Optimization
Shahpour Rahmani; Nasser Yazdani
Abstract
High-speed rail systems, operating at speeds up to 350 km/h, face significant challenges in delivering reliable network connectivity due to frequent handovers, signal degradation, and network congestion. This paper proposes the 5G-R framework, an optimized solution integrating beamforming, network slicing, ...
Read More
High-speed rail systems, operating at speeds up to 350 km/h, face significant challenges in delivering reliable network connectivity due to frequent handovers, signal degradation, and network congestion. This paper proposes the 5G-R framework, an optimized solution integrating beamforming, network slicing, railway-specific Long Short-Term Memory (LSTM) algorithms, and Multi-Access Edge Computing (MEC) to enhance connectivity performance. By leveraging real-time train data, such as speed and GPS location, the framework optimizes handover prediction and traffic management, achieving robust performance in diverse environments. Compared to 4G LTE and standard 5G, the 5G-R framework demonstrates significant improvements, including a 250 Mbps throughput, 15 ms latency, and 95% handover success rate. Network slicing optimizes resource allocation, reducing congestion by approximately 30%, while MEC enables low-latency control for train systems. Field trials along the Beijing-Zhangjiakou railway (174 km, urban/suburban) and simulations validate the framework’s adaptability across urban and rural routes. Designed for compatibility with the Future Railway Mobile Communication System (FRMCS), the 5G-R framework lays a foundation for future advancements, including 6G and satellite communications. Future research should focus on optimizing performance in extreme environments and densely populated routes to support autonomous transport systems. This optimization-driven approach establishes a scalable model for next-generation rail communication systems.
Optimization
Hamed Maleki; Mohammad Sadegh Sepasian; Mohammad Reza Aghamohammadi; Mousa Marzband
Abstract
This study investigates the optimal design configuration of a hydrogen refueling station located in southern Iran, focusing on the integration of renewable energy sources and seawater desalination technology to achieve self-sufficiency. The station integrates various components, including photovoltaic ...
Read More
This study investigates the optimal design configuration of a hydrogen refueling station located in southern Iran, focusing on the integration of renewable energy sources and seawater desalination technology to achieve self-sufficiency. The station integrates various components, including photovoltaic panels, fuel cells, desalination units, natural gas and power-to-hydrogen conversion systems, and storage facilities for water and hydrogen. The primary goals are to achieve an independent power supply from renewable sources and an autonomous water supply through seawater desalination. To determine the most cost-effective configuration, a Mixed Integer Linear Programming (MILP) model is developed, taking into account the water and power consumption of each component. The objective is to minimize the Net Present Cost (NPC) of investment, maintenance, and operation. The model is implemented and solved using the CBC solver within the PYOMO environment. The study's findings reveal that converting natural gas to hydrogen is more economically viable than power-to-hydrogen conversion, with the former accounting for more than 95% of the hydrogen produced. The power demand is effectively met by a combination of photovoltaic systems, fuel cells, and hydrogen storage. Moreover, the study highlights the benefits of integrating water and hydrogen storage systems, which optimizes the utilization of photovoltaic energy. Excess energy generated by the photovoltaic panels is utilized for seawater desalination and the production of green hydrogen
Optimization
Zobeir Raisi; John Zelek
Abstract
Scene text detection frameworks heavily rely on optimization methods for their successful operation. Choosing an appropriate optimizer is essential to performing recent scene text detection models. However, recent deep learning methods often employ various optimization algorithms and loss functions without ...
Read More
Scene text detection frameworks heavily rely on optimization methods for their successful operation. Choosing an appropriate optimizer is essential to performing recent scene text detection models. However, recent deep learning methods often employ various optimization algorithms and loss functions without explicitly explaining their selections. This paper presents a segmentation-based text detection pipeline capable of handling arbitrary-shaped text instances in wild images. We explore the effectiveness of well-known deep-learning optimizers to enhance the pipeline's capabilities. Additionally, we introduce a novel Segmentation-based Attention Module (SAM) that enables the model to capture long-range dependencies of multi-scale feature maps and focus more accurately on regions likely to contain text instances.The performance of the proposed architecture is extensively evaluated through ablation experiments, exploring the impact of different optimization algorithms and the introduced SAM block. Furthermore, we compare the final model against state-of-the-art scene text detection techniques on three publicly available benchmark datasets, namely ICDAR15, MSRA-TD500, and Total-Text. Our experimental results demonstrate that the focal loss combined with the Stochastic Gradient Descent (SGD) + Momentum optimizer with poly learning-rate policy achieves a more robust and generalized detection performance than other optimization strategies. Moreover, our utilized architecture, empowered by the proposed SAM block, significantly enhances the overall detection performance, achieving competitive H-mean detection scores while maintaining superior efficiency in terms of Frames Per Second (FPS) compared to recent techniques. Our findings shed light on the importance of selecting appropriate optimization strategies and demonstrate the effectiveness of our proposed Segmentation-based Attention Module in scene text detection tasks.
Power systems
Farhad Zishan; Ehsan Akbari; Abdul Reza Sheikholeslami; Nima shafaghatian
Abstract
This paper contributes to the design, modeling, and planning of a distributed generation (DG) network with wind and solar by means of the particle swarm algorithm (PSO) algorithm in the IEEE 33-bus network, aiming to minimize The results indicate an adequate performance in a variety of environments, ...
Read More
This paper contributes to the design, modeling, and planning of a distributed generation (DG) network with wind and solar by means of the particle swarm algorithm (PSO) algorithm in the IEEE 33-bus network, aiming to minimize The results indicate an adequate performance in a variety of environments, and the presence of distributed wind/solar energy generators decreases network stress by feeding loads locally. These systems (wind and solar) can be used in remote areas without power networks, or even in areas where there is a tendency to use renewable energy despite the presence of a power network. They can also supply the output load for most of the day and night. Probability distribution functions are used, and the outputs are expressed as probability density distribution functions instead of absolute numbers. In addition, there is a high degree of uncertainty regarding the state of the system, which is an associated renewable energy source within the power system elements. By means of MATLAB software, the proposed method is implemented in order to ensure effectiveness and validate the results.
Optimization
Davoud Mirzaei; Saeid Amini
Abstract
In this paper, an ultrasonic horn based on the PSO algorithm for emulsion homogenization is optimized and fabricated. The application of various ultrasonic instruments such as horns in different industrial procedures is increasingly expanding and developing. Horn is a tool that has played a crucial role ...
Read More
In this paper, an ultrasonic horn based on the PSO algorithm for emulsion homogenization is optimized and fabricated. The application of various ultrasonic instruments such as horns in different industrial procedures is increasingly expanding and developing. Horn is a tool that has played a crucial role in the energy transfer to fluid. Longitudinal frequency, vibration amplitude, length-to-diameter ratio, a distance of frequency from other frequency modes, wide distribution of cavitation along with the horn’s length, and increasing the area of acoustic energy transfer are the main characteristics of the horn design procedure. Therefore considering these important features with using the PSO algorithm and electro-mechanical circuit method to finding resonance frequency, the design procedure of the optimized horn is performed. The definition of the objective function is based on the horn’s amplification factor, and the rest of the other characteristics are defined as design constraints. The simulation results show a 15% improvement in the natural frequency compared to the target frequency and a suitable frequency distance of 2.5 kHz between the previous and next modes. According to the barbell part of the horn, the amplification factor of 14 was obtained for the proposed horn, the frequency and amplitude of the vibration were evaluated. The experimental results were very close in terms of amplification factor and frequency to the simulation results with reasonable accuracy.
Power systems
Ali Masoudi; Mohsen Simab; Hamidreza Akbarj; Seyed Amin Saeed; Tahereh Daemi
Abstract
With an increasing penetration rate of electric vehicles in distribution networks, it is becoming vital to schedule their battery charging/discharging to maintain the network balance and increase the vehicle owners’ profit. Electric vehicles are now considered one of the most important and accessible ...
Read More
With an increasing penetration rate of electric vehicles in distribution networks, it is becoming vital to schedule their battery charging/discharging to maintain the network balance and increase the vehicle owners’ profit. Electric vehicles are now considered one of the most important and accessible sources of revenue for their owners since they can be connected to the grid (V2G) as a power source during peak hours. As such, while flattening the power profile, they can improve the voltage drop across the grid buses. If charging/discharging of the vehicles is scheduled irregularly, the power drawn from the phases will become unbalanced, which can cause global outages and impair system stability in addition to increasing the harmonic volume and decreasing power quality. The present paper uses dynamic programming to reduce operating costs and enhance the profits of vehicle owners who participate in the V2G program. This optimization algorithm eliminates the undesirable paths leading to unconventional responses in the search space, which will greatly increase the speed and accuracy by which the optimal response is achieved. This model, along with multi-part tariffs on electricity prices, can lead to the more active participation of vehicle owners and help improve the power quality indices of the electrical parameters of the grid. The proposed method is simulated on a sample distribution network, and the case studies conducted prove the validity of the proposed algorithm.
Optimization
Nastaran Darjani; Hesam Omranpour
Abstract
Nowadays time series analysis is an important challenge in engineering problems. In this paper, we proposed the Comprehensive Learning Polynomial Autoregressive Model (CLPAR) predict linear and nonlinear time series. The presented model is based on the autoregressive (AR) model but developed in a polynomial ...
Read More
Nowadays time series analysis is an important challenge in engineering problems. In this paper, we proposed the Comprehensive Learning Polynomial Autoregressive Model (CLPAR) predict linear and nonlinear time series. The presented model is based on the autoregressive (AR) model but developed in a polynomial aspect to make it more robust and accurate. This model predicts future values by learning the weights of the weighted sum of the polynomial combination of previous data. The learning process for the hyperparameters and properties of the model in the training phase is performed by the metaheuristic optimization method. Using this model, we can predict nonlinear time series as well as linear time series. The intended method was implemented on eight standard stationary and non-stationary large-scale real-world datasets. This method outperforms the state-of-the-art methods that use deep learning in seven time series and has better results compared to all other methods in six datasets. Experimental results show the advantage of the model accuracy over other compared methods on the various prediction tasks based on root mean square error (RMSE).
Industrial Electronics
Roya Naderi; Ebrahim Babaei; Mehran Sabahi; Ali Daghigh
Abstract
This work proposes a new multilevel inverter consisting of basic and submultilevel units. The basic unit is made-up of four isolated dc voltage sources, two bidirectional switches and ten unidirectional switches. To increase the number of the output voltage levels, a cascaded architecture based on series ...
Read More
This work proposes a new multilevel inverter consisting of basic and submultilevel units. The basic unit is made-up of four isolated dc voltage sources, two bidirectional switches and ten unidirectional switches. To increase the number of the output voltage levels, a cascaded architecture based on series connection of sub-multilevel is proposed. The proposed inverter utilizes two algorithms to determine the values of dc voltage sources. Number of IGBTs, dc voltage sources, gate driver circuits, variety of dc voltage sources and peak standing voltage on the switches are calculated and their optimization to produce maximum number of levels in output voltage is investigated. To examine advantages of the proposed inverter, the topology is compared with other topologies. The results show superiority of proposed topology over most conventional topologies, in number of circuit components. Finally, to confirm the performance of the proposed multilevel inverter, experimental results of a 25-level inverter prototype are provided.
Power systems
Masoud Maleki Rizi; Saeed Abazari; Nima Mahdian
Abstract
This paper presents enhancement of power system dynamic stability while equipped with both unified power flow controller and doubly fed induction generator by using LMI technique. We have used all UPFC (Unified Power Flow Controller) main basic PI controllers and its POD (Power Oscillation Damping) supplementary ...
Read More
This paper presents enhancement of power system dynamic stability while equipped with both unified power flow controller and doubly fed induction generator by using LMI technique. We have used all UPFC (Unified Power Flow Controller) main basic PI controllers and its POD (Power Oscillation Damping) supplementary controller. More complete model of DFIG (Doubly Fed Induction Generator) and both RSC (Rotor Side Converter) and GSC (Grid Side Converter) dynamics with their controllers have considered too. These two devices controllers have simultaneously co-ordinate and optimized with compromising between their control variables parameters. PSO (Particle Swarm Optimization) algorithm has used to optimize an objective function based on Eigen values and damping ratio to reach to best parameters and variables of controllers of both of UPFC and DFIG. LMI (Linear Matrix Inequality) have applied to whole system linearized model to reach to optimally modified eigenvalues. Within steady state and dynamic study we considered practical line thermal capacity and UPFC power rating too. Simulation results in 39-bus 10-machine Ne-England power systems ilustrate the capability of applied method. The results demonstrated that coordinated control of these two devices beside using LMI tend to more damping of system modes oscillation and more stability in power system.
Optimization
Mina Salarian; Milad Niaz Azari; Mostafa Haji aghai
Abstract
One of the important issues in designing high-performance brushless direct current (BLDC) motors is reducing the cogging torque since it results in mechanical vibration, audible noises, and torque ripples, which adversely impact the performance of the motor, which is awkward high-accuracy applications. ...
Read More
One of the important issues in designing high-performance brushless direct current (BLDC) motors is reducing the cogging torque since it results in mechanical vibration, audible noises, and torque ripples, which adversely impact the performance of the motor, which is awkward high-accuracy applications. This paper proposes an optimum design for BLDC motors aimed at reducing the cogging torque based on the capability of metaheuristics algorithms in finding the optimal solution. For this purpose, a simplified cogging torque equation is used as the objective function whose design variables include air gap length, magnet height, slot height, slot opening, and motor axial length. These are the five most influential parameters of cogging torque. On the other hand, we employ not only the old metaheuristics algorithms like the Genetic Algorithm (GA) and Simulated Annealing (SA) but also more recent algorithms such as Keshtel Algorithm (KA) along with the hybrid ones to benefit from their strength. The simulation is performed in the Matlab package. First, five selected optimization algorithms are applied and the results are investigated. The results of all the algorithms show a significant reduction in the cogging torque. Eventually, the proposed algorithms are compared to one another in terms of their value of cogging torque. The results show the superiority of the KASA algorithm in comparison with the others.
Optimization
Hassan Farokhi Moghadam; Nastaran Vasegh; Seyed Mohsen Seyed Moosavi
Abstract
In this paper, an adaptive repetitive controller (ARC) is proposed to reject periodic disturbance with an unknown period. First, a repetitive controller is designed when the disturbance period is known. In this case, the RC time delay is equal to the period of disturbance. Then, the closed-loop system ...
Read More
In this paper, an adaptive repetitive controller (ARC) is proposed to reject periodic disturbance with an unknown period. First, a repetitive controller is designed when the disturbance period is known. In this case, the RC time delay is equal to the period of disturbance. Then, the closed-loop system with the RC controller is analyzed and the effect of RC gain, k, is studied analytically. It is shown that by increasing k, the steady-state error is reduced. It is dependent on the speed of the response convergence. Secondly, an adaptive fast Fourier transform (AFFT) algorithm is proposed to extract the accurate period of disturbance adaptively. Simulation results show that the period is converged to its true value even though varying the period. Also, simulation results about the effect of controller gain are in good agreement with analytical results. Finally, it is shown that the proposed method can decrease the amplitude and energy of output signal significantly.
Power systems
Reza Yazdanpanah
Abstract
Design methodology of the compact size charging system for emergency uses is investigated in this paper. The system consists of a gearing unit, generator, charger and battery cells. The system electrical model has been introduced and a 3-phase axial flux surface mounted PM generator with concentrated ...
Read More
Design methodology of the compact size charging system for emergency uses is investigated in this paper. The system consists of a gearing unit, generator, charger and battery cells. The system electrical model has been introduced and a 3-phase axial flux surface mounted PM generator with concentrated winding is picked as a generation unit. The basic equations needed for the generator design are presented as well as general considerations and constraints for the designs. After validating the design equations using Finite Element Analysis, sizing equations are used to design different machines. The resulted designs are compared based on main characteristics including output power, delivered power to the battery, and generator efficiency. Finally, the performances of valid designs in the system model are analyzed on the wide speed range. The proposed methodology could be used for portable power generation units design for applications such as rescuing, power system outage, camping, so on.
Optimization
Mohammad Dehghani; Zeinab Montazeri; Om Parkash Malik; Ali Ehsanifar; Ali Dehghani
Abstract
Random based inventive algorithms are being widely used for optimization. An important category of these algorithms comes from the idea of physical processes or the behavior of beings. A new method for achieving quasi-optimal solutions related to optimization problems in various sciences is proposed ...
Read More
Random based inventive algorithms are being widely used for optimization. An important category of these algorithms comes from the idea of physical processes or the behavior of beings. A new method for achieving quasi-optimal solutions related to optimization problems in various sciences is proposed in this paper. The proposed algorithm for optimizing the orientation game is a series of optimization algorithms that are formed with the idea of an old game and search operators are an arrangement of players. These players are displaced in a certain space, under the influence of the game referee's orders. The best position is achieved by the laws are there in this game .In this paper, the real version of the algorithm is presented. The results of optimization of a set of standard functions confirm the optimal efficiency of the proposed method, as well as the superiority of the proposed algorithm over the genetic algorithm and the particle swarm optimization algorithm.
