مجله مهندسی برق و الکترونیک ایران
سال بیست و یکم شماره 4 (زمستان 1403)

In recent years, due to the development of distributed generation, the concept of a microgrid has been introduced. A microgrid is a small-scale grid that includes renewable energy sources such as wind turbines and solar energy, conventional energy sources such as microturbines, energy storage units such as batteries, and loads. The microgrid can operate connected to the utility or isolated from the utility. The microgrid can be grouped into AC and DC types. A hybrid microgrid is composed of an AC microgrid, a DC microgrid, and an interlinking converter between them. To optimally operate the linking converter, it should be properly controlled. In this paper, a method to control and manage the interlinking converter is proposed. In the proposed algorithm, the voltage and frequency at the terminal on the AC side, as well as the voltage at the terminal on the DC side, are measured to evaluate the load of both microgrids. Based on the proposed strategies, power transfer occurs between AC and DC microgrids in a manner that maximum efficiency is obtained. The proposed strategy is implemented in MATLAB/Simulink to demonstrate its effectiveness.

Improvements in power electronic technology and devices cause DC microgrids (DC MGs) to be utilized in which the integration of DC distributed generation (DG) sources and loads into the power system is facilitated. Since the power sharing between DG sources and DC bus voltage regulation are two important control objectives in these microgrids, a two-layer hierarchical control strategy is implemented in this article to achieve these goals. The traditional drop control method is employed in the primary layer, but since it is not able to achieve these goals simultaneously, therefore, the proposed secondary control layer which is based on the proportional-integral cooperative distributed averaging (PICDA) secondary control strategy is introduced to overcome the challenges and achieve accurate voltage regulation. The proposed PICDA strategy has high flexibility due to the lack of prior knowledge of the grid topology and has a high convergence speed in reaching the control reference point. An islanded DC MG is simulated in MATLAB/Simulink software environment to verify the effectiveness of the proposed method, and the results of the proposed secondary control strategy are compared and analyzed with another distributed method under different operating scenarios, and the outperformance of the PICDA control method is confirmed.

With the increasing penetration of renewable energy sources (RES) in power systems, small-signal stability (SSS) is challenged due to the decreasing inertia. This paper proposes a sequential generation rescheduling model considering a generic dynamic model for RES, formulated as an optimal power flow (OPF) with SSS constraints (SSS-OPF) to improve the system damping ratio. The dynamic modeling of wind turbine generators, photovoltaic sources, and energy storage systems is possible using this general dynamic model. The sensitivity of eigenvalues with respect to active and reactive powers of generating units has been used to describe the SSS constraints. By employing the semi-definite programming (SDP) relaxation technique, the SSS-OPF model is converted to a convex optimization model. This leads to enhanced convergence reliability and improved computational efficiency in solving the model, rendering the proposed method suitable and applicable for large-scale power systems. This optimization model can be solved using the SDPT3 solver in MATLAB software. Case studies on the IEEE 9-bus and IEEE 39-bus systems are conducted to validate the proposed algorithm. The proposed generation re-scheduling method increases the damping ratio of the power system with a high penetration of RES.

With the increasing penetration of distributed generation resources (DERs) in distribution networks, the strategic placement of these resources has garnered increased attention. Due to the uncertain power generation of renewable based DERs, it is necessary to use stochastic models. Additionally, in recent years, due to the rise in the occurrence of weather-related disasters such as hurricanes and floods, leading to serious damages to power grids, the necessity of considering methods to mitigate these damages in installation planning has become crucial. This study introduces a comprehensive model for optimizing the deployment of distributed renewable (solar and wind) and non-renewable (diesel generator) generation resources, accompanied by energy storage systems (batteries), network reinforcement, and equipment upgrades to enhance its resilience against two specific weather events: hurricanes and floods. The model is applied to the IEEE standard 33-bus network. The planning problem is defined as a costs function while considering the environmental constraints. The optimization problem is structured as a probabilistic programming model and solved using the CPLEX solver in the GAMS software. The results obtained from case studies on the IEEE standard 33-bus network demonstrate the efficiency of the proposed model in minimizing costs and increasing system resilience.

In recent years, with increasing application of electric vehicles, studies about fast-charging stations have received attention. One of the most important factors in the development of electric transportation is the availability of fast-charging stations. This paper examines the sizing and siting of fast-charging stations while taking into account how they will affect the traffic flow of urban transportation networks and the power quality indexes of the distribution network. On the other hand, the optimal usage of the reactive power compensation and active filtering capabilities of fast-charging station have been considered. In the proposed method, the harmonic power flow, staircase cost facility location model (SCFLM), and user-equilibrium traffic assignment for modeling constraints of distribution and transportation networks are used. The proposed model aims to reduce traveling time in the transportation network by considering a penalty cost for incremental road traffic congestion in the objective function. Also, the proposed model aims to improve the power quality of the distribution network. The proposed integer linear programming model is implemented in GAMS software and finally, the effectiveness of the proposed method is shown using numerical results.

The purpose of this paper is to introduce a framework for the simultaneous optimal placement of battery charging and swapping stations for electric vehicles in centralized charging mode. In the centralized charging mode, unlike the decentralized mode, the charging equipment is located in a place other than the battery swapping station, which is called the central charging station. The batteries are charged in this place and are distributed among the battery swapping stations on a regular basis. The objective functions of the presented model include minimization of the batteries transportation cost, improving the voltage index, and reducing the cost of distribution grid losses. For this purpose, a nonlinear multi-objective model is presented to calculate the objective function values of the problem. Non-dominated sorting genetic algorithm (NSGA-II) is used to solve the optimization problem in this article. The output of the genetic algorithm is a set of pareto optimal points, from which the final solution is selected using fuzzy decision-making method. Finally, the model is implemented on the 69-bus IEEE test system, and the results have been examined in different scenarios. The obtained results verify the efficiency of the presented model.

Modern voltage control strategies in the present distribution networks require efficient equipment as well as appropriate communication channels between these equipment, sensors and control centers, which has led to smart distribution networks with a complex cyber-physical nature. One of the efficient equipment for voltage control in modern distribution networks is the DSTATCOM, which uses using multilevel converters in its structure, provides many advantages, such as direct connection to the grid. A DSTATCOM with the multilevel converter requires a cyber-physical network between the controller and its components due to the presence of many controllable components, which makes it vulnerable to cyber-attacks, when connected to the present smart distribution networks. In this paper, a feedback linearization-based controller is developed for the cascaded multilevel DSTATCOM, and a discrete Kalman filter-based method is proposed to detect and compensate false data injection cyber-attacks on voltage sensors of the multilevel converter. The abilities of the proposed nonlinear controller to control the multilevel DSTATCOM as well as the reliable operation against false data injection attacks are verified through the simulation of a test power network in the MATLAB/Simulink environment.

In this paper, a new designing procedure is proposed for the static ground power unit of airplanes. The proposed structure is generally made up of an input and output section, introduced as the dc and ac sides. In order to reach a superior rectifying operation, both six pulse rectifier and injection of third harmonic current have been considered at the dc side of the proposed ground power unit. In the ac side, a neutral point clamped inverter has been implemented to reach a high performance output required by the GPU. Every single part of the proposed configuration has been discussed, designed and presented including third harmonic injection method, inductors, dc link capacitors, the output filter and the isolation transformer. The proposed ground power unit has been simulated in Matlab/Simulink environment. The simulation results confirmed its superior performance by providing a standardized ac voltage at the output for GPU applications.

Publicly releasing local electricity markets data brings a multitude of economic as well as technical and societal benefits. Moreover, public access to these data is instrumental for achieving transparency and more competition in local electricity markets. However, privacy-aware market participants have concerns about the leakage of their private information via releasing of the local electricity market outputs. This paper aims to design a mechanism for local electricity markets that provably guarantees the privacy of market participants and also reflects their privacy preferences by implementing differential privacy. First, the required randomization for achieving differential privacy is embedded in the optimization process of the market-clearing problem via noisy gradient ascent algorithm. Then, for providing personalized level of privacy protection, a subsampling mechanism over the input dataset of the market-clearing problem. In numerical studies, the inherent trade-off between the privacy protection and social welfare in the market is investigated under different privacy regimes.

Power system stability, control, and design studies are time-consuming due to the formation of large size and heavy interconnection networks. Reduced ordered dynamic equivalent methods are thus so desirable for performing these studies. Simple equivalents are obtained by converting large and complex networks into smaller networks. One of the approaches to the problem of model reduction is to find coherency-based generator grouping and aggregation. In this method, first, based on network characteristics, each group of coherent generators is replaced with a dynamic equivalent, and then the dynamic equivalent of the generators is used in the studies of power networks. In this regard, in this article, using graph theory and clustering quality index, a novel method for finding coherent generators in a power network is presented. The proposed method is simple and the network admittance matrix is the only information required by this approach. The IEEE 39-bus network is used to show the effectiveness of the proposed approach. The comparison of the results with other research shows that the proposed method identifies the coherent generators with an acceptable approximation.

In this paper, a new method based on long-short-term memory neural networks is proposed for predicting long-term voltage instability in interconnected transmission and distribution networks. Using the online information of the phasor measurement units, the neural network estimates the stability of the network voltage and in case of any event in the network, it estimates the long-term voltage stability status using the information before and after the event. This structure can be used as an auxiliary tool to quickly inform the network operator of possible risks caused by voltage instability after any contingency in the network. The simulation results of different case studies in offline-mode have been used to create the training dataset. In order to have different case studies, considering load growth in areas prone to voltage instability, (N-1) and (N-1-1) contingency have been simulated. Nordic extended network has been used to evaluate the proposed neural network performance. By using the appropriate time shift, the occurrence of all contingencies has been moved to the tenth second so that the neural network only learn the trajectory of the features. The accuracy of the neural network in the 17th second (7 seconds after contingency) is 99.05%. Finally, the effect of reducing the input dimensions by clustering the data has been investigated.

In this article, a new analytical method for the optimal allocation of active and reactive powers of distributed generation sources (DG) and capacitor banks in distribution networks is presented. The method of doing the work is very simple and lacks the complexity of other methods. The objective functions used in this study include minimizing power losses and improving the voltage profile of the distribution network. To show the effectiveness of the proposed method, simulation results have been performed on 33 distribution networks and its performance has been compared with global search methods. Finally, the presented approach for allocating the optimal power coefficients of DGs and capacitor banks has also been used in the distribution network of 119 buses. The simulation results show that the analytical approach presented in finding optimal results is very strong, fast and easy and can be used on wide distribution systems.

Fast and selective protection brings more security, reliability and stability in power systems. Accomplishing this task by deploying directional overcurrent relays is one of the main challenges in the active distribution network. Due to the presence of two-level fault currents, conventional methods fail to meet selective protection requirements. Tackling this issue renders the need for using the dynamic model for relays operation, which may result in miscoordination or increase the operation time of relays in large scale networks. Therefore, in this article, a new protection coordination approach is presented based on employment of the dynamic model of relays operation and using a new variable K in this model. The presented variable simplifies the optimization space of coordination problem and provides fast-response protection. On the other hand, new coefficients in the operation model of relays may cause mal-operation in normal conditions. Therefore, a new logic is also devised that can be implemented based on numerical relays. Since this logic requires to employ numerical relay, the proposed method is modeled techno-economically to provide fast and selective protection with the minimum number of numerical relays. This model is formulated as a multi-objective optimization problem and TOPSIS algorithm is used as a decision-making approach. Results are discussed in depth.

High-voltage transmission lines (HVTLs) are one of the important parts of power systems exposed to different types of fault. As most of the faults that occur in HVTLs are single line to ground (SLG) and transient type, the operating speed of the single-pole auto-reclosing (SPAR) will play an important role in the stability of power systems. In this paper, an ultra-fast protection strategy is proposed based on the cumulative sum (CUSUM) algorithm to improve the performance of the SPAR. In the proposed protection scheme, first, the terminal voltage of one side of the HVTL is sampled. Then by employing the CUSUM algorithm for the sampled voltage and applying the corresponding settings, the reclosing moment of the faulty phase is determined. The results of extensive simulations in the EMTP-RV software environment on a 400 kV test power system show that the proposed protection scheme has been very accurate in identifying the moment of successful reclosing. The ultra-fast performance and low computational complexity of the proposed method along with its robustness to noise make it useful for practical applications.

Till now, many different methods like heuristic methods, and the method of time difference of arrival (TDoA) have been proposed and implemented for the localization of partial discharge in power equipment, but their weaknesses have made their usage to be limited in practice. Because in complex situations like the presence of noise, scattering, barriers, and inhomogeneous mediums the accuracy of results will be strongly reduced. The new time-reversal method that has been recently proposed in the field of partial discharge localization showed that can accurately localize partial discharge using just a single sensor. In this paper, the implementation of a different version of the time-reversal method in the electromagnetic regime, named inverse filter has been proposed for the localization of partial discharge in GIS. For evaluation of the proposed method, different scenarios are designed and simulated in two pipe-shape and T-shape structures of a GIS in CST-MWS. Results show that in the case of using minimum entropy for determining the moment of refocusing the waves, this method can accurately localize partial discharge in GIS, in all scenarios.