University of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Automatic Detection of Various Epileptic Seizures from EEG Signal Using Deep Learning NetworksAutomatic Detection of Various Epileptic Seizures from EEG Signal Using Deep Learning Networks1122461910.22108/isee.2020.115532.1192FASobhanSheykhivandPhD Student, Faculty of Electrical and Computer Engineering, University of Tabriz, Tabriz, IranSaeedMeshginiDepartment of Biomedical Engineering, Faculty of Electrical and Computer Engineering, University of Tabriz, Tabriz, IranZohrehMousaviPhD Student, Department of Mechanical Engineering, University of Tabriz, Tabriz, IranJournal Article20190209Using an intelligent method to automatically detect epileptic seizures in medical applications is one of the most important challenges in recent years to reduce the workload of doctors in the analysis of epilepsy data through visual inspection. One of the problems of automatic detection of various epileptic seizures is the extraction of desirable characteristics, in such a way that these characteristics can make the most distinction between different phases of epilepsy. The process of finding the right features is usually a matter of time. This research presents a new approach for the automatic identification of epileptic episodes. In this paper, a deep convolutional network with eight convolutional layers and two fully-connected layers is provided to learn the characteristics hierarchically and automatically identify epileptic episodes using the EEG signal. The results show that the use of deep learning in applications such as learning characteristics hierarchically and identification of different stages of epilepsy has a higher success rate than other previous methods. The proposed model presented in this paper provides an average of 100% accuracy, sensitivity and specificity for the classification of three different epileptic seizures.Using an intelligent method to automatically detect epileptic seizures in medical applications is one of the most important challenges in recent years to reduce the workload of doctors in the analysis of epilepsy data through visual inspection. One of the problems of automatic detection of various epileptic seizures is the extraction of desirable characteristics, in such a way that these characteristics can make the most distinction between different phases of epilepsy. The process of finding the right features is usually a matter of time. This research presents a new approach for the automatic identification of epileptic episodes. In this paper, a deep convolutional network with eight convolutional layers and two fully-connected layers is provided to learn the characteristics hierarchically and automatically identify epileptic episodes using the EEG signal. The results show that the use of deep learning in applications such as learning characteristics hierarchically and identification of different stages of epilepsy has a higher success rate than other previous methods. The proposed model presented in this paper provides an average of 100% accuracy, sensitivity and specificity for the classification of three different epileptic seizures.https://isee.ui.ac.ir/article_24619_eaaf7430c94ce6aa69b90c7517808919.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Presenting a New Model of Electric Power Consumption Estimation Based on Parallel Wavelet Converters and Convolutional Neural Networks with Deep Learning for Residential BuildingsPresenting a New Model of Electric Power Consumption Estimation Based on Parallel Wavelet Converters and Convolutional Neural Networks with Deep Learning for Residential Buildings13242453810.22108/isee.2020.116987.1224FANaserKurdDepartment of Computer Engineering, Kerman Branch, Islamic Azad University, Kerman, IranFarshidKeyniaEnergy Department, International Center for Science, High Technology & Environmental Sciences, Kerman, Iran0000-0002-9027-7315Journal Article20190626Considering the increasing rate of electrical energy usage, this energy has become one of the most important resources for human life. So all countries are seeking access to reliable and planned energy resource. Regarding the non-renewability of fossil fuel resources, especially oil and gas, the issue of replacing these types of energy with renewable energy has been considered for decades. Saving and optimal use of electrical energy in important applications such as residential and commercial buildings is critical. One of the most important factors for planning power consumption and optimizing it is accurate forecasting for next hours’ power consumption of residential and commercial buildings. In this paper, first, the data sets of several residential buildings are analyzed using parallel wavelet converters. Then, using an optimal estimator model of the convolutional neural network, the short-term load of the building is estimated. The obtained results show that the proposed method has improved the prediction error about 70, 69 and 73 percent for ARIMA, SVR, and LSTM methods, respectively.Considering the increasing rate of electrical energy usage, this energy has become one of the most important resources for human life. So all countries are seeking access to reliable and planned energy resource. Regarding the non-renewability of fossil fuel resources, especially oil and gas, the issue of replacing these types of energy with renewable energy has been considered for decades. Saving and optimal use of electrical energy in important applications such as residential and commercial buildings is critical. One of the most important factors for planning power consumption and optimizing it is accurate forecasting for next hours’ power consumption of residential and commercial buildings. In this paper, first, the data sets of several residential buildings are analyzed using parallel wavelet converters. Then, using an optimal estimator model of the convolutional neural network, the short-term load of the building is estimated. The obtained results show that the proposed method has improved the prediction error about 70, 69 and 73 percent for ARIMA, SVR, and LSTM methods, respectively.https://isee.ui.ac.ir/article_24538_7a9c6c0e1a21bd4e0478842c3b5570a3.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Design and optimization of an Approximate Full-adder Based on CNTFETs and its application in image processingDesign and optimization of an Approximate Full-adder Based on CNTFETs and its application in image processing25362471910.22108/isee.2020.120390.1313FAMohammad RezaReshadinezhadFaculty of Computer Engineering, University of Isfahan, Isfahan, IranSeyed ErfanFatemiehFaculty of Computer Engineering, University of Isfahan, Isfahan, IranZahraDavari ShalamzariFaculty of Computer Engineering, University of Isfahan, Isfahan, IranJournal Article20191215Novel digital circuit design methods are vital due to the significant increase in data that requires fast processors. No doubt, power consumption is an essential factor in electronic devices. Hence, the design of low-power, area-efficient, and high-performance circuits is crucial. Approximate computing as a promising method for designing efficient circuits in addition to applying CNTFETs can be an excellent solution for the concerns mentioned above. In this article, according to the full adder’s importance in DSP processors, a new approximate full adder based on 32nm Stanford CNTFET model is proposed and optimized in terms of power consumption, delay, PDP, and the number of transistors. HSPICE is applied to compare this new design with state-of-art articles. The simulation results indicate that the proposed design has not only the least delay but also shows an 87% improvement in PDP achieved. Various simulations applying different load capacitors, supply voltages, and process variations demonstrate the acceptable functionality of proposed approximate full-adder in different situations. Image addition simulation using MATLAB is applied to assess the performance of the proposed design in a real error-resilient application.Novel digital circuit design methods are vital due to the significant increase in data that requires fast processors. No doubt, power consumption is an essential factor in electronic devices. Hence, the design of low-power, area-efficient, and high-performance circuits is crucial. Approximate computing as a promising method for designing efficient circuits in addition to applying CNTFETs can be an excellent solution for the concerns mentioned above. In this article, according to the full adder’s importance in DSP processors, a new approximate full adder based on 32nm Stanford CNTFET model is proposed and optimized in terms of power consumption, delay, PDP, and the number of transistors. HSPICE is applied to compare this new design with state-of-art articles. The simulation results indicate that the proposed design has not only the least delay but also shows an 87% improvement in PDP achieved. Various simulations applying different load capacitors, supply voltages, and process variations demonstrate the acceptable functionality of proposed approximate full-adder in different situations. Image addition simulation using MATLAB is applied to assess the performance of the proposed design in a real error-resilient application.https://isee.ui.ac.ir/article_24719_57ae5dd21250159677d04441210e1f74.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Presenting a New Multi-Objective Optimization Method Based on MOPSO-SQP Algorithm in Order to Coordinate the Protective Overcurrent Relays in Power SystemsPresenting a New Multi-Objective Optimization Method Based on MOPSO-SQP Algorithm in Order to Coordinate the Protective Overcurrent Relays in Power Systems37502449410.22108/isee.2020.119289.1281FASamiraSadeghiDept. of Electrical Engineering, University of Kurdistan, Sanandaj, IranNavidRezaeiDept. of Electrical Engineering, University of Kurdistan, Sanandaj, Iran0000-0002-8570-8301AliHesami-NaghshbandyDept. of Electrical Engineering, University of Kurdistan, Sanandaj, IranParhamMoradiDept. of Electrical Engineering, University of Kurdistan, Sanandaj, IranJournal Article20190923One of the important issues in coordination of protective relays is minimization of the time interval between the operation of main and backup overcurrent (OC) relays. Overcurrent relays coordination problem due to the large number of variables and the nature of the objective functions can be introduced as a complex optimization problem that necessitates the need for an efficient optimization method with appropriate accuracy and speed. Given that for the realization of protective purposes including increased relays operation speed, selectivity, support, reliability and stability, different objective functions can be described, therefore, providing a multi-objective mathematical problem in optimizing protective problems may be necessary. In this regard, this article due to the capabilities of the MOPSO, proposes a multi-objective optimization structure, in which the SQP by increasing the speed and not increasing the search space is added to MOPSO. In this paper, several objective functions are proposed based on protection goals and optimal adjustment points have been extracted using the proposed multi-objective MOPSO-SQP. The simulations have been implemented and carefully analyzed on several typical power systems. The simulation results confirm the efficiency of the proposed algorithm in ensuring the optimal coordination of protective overcurrent relays in power system.One of the important issues in coordination of protective relays is minimization of the time interval between the operation of main and backup overcurrent (OC) relays. Overcurrent relays coordination problem due to the large number of variables and the nature of the objective functions can be introduced as a complex optimization problem that necessitates the need for an efficient optimization method with appropriate accuracy and speed. Given that for the realization of protective purposes including increased relays operation speed, selectivity, support, reliability and stability, different objective functions can be described, therefore, providing a multi-objective mathematical problem in optimizing protective problems may be necessary. In this regard, this article due to the capabilities of the MOPSO, proposes a multi-objective optimization structure, in which the SQP by increasing the speed and not increasing the search space is added to MOPSO. In this paper, several objective functions are proposed based on protection goals and optimal adjustment points have been extracted using the proposed multi-objective MOPSO-SQP. The simulations have been implemented and carefully analyzed on several typical power systems. The simulation results confirm the efficiency of the proposed algorithm in ensuring the optimal coordination of protective overcurrent relays in power system.https://isee.ui.ac.ir/article_24494_486577502f9d65296bda8208b3d60be8.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Total Harmonic Distortion Optimization of Delta-Connected H-Bridge Multilevel InvertersTotal Harmonic Distortion Optimization of Delta-Connected H-Bridge Multilevel Inverters51642453910.22108/isee.2020.118116.1256FASeyedeh ZahraHosseini MallaiFaculty of Electrical and Computer Engineering, Jundi-shapur University of Technology-Dezful, Khuzestan, IranAlirezaNamadmalanFaculty of Electrical and Computer Engineering, Jundi-shapur University of Technology-Dezful, Khuzestan, IranAhmadBehzadinezhadFaculty of Electrical and Computer Engineering, Jundi-shapur University of Technology-Dezful, Khuzestan, IranJournal Article20190711Presence of triplen voltage harmonics and circuiting current in a delta-connected cascaded H-Bridge multi-level inverter causes extra power losses. In this paper, a low frequency optimization method is proposed to reduce triplen voltage harmonics in delta-connected cascade H-Bridge multilevel inverters. Using this method, in addition to reducing these harmonics, the total harmonic distortion (THD) is also maintained at optimum level and the requirements of standards such as EN50160, WG36-05 and IEC61000-3-6 are also satisfied. In this method, using optimal minimization of total harmonic distortion (OMTHD), the optimal values of the switching angles and DC sources are obtained using the Particle Swarm Optimization (PSO) algorithm. The simulation and experimental results show the effectiveness of the proposed method in reducing triplen voltage harmonics and circuiting currents while keeping the THD at optimal levels.Presence of triplen voltage harmonics and circuiting current in a delta-connected cascaded H-Bridge multi-level inverter causes extra power losses. In this paper, a low frequency optimization method is proposed to reduce triplen voltage harmonics in delta-connected cascade H-Bridge multilevel inverters. Using this method, in addition to reducing these harmonics, the total harmonic distortion (THD) is also maintained at optimum level and the requirements of standards such as EN50160, WG36-05 and IEC61000-3-6 are also satisfied. In this method, using optimal minimization of total harmonic distortion (OMTHD), the optimal values of the switching angles and DC sources are obtained using the Particle Swarm Optimization (PSO) algorithm. The simulation and experimental results show the effectiveness of the proposed method in reducing triplen voltage harmonics and circuiting currents while keeping the THD at optimal levels.https://isee.ui.ac.ir/article_24539_93602ba9a959a293b154cf0f7dcf66b0.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Improving the performance of three dimensional wireless sensor networks with nodes displacement capabilityImproving the performance of three dimensional wireless sensor networks with nodes displacement capability65822454010.22108/isee.2020.118189.1258FAMehdiSalkhordeh HaghighiFaculty of Computer Engineering, Sadjad University of Technology, Mashhad, IranPayamAminolsharieh NajafiFaculty of Computer Engineering, Sadjad University of Technology, Mashhad, IranJournal Article20190715Wireless sensor networks are one of the most important tools for information acquisition and environment identification in many application areas. Recent advances in the field of electronics and wireless telecommunications have led to the design and manufacture of sensors with low power consumption, small size, reasonable price and various applications. Most research in the area of wireless sensor networks has focused on two dimensional sensor networks while in the real world, most of the applications are three dimensional. Some research in this area has focused on underwater, space, forestry, and environment applications. The main objective of the current research is increasing network lifetime by defining new parameters and embedding them in the fuzzy clustering or fuzzy C-means algorithm that has been adapted for three dimensional wireless sensor networks. One of the parameters that has been used in this research is limited movement of sensors. By adding this ability to the network, there is an expectation of improvement in network performance. The results of the experiments indicate the positive effect of this ability on network performance and lifetime.Wireless sensor networks are one of the most important tools for information acquisition and environment identification in many application areas. Recent advances in the field of electronics and wireless telecommunications have led to the design and manufacture of sensors with low power consumption, small size, reasonable price and various applications. Most research in the area of wireless sensor networks has focused on two dimensional sensor networks while in the real world, most of the applications are three dimensional. Some research in this area has focused on underwater, space, forestry, and environment applications. The main objective of the current research is increasing network lifetime by defining new parameters and embedding them in the fuzzy clustering or fuzzy C-means algorithm that has been adapted for three dimensional wireless sensor networks. One of the parameters that has been used in this research is limited movement of sensors. By adding this ability to the network, there is an expectation of improvement in network performance. The results of the experiments indicate the positive effect of this ability on network performance and lifetime.https://isee.ui.ac.ir/article_24540_6bc789e5c9d2445687e796aa459ef06a.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Developing an Intelligent Pattern Using Principal Component Analysis Method to Detect Eccentricity Faults in Squirrel-Cage Induction MotorsDeveloping an Intelligent Pattern Using Principal Component Analysis Method to Detect Eccentricity Faults in Squirrel-Cage Induction Motors83942478210.22108/isee.2020.120169.1304FAMahdiAhmadiDept. of Electrical Engineering, Esfarayen University of Technology, North-Khorasan, IranJournal Article20191125Eccentricity fault is one of the prevalent faults in rotating machines which can cause other mechanical and electrical faults. This paper focuses on detecting the static and dynamic eccentricity faults in the squirrel-cage induction motors. The induction motor is modeled by the finite element method (FEM), a powerful and accurate method, using FLUX 2D software. The current signal cannot be used to detect the static and dynamic eccentricity faults especially when their severity is small. Therefore, the search-coil based method is employed and the voltage of two symmetrical search-coils is analyzed to detect and diagnose the static and dynamic faults. Since two search-coils are open-circuit, they do not affect the behavior of the induction motor. The analysis based on the principal component analysis (PCA) method shows that there exists an intelligent pattern that, firstly, is sensitive to the occurrence of eccentricity faults, even to a low degree, and secondly, has the ability to distinguish the type of fault (static or dynamic).Eccentricity fault is one of the prevalent faults in rotating machines which can cause other mechanical and electrical faults. This paper focuses on detecting the static and dynamic eccentricity faults in the squirrel-cage induction motors. The induction motor is modeled by the finite element method (FEM), a powerful and accurate method, using FLUX 2D software. The current signal cannot be used to detect the static and dynamic eccentricity faults especially when their severity is small. Therefore, the search-coil based method is employed and the voltage of two symmetrical search-coils is analyzed to detect and diagnose the static and dynamic faults. Since two search-coils are open-circuit, they do not affect the behavior of the induction motor. The analysis based on the principal component analysis (PCA) method shows that there exists an intelligent pattern that, firstly, is sensitive to the occurrence of eccentricity faults, even to a low degree, and secondly, has the ability to distinguish the type of fault (static or dynamic).https://isee.ui.ac.ir/article_24782_b3d21ac602e894e0d7a66efe92727118.pdfUniversity of IsfahanComputational Intelligence in Electrical Engineering2821-068911320200922Performance and Reliability Improvement on 2D-NOC Based on Reducing the Number of Passing LinksPerformance and Reliability Improvement on 2D-NOC Based on Reducing the Number of Passing Links951062448010.22108/isee.2020.117136.1231FASeyed AminAlaviDept. of Electrical Engineering, Mashhad Branch, Islamic Azad University, Mashhad, IranSeyyed JavadSeyyed Mahdavi ChabokDept. of Electrical Engineering, Mashhad Branch, Islamic Azad University, Mashhad, IranJournal Article20190520Network on-chip is a communication subsystem within an integrated circuit that provides communication between processors in the on-chip system. There are several different ways to get from one node to another. Therefore, there must be a routing algorithm to find the route to the destination. This paper presents an algorithm based on the reduction of the passing path to reach a packet from origin to destination which is able to increase the reliability, reduce latency, power consumption and increase network efficiency on the chip. And this is when most of the fault-tolerant networks presented in this field increase parameters such as delay, power consumption and circuit complexity in order to achieve higher reliability. The proposed method improves network performance with minimal hardware changes and circuit complexity. The path passed by the packet is reduced to reach the destination, which means passing through fewer links and routers and less chance of encountering faulty links and routers and increasing network reliability. Also, passing fewer links and routers will reduce network latency and power consumption.Network on-chip is a communication subsystem within an integrated circuit that provides communication between processors in the on-chip system. There are several different ways to get from one node to another. Therefore, there must be a routing algorithm to find the route to the destination. This paper presents an algorithm based on the reduction of the passing path to reach a packet from origin to destination which is able to increase the reliability, reduce latency, power consumption and increase network efficiency on the chip. And this is when most of the fault-tolerant networks presented in this field increase parameters such as delay, power consumption and circuit complexity in order to achieve higher reliability. The proposed method improves network performance with minimal hardware changes and circuit complexity. The path passed by the packet is reduced to reach the destination, which means passing through fewer links and routers and less chance of encountering faulty links and routers and increasing network reliability. Also, passing fewer links and routers will reduce network latency and power consumption.https://isee.ui.ac.ir/article_24480_41037a447ebfefc719418e5d1fcbaa05.pdf