Electrical Engineering最新文献

A novel fuzzy controller based on an input current slope for a three-stage cascaded boost converter is presented for high-voltage DC applications. The control objective is to achieve satisfactory converter performance, facing not only variations in load resistance and renewable energy source voltage but also in the interaction of the individual converter in each stage. The intricate configuration nature of multi-stage cascaded boost converter renders the mathematical model complex, which is perplexing for the design methodology of controllers. In response to these challenging problems, the fuzzy controller can function without a precise mathematical model. It uses an understanding of the converter behavior to develop control rules that imitate human decision making. In addition to controlling the output voltage, the proposed fuzzy controller in this paper can manage the input current along with the reference current slope without overshooting or taking a long time to settle. The input current slope and the output voltage error are determined with suitable membership functions and their intervals. The fuzzy output is a change in duty, linguistically associated with the system inputs through 15 fuzzy rules. The simulations and experimental results indicate that the output voltage of 400 V is finely regulated with robustness. In addition, the results imply that performance performs effectively in both transient and steady states, even when the converter functions under conditions of fluctuating load resistance, input voltage, and desired output voltage.

In the context of a deregulated electricity market, this paper presents a new method for the distribution of electrical losses in distribution systems with local generators, incorporating variations in consumer power factor. Additionally, the paper examines how local generators may gain the utmost possible advantage from their contributions to system performance enhancement. The proposed approach initially analyzes dynamic changes in active and reactive power flow caused by a change in consumer power factor, followed by a loss allocation mechanism. This method is developed based on Kirchhoff’s law and the natural decomposition of injected bundled power flow at a bus. To completely eliminate cross-subsidies, the proposed mechanism follows a two-stage power flow: first, consumer-controlled power flows, and second, determining variations in power flow as compared to the first stage due to interactions with local generators. Consumer load modeling is being considered to improve the proposed method’s adaptability and efficiency across different distribution systems. The effectiveness of the proposed method is validated through testing on 30-bus and 69-bus distribution systems, and the results obtained are compared with those of other methods found in the literature.

Effectively utilizing renewable energy sources while avoiding power consumption restrictions is the problem of demand-side energy management. The goal is to develop an intelligent system that can precisely estimate energy availability and plan ahead for the next day in order to overcome this obstacle. The Intelligent Smart Energy Management System (ISEMS) described in this work is designed to control energy usage in a smart grid environment where a significant quantity of renewable energy is being added. The proposed system evaluates various prediction models to achieve accurate energy forecasting with hourly and day-ahead planning. When compared to other prediction models, the Support Vector Machine (SVM) regression model based on Particle Swarm Optimization (PSO) seems to have better performance accuracy. Then, using the anticipated data, the experimental setup for ISEMS is shown, and its performance is evaluated in various configurations while considering features that are prioritized and user comfort. Furthermore, Internet of Things (IoT) integration is put into practice for monitoring at the user end.

In the current paradigm, integration of distributed generation (DG) has become essential for ensuring the quality, reliability, and security of distribution network operations. Existing literature typically formulates multi-objective problems to quantify the techno-economic assessment of DG placement in a distribution system. Nevertheless, there is a notable lack of a suitable method to assign weight impartially to each objective for optimal decision-making. This paper introduces a novel technique for strategically placing DG units in the distribution network, employing weights calculated based on the importance level of various techno-economic objectives using Shannon’s entropy. The proposed approach has been applied to a 38-node test system to illustrate its efficacy. The numerical findings from four distinct case studies reveal that changes in the physical attributes of the system correspondingly influence the significance of objectives in determining the optimal placement and size of DG. The results show significant reductions in active and reactive power losses and total annualized operational costs, with maximum reductions of 48.17%, 33.30%, and 42.96%, respectively. The minimum voltage magnitude improves from 0.9252 pu in the base case to 0.9384, 0.9695, 0.9369, and 0.9348 for Cases 1, 2, 3, and 4, respectively. Moreover, a comparative statistical analysis underscores the superiority of the proposed method over prevailing weight allocation strategies by achieving a 3.59% reduction in annual expenditure, while maintaining competitive network performance metrics in addressing the multi-objective DG placement problem.

This paper presents the influence of changes in the density of deep (Gaussian) defects, their energetic position, and width on key electrical parameters, including threshold voltage, current on–off ratio, and maximum transconductance in TIPS-pentacene-based organic thin-film transistors (OTFTs). Due to intrinsic disorder, organic semiconductors function with a Gaussian density of states governing the movement and injection of charge carriers within these materials. Our study reveals the presence of deep acceptor and donor density of states within the band gap of the TIPS-pentacene can significantly affect the performance of OTFTs. When the Gaussian acceptor ((N_{textrm{GA}})) value is (1times 10^{15},{textrm{cm}}^{-3},{textrm{eV}}^{-1}), the current on–off ratio ((I_{textrm{on}}/I_{textrm{off}})) is at its peak, reaching (2.3times 10^7), and the mobility is notably high at (0.0270, {textrm{cm}}^{2},{textrm{V}}^{-1},{textrm{S}}^{-1}). In the case of the Gaussian donor ((N_{textrm{GD}})) with a value of (1times 10^{17},{textrm{cm}}^{-3},{textrm{eV}}^{-1}), the current on–off ratio ((I_{textrm{on}}/I_{textrm{off}})) reaches its peak at (7.9times 10^7), and the lowest threshold voltage ((V_{textrm{th}})) is at 1.26 V. For the acceptor-like Gaussian decay energy ((W_{textrm{GA}})) with a value of 0.1 eV, the current on–off ratio ((I_{textrm{on}}/I_{textrm{off}})) peaks at (2.4times 10^5). The dynamic control of charge trapping in this context holds the potential for various applications, including memory-related functions and the emulation of neurons in neuromorphic circuits for deep learning and artificial intelligence.

In high-voltage direct current transmission systems, the unloaded energizations of converter transformers may cause magnetizing inrush currents with high amplitude and slow decaying, as well as the zero-sequence inrush current in the neutral line. This may cause the maloperation of saturation protection based on the amplitude of the neutral line current. Aiming at this problem, a scheme for preventing the maloperation of converter transformer saturation protection based on the waveform prediction of magnetizing inrush current is proposed. The proposed scheme utilizes the differences between the mathematical models of pure magnetizing inrush and magnetizing inrush superimposed with DC bias. Based on the mathematical analytical expression of magnetizing inrush current during the transformer unloaded energization, the complete waveform of the second cycle of phase current is predicted by using the actual measured values of the phase current within 1¼ cycles. The deviation between the predicted and measured peak values of the phase current of the second cycle is calculated. On this base, a modification coefficient is constructed to modify the measured current in the neutral line, which eliminates the influence of magnetizing inrush and prevents the maloperation of saturation protection. Simulation results indicate that the proposed scheme can separate the influences of magnetizing inrush and DC bias on the saturation protection, effectively preventing the maloperation of saturation protection caused by magnetizing inrush.

Acetylene is one of the main fault gases for oil transformers. The rapid and highly sensitive detection of dissolved acetylene in the insulating oil plays an extremely important role in the diagnosis of transformer faults, as it is produced by the decomposition of hydrocarbons due to discharge and overheating. This work describes a rapid real-time online monitoring system for dissolved acetylene in oil, which integrates a highly efficient Teflon-AF2400/ceramic composite degassing module and a high-sensitive laser photoacoustic detector. Real-time online monitoring is feasible as the detection period of the device is as short as 1.5 min, and the equilibrium concentration of acetylene in the oil can be accurately determined from test data at a degassing time of 15 min. When the concentration of acetylene in the oil changes suddenly, the device can report more than 90% of the change within 30 min. The detection accuracy is improved from 0.9 to 0.3 μL L⁻¹ after corrections are made to account for the influence of temperature on the oil–gas separation membrane.

Partial discharge (PD) induces degradation in filament-like channels of electrical trees (ETs), leading to a significant reduction in material strength and lifespan. To enhance the assessment of insulation status in polyethylene (PE) based on PD characteristics during ET growth, a dynamic health monitoring evaluation model is developed. This model enables early critical warnings for insulation states. Findings reveal that during the rapid growth phase of ETs, the Dynamic Health Index of PE fluctuates within the range of [62%, 11%], with a failure rate reaching 54.30%. The transition point from the retardation stage to the rapid growth stage is identified as the critical warning threshold for insulation failure based on PD characteristics during ET growth. Utilizing linear fitting slopes of apparent discharge magnitude and ultra-high frequency amplitude in the time domain, the frequency bandwidth of ultra-high frequency signals, and average energy distribution facilitates effective differentiation between the retardation and rapid growth stages of ETs. This work establishes a robust foundation for developing an intelligent insulation state evaluation system, thereby enhancing the reliable operation of insulation systems.

Modeling and simulating photovoltaic (PV) cells or modules involve using mathematical and computational models to predict their behavior and performance under various conditions. This can include modeling the electrical characteristics of solar cells, as well as the interactions between multiple cells in a PV module. In ISIS-Proteus software, the existing research works have modeled the PV modules either by using a Proteus Spice model of the PV panel without including the effect of climatic conditions variation or by using pure mathematical relations that describe all physical and environmental parameters that lead to a static behavior. Therefore, this paper proposes a new improved ISIS-Proteus model of a PV cell/module for dynamic performance emulation under varying climatic conditions. The proposed model is designed based on the equivalent circuit of a five-parameter single-diode as an electrical part controlled by a numerical part that includes the mathematical expressions corresponding to each parameter. The designed model can capture the impact of solar irradiance and temperature on PV outputs, thereby enhancing real-world PV performance prediction. Also, it can effectively simulate the effect of the partial shading. To validate the accuracy of the proposed model, a comparative study is conducted evaluating the model's performance against PVsyst software models and real-world data brought from a large-scale grid-connected PV station in Ain El-Melh, Algeria. In this study, the simulation tests are carried out using ISIS-Proteus considering several PV module types and under various operating conditions, including uniform test conditions (UTCs) and partial shading conditions (PSCs). The findings, including I–V and P–V curves and several standard metrics, prove the proposed model's effectiveness in accurately predicting the behavior of PV modules under both UTCs and PSCs, aligning closely with real-world performance.

Because of the high current required and the fact that charging stations introduce limits and concerns into the public grid at different times and in different locations, electric vehicles are becoming increasingly popular. Short charging times for electric vehicles (EVs) due to inefficient EV charging infrastructure are the main obstacles to their expansion. This paper proposes a hybrid technique for enhancing electric vehicle charging stations in DC microgrid. The proposed hybrid approach is a combination of both dilated residual network (DRN) and Kepler optimization algorithm (KOA). Hence, it is named as KOA–DRN technique. The main objective of the proposed method is minimizing the total energy loss and charging time. The SAO algorithm is utilized to optimize the charging process, ensuring efficient and optimal use of available resources, and DRN is utilized to provide intelligent control and decision-making capabilities to the EV charging station. The proposed method is executed in the MATLAB and is compared with different existing methods like wild horse optimization (WHO), heap-based optimization (HBO), and particle swarm optimization (PSO). The peak PV power is 11 W; peak grid current is − 195 to 190 in 2 s. DC load voltage is 4.1 W. The proposed approach KOA–DRN obtains loss value of 1.2% and setting time of 0.02 s, which is less than the existing approaches.