual PV dataset as well as the imputed dataset from the GAN framework to learn the optimal coordinated control policy for controlling the optimal reactive power set-points of PV inverters. The effectiveness of the proposed approach is analyzed on a modified IEEE 34-bus test case with added PV inverters. The results are compared and analyzed with a base case model with no VVC and VVC with a local droop control approach, genetic algorithm optimization, and a centralized soft actor–critic-based approach. Moreover, the performance of the proposed approach is compared with that of a multi-agent VVC framework without using the PV generation data and load information as the system state. The results illustrate that the proposed method with more state input improves the voltage profile and reduces the power loss of the network across various loading and PV generation scenarios. • A multi-agent GAN-DRL-based Volt-VAR control of distribution systems. • A generative adversarial network (GAN) is incorporated to impute the unobservable PV data. • A MASAC algorithm is used to control the reactive power set-points of PV smart inverters to control the voltage. • Case studies are performed on IEEE 34-bus and IEEE 123 bus systems and performance is compared with different state-of-the-art control approaches under different scenarios. [ABSTRACT FROM AUTHOR]

With the growing adoption of renewable energy resources in the power grid, co-optimizing the operation of water and power distribution networks (WDNs and PDNs) increases the flexibility of both systems. The optimal water power flow problem (OWPF) is formulated to manage resource utilization across the WDNs and multi-phase PDNs. The nonlinear WDN hydraulic constraints and PDN ac power flow equations render the OWPF problem nonconvex. This paper formulates a convex optimal water power flow (C-OWPF) problem adopting successive approximations in WDNs and branch-flow relaxations in PDNs. The C-OWPF problem is a mixed-integer semidefinite program, which is computationally challenging even for small instances. Additionally, privacy considerations of WDN and PDN operators motivate the need for solving the C-OWPF in a distributed fashion. This paper develops a C-OWPF solver based on Benders decomposition that overcomes computational complexity challenges and preserve the privacy of the respective operators. The merits of the Benders decomposition-based solver are demonstrated on the IEEE 4-bus PDN coupled with a 3-node WDN and the IEEE 123-bus PDN coupled with the 36-node WDN. • Optimal pump scheduling in multi-phase distribution networks is studied. • Monomial approximations for water flow hydraulics are considered. • SDP relaxations and approximations for power flows are compared. • Optimal reactive power dispatch for PV inverters is included. • A novel multi-period optimization method is put forth via Benders' decomposition. [ABSTRACT FROM AUTHOR]

• For islanded ADM, a novel supervisory control method is presented. The challenge of restoring frequency and voltage magnitude while lowering generating cost is formulated using the sensitivity matrix generated from the Jacobian matrix incorporating droop-based DGs as a multi objective function. • PV and wind units and uncertain nature of their output are considered with Monte Carlo simulation and scenario reduction technique. • The effect of the reactive power on the frequency is considered with using the frequency sensitivity. Load demand has locally met the distributed energy resources (DERs) under an infrastructure that is called microgrid. In this paper, a new supervisory control approach is presented based on MPC method, using a stochastic optimization model. The objective function (OF) of the proposed central controller consists of the combined cost-based and system-based parts. Indeed, the setpoints of DERs are updated by the supervisory control method to regulate voltage and frequency to their nominal values as well as to decrease the generation costs. The stochastic optimization formulation is conducted based on MPC structure. Moreover, the uncertain nature of DERs is considered in the optimization formulation and MPC framework using several scenarios. In addition, the effect of reactive power on frequency is considered in this paper and the results show better voltage and frequency regulations. The algorithm is implemented on a modified IEEE 33-bus test system with AC and DC parts. Furthermore, several cases and the sensitivity analysis are accomplished to show the efficiency of the proposed model. [ABSTRACT FROM AUTHOR]

• The XGBoost junction temperature calculation model is used to improve the junction temperature calculation efficiency and reduce the influence of IGBT parameter deviation on the junction temperature calculation accuracy. • A data-driven IGBT reliability evaluation method is proposed to realize the quantitative evaluation of IGBT reliability in PV inverter when photovoltaic power supply participates in reactive voltage regulation in distribution network. • The voltage/var optimization model of the distribution network considering the reliability of the PV inverter is established, which significantly improves the reliability, minimum lifetime and average lifetime of the PV inverter. • DDPG algorithm is used to realize the voltage/var optimization of distribution network considering the reliability of the PV inverter, which effectively improves the solution speed of the voltage/var optimization model and lays a foundation for realizing online optimal voltage/var control. Fully exploiting the reactive power support capability of the distributed photovoltaic power supply is helpful to solve the problems of voltage fluctuation, voltage overlimit and new energy consumption in the distribution network. However, the reactive power output of the photovoltaic power supply will seriously threaten the reliable operation of the photovoltaic inverter. Therefore, this paper proposes a data-driven voltage-reactive optimization control strategy considering the reliability of the photovoltaic inverter. Firstly, the data-driven model is used to calculate the insulated gate bipolar transistor junction temperature, which improves the calculation efficiency of the insulated gate bipolar transistor junction temperature and reduces the dependence of the evaluation accuracy on the insulated gate bipolar transistor parameters. Then, the voltage-reactive optimization control model of the distribution network considering the reliability of the photovoltaic inverter is established, and the average junction temperature and junction temperature fluctuation of the insulated gate bipolar transistor are introduced into the model optimization goal. The optimization model is transformed into the reinforcement learning task, and then the deep deterministic policy gradient algorithm is used to realize voltage-reactive control. According to the simulation of IEEE 33 node distribution system, the proposed strategy can increase the minimum and average insulated gate bipolar transistor lifetime of all nodes by 6 years and 4 years. At the same time, it can reduce the minimum and average levelized cost of energy of all nodes by 0.1095 $/W and 0.0318 $/W. [ABSTRACT FROM AUTHOR]

• To detect islanding, the 4th harmonic is injected into the output current. • A modified Goertzel filter is employed to measure the harmonics of voltage. • The method is robust to variation of load quality factor and resonant frequency. • The impact of reactive power on frequency after islanding is compensated. • The method is fast and detects islanding within two cycles of fundamental frequency. Islanding occurs when a utility-interactive distributed generator (DG) is disconnected from the utility grid and continues to supply power to local loads. This condition has several consequences including damage to the equipment and personnel safety risks. Therefore, it is required by utility interconnection standards for the DG to stop energizing local loads in the case of islanding. In this paper, an active islanding detection method (IDM) based on injecting a disturbance into the phase-locked loop (PLL) of a grid-connected photovoltaic (PV) inverter and monitoring the harmonic components of the point of common coupling (PCC) is proposed. A modified Goertzel algorithm is utilized to detect the injected harmonic which is effective for a wide range of load conditions in terms of load quality factor, resonant frequency, and reactive power. The simulation results in MATLAB/Simulink show that the proposed technique is fast with minimum side effects on power quality features such as output current total harmonic distortion (THD). This method is implemented on a single phase 5 kW grid-tied PV inverter and is tested experimentally. The experimental results verify that the proposed technique is fast with zero non-detection zone (NDZ) which is effective for a variety of load conditions. [ABSTRACT FROM AUTHOR]

• This paper focuses on boosting the diesel generator (DG) set's utilization by exploiting the solar photovoltaic plant (SPVP) power converter's available capacity as DSTATCOM. The key highlights of this work are as follows. • The power factor of the DG set is regulated at unity irrespective of the SPVP generation or load behavior, enhancing the efficiency of the diesel engine and electric generator. Consequently, fuel consumption is decreased, lowering the cost of electricity. • The SPVP carries the burden of the entire load's reactive power. Hence, the DG set's surplus capacity, initially reserved to supply the reactive power to the load, can be utilized now to commit the additional load on the same rating system. • Due to the DSTATCOM operation, the SPVP operates 24 hours, raising the overall utilization factor of the SPVP. Hence, the payback time is less compared to the existing system. • The DSTATCOM operation of the SPVP boosts the power quality performance of the microgrid without causing any high additional costs to the system. Thus, the IEEE std. 519 [21] is followed at POI, with improved utilization of the microgrid. • A least mean 2L th (LM2L) rule-based adaptive filter has been developed for a quick and accurate estimation of the load's fundamental component current. It is a hybrid adaptive filter algorithm that uses the conventional normalized least mean square (N-LMS) and normalized least mean fourth (N-LMF) algorithms to improve the steady state and transient response. In the existing solar photovoltaic (SPV) and diesel generator (DG) based standalone microgrid, the power factor of the DG set reaches around 0.5-0.6. Further, it encounters poor power quality performance due to harmonics currents of the local loads, coupled at the microgrid's point of interconnection (POI). All these issues significantly affect the DG set performance, causing additional losses and reducing its fuel efficiency. In addition, the SPV plant does not provide any ancillary services in the existing system and remains ideal most of the time, causing the under-utilization of microgrid's power converters. In this context, this work presents a least mean 2L th (LM2L) rule-based adaptive filter-controlled SPV system, where its power converters are regulated to deliver multifunctional capabilities such as enhancing the DG set's power factor and power quality performance. As a result, the DG set power factor improves to unity, and the %THD value of its voltage and current reduce to below 5%. Hence, the fuel efficiency of the DG set is improved, line losses are decreased, and the additional load can be committed to the DG set capacity, which is kept reserved to supply reactive power to the load. Moreover, the modifications, which are earlier required in the DG set design to accommodate the distortion in its voltages and currents due to the nonlinear load, can be avoided. The LM2L algorithm is used as a prefilter for the measured load currents and determines their fundamental component. It has two operating rules, which decide the fundamental component's estimation depending on the system conditions. As a result, it demonstrates an improved steady state and dynamic performances compared to the existing adaptive filter algorithms. Test results verify the control and design of the microgrid on a laboratory prototype. [ABSTRACT FROM AUTHOR]

• We proposed an adaptive (two-stage) robust energy management for an isolated microgrid considering renewables uncertainty. • Minimization of the expected total operating costs of the microgrid is the objective function. • Reactive power costs from diesel generators are considered and are co-optimized with active power costs • The reactive power capability of the inverter-interfaced distributed energy resources is considered. • Detailed modeling of the diesel generators and inverter interfaced distributed energy resources by the consideration of their capability curves is provided. • The results obtained validate the effectiveness of the proposed solution. This paper presents an adaptive robust Energy Management (EM) model for an isolated Microgrid (MG) to decide for the day-ahead optimal dispatch to effectively manage the MG power sources considering uncertainties in Renewable Energy Sources (RESs). The aim of this model is to minimize the operation costs, reactive power costs, spinning reserve and load shedding. Usually, fuel consumption costs of diesel generators are considered to be dependent on active power generation only. However, neglecting the related reactive power costs might result in increased operation costs and deviations from optimal dispatches. Hence, this paper optimizes the costs related to both active and reactive powers of diesel generators. Furthermore, simultaneous active/reactive power dispatch can lead to accurate operation decisions compared to separate dispatch for active or reactive power alone. In addition, this study considers the reactive power capability of inverter-interfaced Distributed Energy Resources (DERs) so that reactive powers can be supplied from inverter-interfaced DERs. Moreover, detailed models for the different resources are presented, especially for diesel generators and inverter interfaced DERs where actual capability curves are used instead of the widely used box constraints. The problem is formulated as nonlinear programming problem using GAMS program and solved by the CONOPT solver. [ABSTRACT FROM AUTHOR]

• A new control scheme is proposed to benchmark performance of the power control strategies. • The DC-link voltage oscillations originated from active power oscillations are eliminated. • Increasing lifetime of inverter as well as eliminating third order current harmonics at AC side. • Protection of overvoltage at DC-side of inverter and overcurrent at AC-side of inverter. • Comparative analyses for state of the art proposed control scheme are performed. Voltage and current oscillations have become a primary power quality issues for DC and AC systems. In this respect, this paper presents a new control scheme for analytical implementation of reference current generator (RCG) based power control strategies to eliminate oscillations both DC and AC sides. Benchmarking of the RCG based power control strategies are examined in terms of elimination of the power oscillations, reducing DC-link voltage oscillations, avoiding DC-side overvoltage and preventing AC-side overcurrent. Compared to the existing strategies, controlling reference active power is provided with a closed-loop control system. A noteworthy contribution of this paper is to eliminate active power oscillations, which is an effective method to suppress the DC-link voltage oscillations resulting in third-order harmonic at the grid side. This helps to reduce the size of the DC-side capacitor and also improves the lifetime of the inverter, safety and stability of the grid connected primary energy resources (PERs). Detailed comparative analysis with the most recent strategies have been performed for state of the art proposed control scheme. Finally, various case studies are presented to support the proposed control scheme can mitigate the DC-link voltage oscillations, effectively. [ABSTRACT FROM AUTHOR]

This paper proposes a voltage prevention and emergency control strategy that consists of coordinately arranging multiple reactive power sources in order to handle the point of common coupling (PCC) voltage fluctuation and stability in large-scale PV power plants. When a disturbance occurs at the PCC, dynamic reactive power compensation devices are coordinated preferentially to support the PCC voltage. After the disturbance is cleared, the reactive power in dynamic and fast devices is transferred into static and slow devices so that the static VAR generation (SVG) can maintain a large power margin for coping with the next disturbance. Moreover, the reactive power output of the individual inverter in PV power plants is coordinately allocated using a model to optimize the in-station voltage distribution. Finally, the effectiveness of the proposed control strategy is verified by an example simulation of a practical large-scale PV power plant. [ABSTRACT FROM AUTHOR]

• Cascading jam-congestion and source-load mechanism in ET-DN was demonstrated. • Combined the DTA model of TN based on DVI with the multi-period AC OPF model of DN based on MISOCP. • Jam/Congestion in ET-DN may cause undesirable chain propagation and source-load mismatch problem. • A joint optimization scheme of ET-DN with dynamic distribution network reconfiguration is proposed, to eliminate jam-congestion deterioration and improve the operation of ET-DN system. The development of dynamic wireless charging technology and photovoltaic distributed generation sources provides a solution to eliminating the mile anxiety of electric vehicles (EVs), however which strengthens the spatio-temporal coupling in electrified transportation-distribution network (ET-DN). In this paper, by combining the dynamic traffic assignment model of transportation networks (TNs) with the multi-period AC optimal power flow model of distribution networks (DNs), the dynamic coupling model for ET-DN is established to depict the spatio-temporal distribution of EV charging load and jam-congestion cost. We found that coordinating the EVs charging and active/reactive power resources helps to alleviate jam in TNs and congestion in DNs, while source-load spatio-temporal mismatching may lead to the deterioration of jam-congestion in ET-DN, even there are sufficient active/reactive power resources, which has an adverse impact on the system security and economy. To overcome the source-load spatio-temporal mismatching problem, a joint optimization scheme of ET-DN with dynamic distribution network reconfiguration is proposed to eliminate jam-congestion deterioration and improve the operation of ET-DN system. [ABSTRACT FROM AUTHOR]

• A multivariate approach to probabilistic industrial load forecasting is proposed. • Univariate quantile regression and quantile regression forests generate individual forecasts. • A multivariate quantile regression model combines individual forecasts of active and reactive power. • Multivariate quantile regression on quantile regression forests is the best pick for aggregate load. • Multivariate quantile regression on univariate quantile regression is the best pick for a single load. Industrial load takes a big portion of the total electricity demand. Skilled probabilistic industrial load forecasts allow for optimally exploiting energy resources, managing the reserves, and market bidding, which are beneficial to transmission and distribution system operators and their industrial customers. Despite its importance, industrial load forecasting has never been a popular subject in the literature. Most existing methods operate on the active power alone, partially or totally neglecting the reactive power. This paper proposes a multivariate approach to probabilistic industrial load forecasting, which addresses active and reactive power simultaneously. The proposed method is based on a two-level procedure, which consists of generating probabilistic forecasts individually for active and reactive power through univariate probabilistic models, and combining these forecasts in a multivariate approach based on a multivariate quantile regression model. The procedure to estimate the parameters of the multivariate quantile regression model is posed in this paper under a linear programming problem, to facilitate the convergence to the optimal solution. The proposed method is validated using actual load data collected at an Italian factory, under comparison with several probabilistic benchmarks. The proposed multivariate method enhances the skill of forecasts by 6% to 13.5%, with respect to univariate benchmarks. [ABSTRACT FROM AUTHOR]

• This paper presents a new analytical method to optimal siting and sizing of DERs. • This method is based on a minimizing the total daily power losses. • In the formulation, the new factor in order to consider time-varying DG defined. • This paper is considered several types of time-varying load and DG model. • This method identifies optimal buses to be placed on DERs and sizes of this units. In the most studies related to allocation of distributed generation the load model assumed in constant. Such assumptions may lead to inconsistent and misleading results in the reduction of energy losses, improve system stability, etc. However, this paper is considered several different types of time-varying to allocation of units. In this study, Optimal Sizing and Siting of Distribution Generation units (Consisting of dispatchable, nondispatchable and photovoltaic) Problem are provided at the same time with considering the role of demand response and time-varying voltage-dependent load models (Consisting of commercial, industrial and residential). This method first identifies a set of buses to be placed of DG units. The optimal sizes of DG units at the identified buses are then evaluated by minimizing Multiobjective Index (Consisting of active, reactive power losses). The proposed method has been tested and validated on 33-buses test distribution system. The results show that demand response and different load models have a significant impact on the planning of Distribution Generation units in radial distribution network. [ABSTRACT FROM AUTHOR]

• An emulated wind turbine (WT) drive system is used. • The THD values for source current are 3.03% for PI and 1.93% for GA-ANN. • Landsman converter has proved to be more efficient. • The gain of 4 and an output voltage of 1200 v which is more when compared to Zeta and Flyback converters. Because of global warming concerns and the aim of decreasing nonrenewable energy sources, the direction of power production is being steered towards sustainable power sources rather than conventional ones. Hence, a n emulated wind turbine (WT) drive system is used in this study to assess the behavior of a grid-connected Doubly Fed Induction Generator (DFIG) based wind energy system. In order to ensure uniform and continuous power supply from renewable sources to the grid, a hybrid microgrid system has been created by taking into account a PV fed Landsman converter as part of the design. The control of wind generators is divided into two parts: rotor control and grid control. The grid side control is achieved via the use of a PI controller, and the rotor side control is achieved through the combination of the stator flux and the D -q reference frame. The suggested system manages actual and reactive power, as well as DC link voltages, in an efficient manner, and this has been shown at a variety of various speeds (sub and super synchronous). This paper presents a new Artificial Neural Network (ANN)-based tracking of the Maximum Power Point (MPP) for PV-fed Landsman converters that is based on the Genetic Algorithm (GA). The controlling mechanism is tested on an emulated wind turbine that is driven for concurrent control of the DC tie potential, active and reactive powers using MATLAB and an experimental bench configuration. By using this methodology, the THD values for source current utilizing the PI and GA-ANN based systems, are obtained as 3.03% and 1.93%, respectively. Also, the employment of Landsman converter has proved to be more efficient with the gain of 4 and an output voltage of 1200 V which is more when compared to Zeta and Flyback converters. [ABSTRACT FROM AUTHOR]

• Illustrates the contribution of wind generation, and investigates the influence of the grid-forming mode-based VSWT-DFIGs. • Coordinated design of wide-area damping controller for utility-scale wind generation power plants considering time delays uncertainty. • Optimization is used to optimize the parameters of damping controller. • Contains case studies on a large power system, including FACTS, and wind generation. • MATLAB/Simulink-based nonlinear simulations are performed, for variable wind speed, 3-phase faults, and dynamic load model. The large integration of wind generation resources into power systems may causes low-frequency oscillations and deterioration in power system stability. This paper presents a novel coordinated approach-based Wide-Area Damping Control System (WADCS) for large wind generation integrated power system, while taking into consideration of the operational uncertainties. The coordinated WADCS approach is used to enhance the damping of inter-area oscillations while maintaining the voltage and frequency in the prescribed range. In a coordinated WADCS, a wide-area damping controller is coordinated with the rotor side converter model of wind generation resources (WGRs). The Static compensator (STATCOM) is used to enhance voltage and reactive power support for WGRs. To provide the additional damping, a static synchronous series compensator (SSSC) is also employed with the WADCS. The geometric measures of controllability/observability and residues method is utilized to select the appropriate input control signals and optimal location of wide-area damping controller. A multi-objective grey wolf optimizer (MOGWO) algorithm is used to optimize the parameters of the wide-area damping controller to enhance the damping. MATLAB/Simulink-based nonlinear simulations are performed considering operational uncertainties such as variable wind speed, 3-phase faults, variable load, and time delays to demonstrate the efficacy of the proposed coordinated control for modified WGRs integrated test power system. [ABSTRACT FROM AUTHOR]

Real-time coordination of distributed energy resources (DERs) is crucial for regulating the voltage profile in distribution grids. By capitalizing on a scalable neural network (NN) architecture, one can attain decentralized DER decisions to address the lack of real-time communications. This paper develops an advanced learning-enabled DER coordination scheme by accounting for the potential risks associated with reactive power prediction and voltage deviation. Such risks are quantified by the conditional value-at-risk (CVaR) using the worst-case samples only, and we propose a mini-batch selection algorithm to address the training speed issue in minimizing the CVaR-regularized loss. Numerical tests using real-world data on the IEEE 123-bus test case have demonstrated the computation and safety improvements of the proposed risk-aware learning algorithm for decentralized DER decision making, especially in terms of reducing feeder voltage violations. • Scalable NN architecture for decentralized DER operations in real-time. • Risk-aware learning by using CVaR to systematically improve worst-case performance. • Improving the safety of DER decision rules by using the CVaR regularization. • Accelerating the training speed by proposing a statistical based mini-batch selection. • Numeric tests demonstrate computation and safety improvements over MSE based learning. [ABSTRACT FROM AUTHOR]

This paper first reviews today's hierarchical control and highlights modelling and control assumptions which may no longer hold in the changing electric energy systems. To overcome this fundamental problem, we conceptualise electric energy systems as complex dynamical systems using physically intuitive multi-layered energy modelling as the basis for systematic diverse technology integration and control. Such modelling exhibits unique structure which comes from the conservation of instantaneous power (P) and the rate of change of instantaneous reactive power (Q ̇). We name P and Q ̇ as the rate of change of interaction variables (intVar) for characterising the interface of each system module with the rest of the system. The intVars are used as a means to model and control the interactive zoomed-out inter-modular (inter-area, inter-component) system dynamics. The control is distributed and modular as it utilises only a low-level detailed technology-specific model and intVar info exchange with the neighbours. As a result, the proposed approach can be used to support orderly and natural evolution from today's hierarchical control to flexible interactive protocols for electricity services, key to implementing decarbonisation. • T&D system needs to be operated in a flexible data-enabled way in order to manage uncertainties in an efficient manner. • Efficient energy service requires temporal, spatial and functional alignment of energy resources and demand. • Models of interactions subject to conservation of instantaneous power and reactive power dynamics can be used to formulate optimization problems in the energy space as convex linear problems. [ABSTRACT FROM AUTHOR]

Inverter-based distributed energy resources provide the possibility for fast time-scale voltage control by quickly adjusting their reactive power. The power-electronic interfaces allow these resources to realize almost arbitrary control law, but designing these decentralized controllers is nontrivial. Reinforcement learning (RL) approaches are becoming increasingly popular to search for policy parameterized by neural networks. It is difficult, however, to enforce that the learned controllers are safe, in the sense that they may introduce instabilities into the system. This paper proposes a safe learning approach for voltage control. We prove that the system is guaranteed to be exponentially stable if each controller satisfies certain Lipschitz constraints. The set of Lipschitz bound is optimized to enlarge the search space for neural network controllers. We explicitly engineer the structure of neural network controllers such that they satisfy the Lipschitz constraints by design. A decentralized RL framework is constructed to train local neural network controller at each bus in a model-free setting. • A safe learning approach for voltage control in power distribution systems. • Establishing Lipschitz constraints to guarantee exponentially stabilizing controllers. • Designing neural networks that satisfy the Lipschitz constraints by construction. • A decentralized RL framework for training these neural network controllers. [ABSTRACT FROM AUTHOR]

• In addition to load compensation, UPQC is enhanced to supply reactive power to grid. • The reactive power supplied is flexible as per the requirement from the grid. • UPQC-DG can participate in providing ancillary service in smart grid scenario. • Reactive power is regulated using additional PI controller in UPQC-DG control. • Power angle control is used to reduce additional VA burden on UPQC-DG. Unified Power Quality Conditioner with Distributed Generation (UPQC-DG) is combined equipment that not only compensates power quality issues but also integrates distributed generation into the grid. This paper proposes to incorporate additional functionality in UPQC-DG to provide reactive power support to the grid. The proposed functionality enables the UPQC-DG to not only compensate for load reactive power but also to export it to the grid in a controllable and regulated manner. The control method of the proposed UPQC-DG system is based on a combination of Synchronous Reference Frame (SRF) theory and Unit Vector Template Generation (UVTG) and utilizes an additional PI controller to control the reactive power flow to the grid. Since the UPQC-DG in the proposed research handles an increased quantity of reactive power, the work incorporates the Power Angle Control (PAC) method to share reactive power burden between series and shunt Active Power Filters (APFs) of UPQC-DG for reducing the overall rating. The proposed UPQC-DG system is validated using exhaustive real-time simulations in Opal-RT for steady-state and dynamic performance. [ABSTRACT FROM AUTHOR]

• Two-timescale volt/var optimization that utilizes distributed energy resources. • Coordinating distributed energy resources along with utility-owned assets. • Engaging utility operator in the decision-making process. • Active distribution network with high distributed energy resources penetration. • Incorporate all resources and assets for enhanced network performance. The high penetrations of distributed energy resources of renewable energy sources, battery energy storage systems, and electric vehicles introduced several challenges to power networks. It can lead to high voltage variation, sudden over/under-power generation, high losses issues, and negatively impact distribution assets. Thus, there is a vital need for volt/var optimization schemes that integrate conventional volt/var devices with inverter-interfaced resources to overcome these challenges. This paper demonstrates a performance assessment of the volt/var optimization scheme utilizing different utility-owned assets and resources based on a two-timescale multi-objective algorithm. At the slow timescale, it utilizes capacitor banks and on-load tap changers to minimize system losses and maximize asset lifetime based on their conditions. The algorithm incorporates the utility operator's direct input for setting the conventional volt/var devices. The operation of these devices is optimized, in three steps, by limiting the number of switching operations based on the operator preferences. At the faster scale, it considers renewable energy sources, electric vehicle charging stations, and battery energy storage systems to be optimized to limit voltage variation and system losses further. The IEEE 33-bus distribution network is modified and used to demonstrate the effectiveness of the algorithm. [ABSTRACT FROM AUTHOR]

• We propose a multi-layer scheme for the power balancing of hybrid AC-DC distribution grids. • The grid is assumed to be divided in areas, named clusters, controlled by decentralized MPC controllers, to locally and promptly compensate active power imbalances. • A supervisory layer manages the power exchanges among clusters to ensure that they have enough power margins to operate independently. • An Optimal Reactive Power Flow is periodically performed to keep voltages and currents within regulatory limits, and to minimize power losses. • The proposed approach has been tested on a benchmark system consisting of the IEEE 37- bus and IEEE 13-bus networks interconnected by a multi-terminal DC grid, witnessing promising results. In this paper, the flexibility of hybrid AC-DC distribution networks is exploited to coordinate multiple Distributed Energy Resources (DERs) with the aim of promptly restoring unexpected power imbalances caused by intermittent Renewable Energy Sources (RESs) and loads. Given the potential large-scale nature of the problem, the AC distribution network is decomposed into non-overlapping areas named clusters, equipped with MicroGrids (MGs) and non-dispatchable units, and interconnected also by the DC network. Each cluster is endowed with a Model Predictive Controller designed to compensate the local active power variability by requesting balancing services to the local MGs. A supervisory layer is designed and activated to optimally transfer power through the controllable DC links guaranteeing enough operative margins to each cluster. The designed architecture is tested on a benchmark grid composed of the IEEE 37-bus and 13-bus systems, connected by a multi-terminal DC network. The reported numerical results witness the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

• A novel MI-QCQP for optimizing reactive power from domestic-scale DERs is proposed • The approach considers both inverter and network losses, and load-voltage sensitivity • Unbalanced European and North American networks are studied, with and without taps • Total demand reduction, of up to 0.5%, depends strongly on inverter loss coefficients • Control with reduced communication requirements can realize 98% of potential benefits This paper proposes a method for studying the potential benefits of reactive power control by domestic-scale inverters, considering network losses, inverter losses, and load-voltage sensitivity. The model is developed as a mixed-integer quadratically constrained quadratic program (MI-QCQP), using a linearization of the unbalanced distribution load flow equations. Networks both with and without on-load tap changers are studied, with the test cases covering both European and North American-style circuits. The use of domestic inverter control is shown to increase benefits compared to conventional tap control by 20%, reducing the total feeder power draw by as much as 0.5% of the feeder load. In contrast, the minimization of either load or losses in isolation is shown to increase the amount of power that a feeder draws in five of the seven circuits. To reduce the communications overhead of the approach, a control scheme is proposed that specifies the reactive power of all inverters on each phase identically; this approach is shown to realise up to 98% of the potential benefits of inverter reactive power control. [ABSTRACT FROM AUTHOR]

• Competitive Volt-VAr optimization of transmission and distribution systems is proposed. • An efficient decentralized algorithm based on the decomposition of boundary bus variables is proposed. • Sub-problems are formulated using linear or convex relationships. • The case of unbalanced distribution system is also investigated. A more efficient Volt-VAr optimization (VVO) for transmission and distribution systems can be realized with considering the interaction between distribution system operators (DSOs) and transmission system operator (TSO) and using flexibility resources such as Distributed Generations (DGs) with high penetration in distribution systems. In this paper, a competitive decentralized framework is proposed for interactive VVO of transmission and distribution systems, in the presence of high penetration of distributed energy resources (DERs). In the proposed scheme, while TSO and DSOs are self-interested and individually solve VVO sub-problem according to their own objectives and constraints, they are interactive and receive feedback from each other to understand the mutual effects of their decisions and give an appropriate response. Each optimization sub-problem minimizes power losses of its own network for each load level subjected to the network and reactive power resources constraints. The defined interactive VVO problem is solved by a new decentralized algorithm based on the decomposition of boundary bus variables between transmission and distribution systems. Furthermore, some heuristic techniques are applied to convert the sub-problems formulation to a linear or at least convex one. Some case studies on IEEE 14-bus test system connected to three IEEE 69-bus distribution systems and IEEE 118-bus test system connected to thirty IEEE 69-bus distribution systems confirm the efficiency of the proposed framework in view of power loss, voltage profile and reactive power generation of transmission and distribution networks in comparison with isolated and interactive cooperative methods. [ABSTRACT FROM AUTHOR]

• Under high penetration levels, uncoordinated rooftop PVs may negatively affect node voltages. • An analytical optimization model is proposed for control of PV resources. • Centralized control of PV and other voltage regulating devices allows for higher PV penetration levels. • Probabilistic assessment of the impact of PV on the distribution grid. High penetration of rooftop photovoltaic (PV) panels in the distribution grid can affect the node voltages, and under light loading conditions, may even result in reverse power flow and localized overvoltage instances. While voltage control at the distribution grid level has been traditionally addressed using voltage and var control (VVC) techniques, with the introduction of PV resource as well as active demand, the problem needs to be reformulated as voltage, var and watt optimization (VVWO). A mixed-integer nonlinear multi-objective optimization solution has been proposed in this paper for the control of active and reactive power resources in an unbalanced distribution grid with high PV penetration. The goal of the proposed methodology is to simultaneously minimize line losses, active power curtailment of PVs, demand curtailment, and operation instances of voltage regulating transformers and shunt capacitors. Due to the stochastic nature of load and solar irradiance, at each PV penetration level the problem is solved probabilistically by using random samples for network loading levels and cloud coverage in the sky. A case study is presented using a modified version of the IEEE 123-bus test distribution feeder. Simulation results indicate that if voltages and reactive powers are optimally controlled, high PV penetration levels can be achieved at the distribution grid without any complications. It is also observed that allowing PVs to exchange reactive power with the grid can significantly affect the operation of voltage regulators and shunt capacitors. [ABSTRACT FROM AUTHOR]