Mohamed A. Hassan

Electric vehicles (EVs) offer substantial environmental benefits but introduce challenges to power grids due to capacity limitations. Effective deployment and coordination of EV Charging Stations (EVCS) are crucial, as improper placement can lead to increased power losses and voltage fluctuations. This paper presents a comprehensive approach to address these challenges by proposing a deep learning-based, two-tier optimization strategy for EVCS placement within distribution networks. A hybrid forecasting model that combines autoregressive integrated moving average (ARIMA) and long short-term memory (LSTM) techniques is employed to accurately predict EV charging load. The ARIMA model captures seasonality and transient patterns, while the LSTM model identifies long-term dependencies, enhancing forecasting accuracy. Additionally, the paper introduces a Hybrid Archimedes-Genetic Algorithm (HAG), which merges the Archimedes Optimization Algorithm (AOA) with the Genetic Algorithm (GA). This multi-objective approach aims to minimize both active and reactive power losses and control voltage deviations. Tested on IEEE 33-bus and 69-bus systems using MATLAB and Python, the HAG method demonstrated significant improvements, with reductions in active losses of up to 33% and 10.7%, and reactive losses of up to 33% and 10.8% compared to GA and AOA alone. These findings underscore the effectiveness of combining accurate EV charging load forecasting with advanced optimization techniques to optimize EVCS placement while minimizing network impact.

The integration of inverter-interfaced sources into microgrids has led to new protection challenges due to fluctuating fault currents, bidirectional power flow, and low inertia. Traditional protection methods are inadequate for these evolving systems, necessitating the development of novel, fast-acting, and cost-effective strategies that are independent of fault current levels. Machine learning-based protection schemes offer potential for improved performance over traditional methods, but their success is highly dependent on the quality and nature of the input data. Raw fault signal data are often noisy and high-dimensional. Processing it without feature extraction or selection can result in poor performance, longer processing times, higher storage needs, overfitting, and reduced generalizability. The dimensionality reduction step, which encompasses feature extraction and/or feature selection algorithms, is crucial for machine learning-based protection schemes. Many such methods have been proposed in the last decade. This research aims to investigate available techniques for reducing the dimensionality of data used for fault diagnosis in AC microgrids. The research will review and compare different dimensionality reduction methods, highlighting their strengths and weaknesses. It will also explore the most used features and their effectiveness in different scenarios. This article highlights the key gaps in AC microgrid fault diagnosis, which include scalability, real-time performance, and generalization, among many others. It emphasizes that future research should focus on unexplored dimension reduction techniques, hybrid approaches, and improving deep learning interpretability for robust and accurate solutions.

AC microgrids are becoming increasingly important for providing reliable and sustainable power to communities. However, the evolution of distribution systems into microgrids has changed the way they respond to faults and hence their protection requirements. Faults in microgrids could hinder operation stability and damage the system components. The types, locations, and resistances of faults, as well as microgrid operation modes, distributed generation penetration levels, load changes, and system topologies, all affect how the microgrid responds to faults. In order to offer quick restoration and to protect the microgrid components, fault detection and classification are therefore essential for microgrids. In this direction, unconventional methods such as artificial intelligence have been increasing in popularity over the last years. Pattern recognition is a methodology that machine learning as an approach to artificial intelligence is concerned with. The combination of protection with machine learning may be motivating in order to achieve the goal of intelligent operation in the smart grid. In this paper, fault detection, classification and location methods are reviewed for microgrid application. Different methods applied for both fault location and fault classification are being classified by the implemented technique. Such methods are explained and analyzed providing the main advantages and disadvantages of each category. Additionally, the research trends in both fields are analyzed and state–of–the–art methods from each category are thoroughly compared. Finally, the research gaps and future directions are identified.

Accurate load forecasting is of utmost importance for modern power generation facilities to effectively meet the ever-changing electricity demand. Predicting electricity consumption is a complex task due to the numerous factors that influence energy usage. Consequently, electricity utilities and government agencies are constantly in search of advanced machine learning solutions to improve load forecasting. Recently, deep learning (DL) has gained prominence as a significant area of interest in prediction efforts. This paper introduces an innovative approach to electric load forecasting, leveraging advanced DL techniques and making significant contributions to the field of energy management. The hybrid predictive model has been specifically designed to enhance the accuracy of multivariate time series forecasting for electricity consumption within the energy sector. In our comparative analysis, we evaluated the performance of our proposed model against ML-based and state-of-the-art DL models, using a dataset obtained from the Distribution Network Station located in Tetouan City, Morocco. Notably, the proposed model surpassed its counterparts, demonstrating the lowest error in terms of the Root-Mean-Square Error (RMSE). This outcome underscores its superior predictive capability and underscores its potential to advance the accuracy of electricity consumption forecasting.

To operate photovoltaic (PV) systems efficiently, the maximum available power should always be extracted. However, due to rapidly varying environmental conditions such as irradiation, temperature, and shading, determining the maximum available power is a time-varying problem. To extract the maximum available power and track the optimal power point under these varying environmental conditions, maximum power point tracking (MPPT) techniques are proposed. The application of MPPT for extracting maximum power plays a crucial role in developing efficient PV systems. These MPPT techniques face several issues and limitations, particularly during partial shading conditions caused by non-uniform environmental conditions. Researchers have been focusing more on mitigating the partial shading condition in PV systems for the last few years due to the need to improve power output and efficiency. This paper provides an overview of MPPTs proposed in the literature for uniform and non-uniform environmental conditions broadly categorized as MPPT-based and circuit-based methods. The MPPT-based methods are classified as conventional, soft computing, and hybrid techniques. A critical analysis of each approach regarding tracking speed, algorithm complexity, and dynamic tracking under partial shading is discussed. The literature shows hybrid strategies provide fast-tracking speed and are efficient with a tracking efficiency of around 99% compared to conventional methods; however, their design and practical implementation are complex. This comprehensive review of MPPT methods aims to provide power utilities and researchers with a reference and guideline to select the best MPPT method for normal operation and partially shaded PV systems based on their effectiveness and economic feasibility.

Microgrids that are integrated with distributed energy resources (DERs) provide many benefits, including high power quality, energy efficiency and low carbon emissions, to the power grid. Microgrids are operated either in grid-connected or island modes running on different strategies. However, one of the major technical issues in a microgrid is unintentional islanding, where failure to trip the microgrid may lead to serious consequences in terms of protection, security, voltage and frequency stability, and safety. Therefore, fast and efficient islanding detection is necessary for reliable microgrid operations. This paper provides an overview of microgrid islanding detection methods, which are classified as local and remote. Various detection methods in each class are studied, and the advantages and disadvantages of each method are discussed based on performance evaluation indices such as non-detection zone (NDZ), detection time, error detection ratio, power quality and effectiveness in multiple inverter cases. Recent modifications on islanding methods using signal processing techniques and intelligent classifiers are also discussed. Modified passive methods with signal processing and intelligent classifiers are addressing the drawbacks of passive methods and are getting more attention in the recently published works. This comprehensive review of islanding methods will provide power utilities and researchers a reference and guideline to select the best islanding detection method based on their effectiveness and economic feasibility.

Nowadays, behaving as constant power loads (CPLs), rectifiers and voltage regulators are extensively used in microgrids (MGs). The MG dynamic behavior challenges both stability and control effectiveness in the presence of CPLs. CPLs characteristics such as negative incremental resistance, synchronization, and control loop dynamic with similar frequency range of the inverter disturb severely the MG stability. Additionally, the MG stability problem will be more sophisticated with a high penetration level of CPLs in MGs. The stability analysis becomes more essential especially with high-penetrated CPLs. In this paper, the dynamic stability performance of an MG involving a high penetration level of CPLs is analyzed and investigated. An autonomous MG engaging a number of CPLs and inverter distributed generations (DGs) is modeled and designed using MATLAB. Voltage, current, and power controllers are optimally designed, controlling the inverter DGs output. A power droop controller is implemented to share the output DGs powers. Meanwhile, the current and voltage controllers are employed to control the output voltage and current of all DGs. A phase-locked loop (PLL) is essentially utilized to synchronize the CPLs with the MG. The controller gains of the inverters, CPLs, power sharing control, and PLL are optimally devised using particle swarm optimization (PSO). As a weighted objective function, the error in the DC voltage of the CPL and active power of the DG is minimized in the optimal problem based on the time-domain simulation. Under the presence of high penetrated CPLs, all controllers are coordinately tuned to ensure an enhanced dynamic stability of the MG. The impact of the highly penetrated CPLs on the MG dynamic stability is investigated. To confirm the effectiveness of the proposed technique, different disturbances are applied. The analysis shows that the MG system experiences the instability challenges due to the high penetrated CPLs. The simulation results confirm the effectiveness of the proposed method to improve the MG dynamic stability performance.

In this article, a new soft starting control scheme based on an artificial neural network (ANN) is presented for a three-phase induction motor (IM) drive system. The main task of the control scheme is to keep the accelerating torque constant at a level based on the value of reference acceleration. This is accomplished by the proper choice of the firing angles of thyristors in the soft starter. Using the ANN approach, the complexity of the online determination of the thyristors firing angles is resolved. The IM torque-speed characteristic curves are firstly used to train the ANN model. Secondly, the IM- soft starter system is modeled using MATLAB/SIMULINK. To prove the effectiveness of the proposed ANN-based acceleration control scheme, different reference accelerations and loading conditions are applied and investigated. Finally, a laboratory prototype of 3 kW soft starter is implemented. The proposed control scheme is executed in a real-time environment using a digital signal processor (Model: TMS320F28335). The simulation and real-time results significantly confirm that the proposed controller can efficiently reduce the IM starting current and torque pulsations. This in turn ensures a smooth acceleration of the IM during the starting process. Moreover, the proposed control scheme has the superiority over several soft starting control schemes since it has a simple control circuit configuration, less required sensors, and low computational burden of the control algorithm.