فصلنامه روشهای هوشمند در صنعت برق
پیاپی 61 (بهار 1404)

With the expansion and success of convolutional networks, the topic of deep learning has attracted increasing attention in recent years; Since convolutional networks include many layers, optimal learning of network layers is of great importance. In this paper, a new model, called the 4-stream model, is presented with the aim of helping to linearize the data space using representational dissimilarity transformation, and the effects of this transformation on standard classifications for artificial data and Cifar10 images are investigated. Then, two models based on data preprocessing with dissimilarity transform representation and Sobel and Edge Detector filters are analyzed. The 4-stream model increased the accuracy by 3.2% due to the increase in the number of model parameters, and hence the capacity of the network. Besides, adding the dissimilarity representation wherever the classifier cannot perform a high-resolution classification by merely using the main features, can help to increase the discriminability of classes by adding linear features.

To overcome the problem of channel estimation in massive multiple-input multiple-output (M-MIMO) systems, in this paper we propose a downlink link channel estimation scheme in frequency-division duplex (FDD) based on structured compressive sensing to reduce the pilot required by which Intrinsic spatial sparsity of M-MIMO delay channels are amplified. For this purpose, first, after discussing the different methods of channel estimation and examining the existing challenges, we define our roadmap and propose our algorithm, in which we estimate the channel based on the greedy orthogonal matching pursuit (OMP) algorithm. In this algorithm, spatial correlation between the channel impulse response of different transmitter antennas is used for accurate channel estimation. This algorithm obtains the channel sparsity in an adaptive way, which negates the ideal assumption of the previous works that the channel sparsity is in hand. In this case, this algorithm estimates the channel with good accuracy in cases when the exact amount of channel sparsity is not known. Finally, we present simulations that demonstrate the ability of this method to reduce the required pilot. The simulations show that the proposed channel estimation reliably obtains the channel sparsity level and the support set compared to similar methods.

Since most space-based synthetic aperture radar systems have an accuracy of less than one meter in the resolution of images, very accurate processing of synthetic aperture radar data to produce images with high-resolution accuracy is of particular importance. In this article, methods for actual modeling and simulation of the space-based synthetic aperture radar system are presented and the raw data were obtained. For simulation and modeling, the main characteristics of the real satellite synthetic aperture radar system related to sensor mode/dynamics, target observation, antenna beam patterns, pointing errors on the antenna beam, and raw data generation are reflected. Analyzes based on simulations show the effectiveness of the presented methods. In the simulation, the presented method compensates for the phase errors induced by the aiming errors of the antenna beam. The results of the centralization of raw data, the calculated value of the resolution accuracy of the slant range is equal to 1.89 meters. Also, the average values of the measured slant range resolution accuracy, peak side-lobe ratio (PSLR), and integrated side-lobe ratio (ISLR) for the Interrogation Rate Frequency (IRF). An unweighted point in the focused image was obtained around 1.94 m, 13.57 dB and -10.26 dB respectively. The calculated value of azimuth resolution accuracy is 2.24 meters and the average values of measured azimuth resolution accuracy, PSLR, and ISLR for unweighted point target IRFs are 2.29 meters, -12.57 dB and -9.68 dB, respectively. These results show the effectiveness of the proposed method. In other words, the performance of space-based synthetic aperture radar image formation using the proposed method for raw data is very good, so the various effects induced by the real synthetic aperture radar sensor are reflected. Therefore, these results confirm the proposed methods for forming the space-based synthetic aperture radar image.

Quantum-dot cellular automata (QCA) is a modern technology, which has higher speed, lower power consumption, higher density, and lower complexity than conventional technologies, such as CMOS. Moreover, the gate diffusion input (GDI) technique has been successful in reducing complexity, area, and energy consumption in low-power circuit designs. In this technique, a wide range of complex logic functions can be implemented using only two transistors as the main block. In this study, a QCA-based GDI block is proposed using only 11 cells as a standard design unit that can be used to implement basic functions such as AND, OR, MUX, BUFFER, NOT and XOR in digital circuits. QCADesigner simulations of the functions in 18 nm technology indicate the superior performance of the proposed block with only one clock cycle delay in performing the operations. Moreover, the power consumption analysis of the designed circuits is performed using QCADesigner. The advantages of the proposed circuit compared to previous designs are 31% reduction in cell count, 50% smaller surface area, and 17% reduction in total energy loss.

Edge computing is a computing paradigm that extends cloud services to devices at the edge. This processing model refers to technologies that allow computing and storage to be performed on devices at the edge of the network. In this architecture, computing and storage operations take place close to objects and data sources. In order to reduce latency and network traffic between end devices and cloud centers, groups at the edge have processing capabilities, perform a large number of processing and computing tasks, including data processing, temporary storage, device management, decision making, and privacy protection. Since the number of edge devices is large, there must be a mechanism to select these tasks and offload them to the cloud. The problem to be decided is that which one of the available edge devices should be selected for unloading and then unloaded. This problem is classified as one of the hard non-polynomial problems and by using deterministic algorithms simply and in polynomial time, it is not possible to find a suitable and efficient solution for it found.

In this paper, the far- and near-fields of a vessel are discussed in the full-wave CST software environ ment. In this regard, the far-field is extracted with the help of the data obtained from the scattered near-field. First, to explain the method used to calculate the scattered near-field, a simple structure (metal cube) is simulated. Then, by simulating the full-wave of a vessel with the dimensions of 130.8×20×23.1 cm3 or 1.54×103 λ3 at 8.5 GHz from the X-band, its far- and near-fields were calculated and reported according to the mentioned method. In the following, the far-field of this vessel is obtained with the help of the available data from the scattered near-field. The characteristic of the radar cross section (RCS) of this vessel is also calculated using the Asymptotic solver of the CST software. The maximum RCS of the vessel at 8.5 GHz is equal to 2.51 m2. In order to calculate all near- and far-electric fields, the Time Domain solver has been used. The simplicity of calculating the far-field from the near-field presented in this paper makes it possible to analyze similar structures using the same method.

In many applications, it is necessary to transfer electrical energy to the load as a pulse with a high voltage level. Conventional switches such as semiconductors cannot be used for these applications due to voltage or speed limitations, and therefore saturable inductors must be used. These inductors act like a short circuit at the moment of saturation and play the role of a connected switch. By designing a circuit called magnetic pulse compressor (MPC), this feature can be used to produce voltage pulses with a small width and a large amplitude. The study of the structure of magnetic pulse compressor is mentioned in this article and an algorithm for its design is proposed. By using this algorithm, an efficient compressor circuit can be designed for any application and with the characteristics of magnetic materials. To show the correct operation of the algorithm, a compressor circuit sample has been designed and simulated. The output pulse of the desired circuit for a 5 Ω resistive load has an amplitude of 500 volts and a width of 500 nanoseconds. To validate the proposed method, a laboratory sample is made. The results show that the magnetic switches have an acceptable performance and the proposed algorithm is also successful in designing the compressor circuit.

Vibrations in rotating equipment are inevitable phenomena and its analysis is necessary and useful to study the condition of the equipment. In a coupled hydroturbine-generator set, the vibrations are caused by the dynamic forces of both machines. Due to the coupling of the turbine and the generator the vibration analysis of each machine separately can not reveal all the phenomena and dynamic behaviors. The change in speed and torque of the turbine and the generator affect each other, so the results of the coupling analysis are closer to reality. In this paper, dynamic modeling of the generator-turbine coupling set is performed and the governing equations of the system are solved using numerical methods in MATLAB software. For more accurate modeling and analysis of the system, the rotational speed of the shaft is considered as variable. The gyroscopic effect and arcuate whirl are also included in the modeling of the set. The frequencies in the vibrations of the set are extracted and the dynamic phenomena related to each are identified. The results show that new dynamic and vibrational phenomena in this system. Also, considering that imbalance mass is never zero, the resulting force is considered in the modeling and the effect of the amount of imbalance mass is also studied and analyzed. Due to the nonlinearity of the governing equations, the analysis of vibrations using phase plane diagrams shows that in some operating conditions, the vibration behavior is chaotic, which can be used in the processing of experimental signals.

The use of electric vehicles (EVs), to reduce greenhouse gases and air pollution caused by the fossil fuel consumption, seems to be an inevitable solution. The increased penetration of EVs imposes a large variable load to the grid. However, the battery of EVs in an aggregator provides a large source of energy storage. Therefore, EVs depending on the charging or discharging modes can act as flexible loads or as flexible energy sources. Then, the proper coordination and control charging and discharging of EVs, using vehicle-to-grid (V2G) technology, not only can minimize the undesirable effects resulting from the increased penetration of EVs, but also can improve the voltage profile. In this paper, a new algorithm with variable-objective function is proposed to control variable quantities of generation and consump tion. In the proposed method, the control of the point of common connection (PCC) voltage in a specific value of a determined permissible range, which depends on different operating conditions, can be considered as a variable-objective function. Moreover, there are constraints to state of charge (SOC) of electric vehicles (EVs) batteries and charging/discharging time. Other advantages of using the proposed method are the reduction of network losses in peak load hours and the establishment of an appropriate coordination between charging and discharging EVs. The simulations of the IEEE 14-Bus distribution system with V2G capabilities based on the proposed variable-objective function (VOF) algorithm are implemented for both charging and discharging modes using MATLAB/PSAT software and tested for the various scenarios. Finally, the results of the proposed method are compared with the traditional method and the merit of that is proved.

This research proposes a novel method for utilizing renewable energies in electric vehicle charging stations. In the proposed method, the required energy for charging the electric vehicles, the needed energy for supplying the cryptocurrency mining farm as well as the energy for supplying the critical loads of the upstream grid are afforded by 100% renewable resources, and the electric vehicle parking station (i.e., charging station) receives no energy from the utility grid. In this model, the uncertainty of renewable energies and the mismatch between generation and demand are resolved by optimal charging-discharging of electric vehicles. While the utility grid is available, the electric vehicles are fully charged, the cryptocurrency mining farm continues operation at full capacity, and excess renewable energy is sent to the upstream grid to supply the critical loads. The critical loads are supplied from two directions of the grid and parking station. During an outage or blackout of the utility grid, the critical loads are fully supplied by an electric vehicle parking station. In this situation, the cryptocurrency miners are modeled as a responsive load. In the off-grid operation, the proposed planning manages the energy of cryptocurrency miners as well as the charging-discharging pattern of electric vehicles for reaching two objectives of fully charging all the electric vehicles and fully supplying the critical loads. If there is any surplus of energy, it is used to run the cryptocurrency miners. In the proposed model, since no energy is received from the utility grid, the issues related to cryptocurrency mining are inconsequential. The objective function of the proposed method is to minimize the number of discharge cycles on electric vehicles to avoid battery degradation. The problem is formulated as mixed integer linear programming and solved by GAMS software. The wind and solar energy uncertainty are incorporated in the model and an optimal charging-discharging pattern is designed for electric vehicles to confirm feasible operation under all energy variations.