2022
Browse
53 results
Search Results
Item Prospects and Design Assessment of a Hybrid Renewable Energy Microgrid for an Indigenous Community in Bangladesh(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Hossain, Muhit; Kayes, Abdullah Al; Suny, RayhanThe Thuisa Para indigenous community living in the hill tracts of Bandarban, Bangladesh, has never experienced the miracles of electricity, as most of the remote hilly areas of Bandarban are still not under the National Energy Grid coverage. These indigenous people are deprived of the blessings of electricity and, their socio-economic advancement is being obstructed. The Sustainable Development Goal 7 (SDG7) program under the UN aims to diminish such energy access inequalities. It is possible for communities like Thuisa Para, which are located in remote areas, to acquire an adequate supply of electricity by utilizing the available renewable energy resources. But before implementation, thorough analyses regarding the geographical factors, cost-effectiveness and durability for a particular location is required to ensure that the energy system fulfills the demands adequately. Therefore, this paper aims to propose the most affordable and most reliable hybrid renewable energy microgrid design for the Thuisa Para Community upon completion of thorough comparative analyses of the available design options. For design and simulation purposes, HOMER software has been used. Elements considered for this microgrid are specifically solar-PV panels, kinetic batteries, wind turbine, diesel generator and converter. The results obtained from the simulation were used to compare the viable design choices in terms of their respective energy production capabilities, per-unit electricity costs, net present costs (NPC) and a few other important factors. Additionally, a multiyear sensitivity analysis regarding the net present cost has been conducted for the ease of choosing the suitable project lifetime of the Thuisa Para microgrid.Item Young’s modulus distribution prediction for elasticity imaging technique(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Munira, Sidratum; Muskan, Tasfia Akter; Tajin, TamannaElastography is a noninvasive method of determining the stiffness of a tissue. The basic premise is that whether the tissue is hard or soft can provide diagnostic information regarding the presence or absence of illness. Breast elastography is a modern sonographic imaging technology that, in addition to standard ultrasonography (US) and mammography, contains data on breast lesions. The primary goal of elastography in breast imaging is to detect tumors at an early stage by offering a non-invasive approach to determine the mechanical characteristics of breast tissue. Young's modulus (YM) in biological tissues is frequently used to predict the start of pathological diseases. Understanding of this parameter has been shown to be extremely useful in the diagnosis, prognosis, and therapy of cancers. In this work, imaging approaches has been provided based on the ways for producing a stress in the tissue (external mechanical force) and measuring the tissue response. In theory, elastography data may be used to determine the Young's Modulus distribution of the desired soft tissue area. On the other hand, Elastography methods can only extract the distribution of strain in a certain region. The stress distribution profile of the top surface is considered to be available in practice. Young's modulus of the tissue is based on immediate strain in reaction to stress distribution. In this study, the difference between simulated and actual stress distribution values from different lateral surfaces of the tissue have been observed, as well as the position of the tumor. Depending on the position, this error displays the sequence of problems that must be solved in order to anticipate the stress distribution. This work clearly shows the minimization of error using logical operation to improve in the biopsy using diagnostic procedures.Item Hybrid Algorithm-Based Approach For Optimized PID Controller for Stability Analysis of DC-DC Interleaved-Buck Converter(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Rahman, Kazi Asif; Sakib, Sadman; Ahmed, MoshiurThis thesis is an investigation of the stability analysis of the Interleaved Buck converter by inducing Nature-Inspired Algorithms to create an optimal PID controller. The applicability and compatibility of some algorithms such as PSO, Firefly Algorithm (FA), GWO and Cuckoo Search are used for optimizing the control mechanism of power converters. Improvements in performance parameters are noted, and the results are compared using multiple fitness functions. The thesis focuses on Interleaved converter which has the benefit of lowering ripple currents, simplifying EMI transmission, and faster transient response. The converters are constructed using the State Space Averaging (SSA) technique to provide promising feedback control and to evaluate the transfer functions. The Nature-Inspired Algorithms are non-linear optimization methods based on artificial intelligence. The above algorithms are based on swarm intelligence and operate in accordance with swarm creature customs. Swarm intelligence ensures better data exploitation, which concentrates the search method within the vicinity of optimal solutions while also assisting the procedure to escape the restriction of the local minima, resulting in successful exploration of the search space. With the help of these algorithms, we approached a hybrid algorithm named Hybrid Firefly Particle Swarm Algorithm (HFPSO). We also compared the performance according to the standard. As a result, the algorithms are evaluated for improved system performance using two fitness functions (ITAE, and ITSE) and performance parameters such as percentage of overshoot, rise time, settling time, and peak amplitude are all variables to consider. The procedure is carried using MATLAB. After studying the results, It is discovered that the percentage of HFPSO overshoot (ITSE) offers a lower for each of the error functions than PSO, CSA, FA, GWO, WOA.Item Real-time Traffic Object Detection Using DNN Frameworks Centering Bangladesh(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Munir, Nafis Shahriar; Hossain, Nazia; Zame, Raghib RyyanIn this thesis, the research is mainly focused on traffic object detection by different computer vision algorithm like MobileNetV2 and YOLOv5. The main focus is on real-time detection and identification of various types of automobiles specific to Bangladesh. Vehicle detection is a necessary step for traffic surveillance and autonomous vehicles. Vehicle counting in complex transportation conditions requires the detection and tracking of mobile vehicles. Vehicle detection on the road are used for vehicle tracking, vehicle traffic assessment, average velocity of each individual vehicle, motion analysis, and vehicle classification, and may be applied in a variety of contexts. Because of the irregular traffic, the variety of vehicles, and the absence of a good dataset, it is more difficult in Bangladesh to adopt a smart traffic. One of the most important aspects of algorithm training is data quality. Here, the dataset, “Dhaka Traffic Detection Challenge Dataset” was cleaned and augmented to get better results. The dataset was trained on two neural network architecture, YOLOv5 and MobileNetV2. YOLOv5 ran on PyTorch and the model was trained on Google Colaboratory, a cloud-based platform. Codes were written in Python 3.8 and Python 3.9. Roboflow, an online-based computer vision application, was used to organize the dataset for training. For image categorization and mobile vision, MobileNet is built on the CNN architectural model. There are other models available, but MobileNet is unique in that it uses very minimal compute resources to operate and also applies transfer learning. As a result, MobileNet is ideal for both mobile devices and web browsers. There are 28 levels in MobileNet. MobileNet contains 4.2 million parameters by default. YOLOv5 performed better than MobileNetV2, in terms of accuracy and inference time. This research will be a vital step towards intelligent traffic detection system that can detect unauthorized vehicles like rickshaw/CNGs in highways, or to develop a traffic plan that minimizes traffic congestion on the road.Item Driver audacity evaluation based on environment-specific embedded cyber-physical systems using low-level computation platform.(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Khan, Md. Asifuzzaman; Ashfaq, Md. Arean; Ahmed, Md. Kazi SazzadReckless driving of public transport in third world countries like Bangladesh has long been an unsolved issue. One of the reasons being the lack of constant monitoring of the driver’s driving style. Present solution to this is to use a constant monitoring device onboard the vehicle to assess the driver’s performance. They use onboard neural networks to analyze various driving data pattern collected from the suite of sensors on board the car. However the use of onboard neural network increases the computational complexity and demands for expensive processor. Moreover the suite of sensors used such as GPS, Steering angle sensor, IMU adds to the cost. This makes them a not so viable option for mass use in the roads of a underdeveloped country like Bangladesh. Here in this paper we propose a plug and play device that can analyze the driving behavior in real-time using an onboard low level processor and 2 low cost sensors. This process eliminates the need of onboard neural network. A test bench is setup for simulating the various driving patterns and the collected data is compared with the published studies to validate the method. Then the methodology is explained in details. Finally the whole system is tested with the collected data and result is analyzed.Item Estimation and Optimization of Attenuation of High Frequency mmWave within 5G Spectrum(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Haque, Md. Monzurul; Eisham, Zubayer Kabir; Rahman, Md. SamiurFifth-generation (5G) introduces the use of millimeter waves (mmWave) in cellular technology, and thus poses a great challenge in proper radio coverage. One of the difficulties in high-frequency coverage is the outdoor to indoor (O2I) penetration loss for indoor users. An estimation in O2I penetration losses can help operators decide to ensure proper usage of available radio resources in the range of 5G. The first part of the work presents an estimation of the variation pattern of penetration losses with varying frequencies, within 5G supported range, for different building exterior conditions. For the purposes of simulation two simulators have been used, mainly NYUSIM and a MATLAB based simulator developed using 3GPP TR 38.901. This paper also compares the simulation results from these two simulators. The use of high frequency mmWave generates yet another issue of attenuation due to different environmental factors such as temperature, rain rate, humidity etc. This attenuation causes significant loss of transmission power, resulting in poor service and radio coverage quality. To mitigate this issue, it is of utmost necessity for the operators to be concerned about the optimum operating frequencies for certain environmental situation based on the loss due to environmental attenuation. The total problem becomes a multidimensional optimization problem which can be readily optimized using nature inspired metaheuristic algorithms. In the second part of this work, the multidimensional optimization problem of environmental attenuation is optimized by different well known optimization algorithms to investigate the optimum and worse operating points of operation for proper radio coverage.Item Multi-variate Time-series Load Forecasting using Deep Learning(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Arnob, Saadman Sakif; Saqalain, Ahmed Syed; Sakib, Najmus SadatTo ensure the stable and reliable operation of a power system, load forecasting is required. Accurate forecasting leads to efficient dispatch, unit commitment, and energy security. Smart power management in the generating, transmission, and distribution network, as well as the accompanying energy demand, can be realized with accurate forecasting approaches. This paper analyses the short-term load forecasting of the Bangladesh power system. Various deep neural network models- XGBoost, LSTM, Stacked LSTM, CNN, CNN-LSTM, Time Distributed MLP, and Encoder-Decoder are used to forecast the load. The load is predicted based on previous load data and various features like temperature, Weekdays, Weekends, and Peak Business Hours are taken to ensure the accuracy of the results. This study reports the advantages and disadvantages of each model.Item Distribution System Performance Improvement using Meta-heuristic Optimization Algorithms(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Apon, Hasan Jamil; Morshed, Khandaker Adil; Abid, Md. ShadmanThe critical challenge for an efficient islanding operation of a distribution system having Distributed Generation (DG) is preserving the frequency and voltage stability. Contemporary load shedding schemes are inefficient and do not adequately assess the optimum amount of load to shed which results in either excessive or inadequate load shedding. Appropriate installation of renewable energy-based distributed generation units (RDGs) is one of the most important challenges and current topics of interest in the optimal functioning of modern power networks. Due to the intermittent nature of renewable energy sources, optimal allocation and sizing of RDGs, particularly photovoltaic (PV) and wind turbine (WT), remains a critical task. Additionally, maintaining frequency and voltage stability is crucial for optimal functioning of an islanded network connected to DGs. Conventional load shedding schemes do not effectively identify the optimal amount of load to shed, culminating in either excessive or insufficient load shedding. Hence, the first part of this work presents an optimal load shedding technique using Chaotic Slime Mould Algorithm (CSMA) with sinusoidal map in order to achieve greater efficiency. A constrained function with static voltage stability margin (VSM) index and total remaining load after load shedding was applied to accomplish the evaluation. A total of three islanding scenarios of IEEE 33 bus and IEEE 69 bus radial distribution systems were used as test systems to assess the efficacy of the proposed load shedding approach using MATLAB software. To identify performance enhancements, the developed method was compared to Backtrack Search Algorithm (BSA) and the original SMA. According to the results, CSMA outperforms both BSA and SMA in terms of remaining load and voltage stability margin index values in all the test systems. Moreover, the second part of this work proposes Chaotic Equilibrium Optimizer (CEO) with iterative map to achieve an optimal solution for multiple DG sizing and placement in distribution networks, as well as an optimal load shedding approach. Regarding DG placement, the objective function was to minimize total active power loss and voltage deviation of the network nodes. The proposed method was compared with Modified moth flame optimization (MMFO), Teaching learning based optimization (TLBO) and the original Equilibrium optimizer (EO). Moreover, to assess the optimal load shedding technique, a constrained function with total remaining load and static voltage stability margin (VSM) index was used. In addition, the proposed CEO algorithm is compared with some of the recent metaheuristics algorithms applied in this domain such as Grasshopper optimization algorithm (GOA), Backtrack search algorithm (BSA) and the original Equilibrium optimizer (EO). In the last part of the work, based on a new metaheuristic known as the Artificial hummingbird algorithm (AHA), this work provides a novel approach for addressing the problem of RDG planning optimization. Considering various operational constraints, the optimization problem is developed with multiple objectives including power loss reduction, voltage stability margin (VSM) enhancement, voltage deviation minimization, and yearly economic savings. Furthermore, using relevant probability distribution functions, the ambiguities related with the stochastic nature of PV and WT output powers are evaluated. The proposed algorithm was compared to two of the recent metaheuristics applied in this domain known as improved harris hawks and particle swarm optimization algorithm (HHO-PSO) and hybrid of phasor particle swarm and gravitational search algorithm (PPSOGSA). The IEEE 33-bus and 69-bus systems are assessed as the test systems in this study. According to the findings, AHA delivers superior solutions and enhances the techno-economic benefits of distribution systems in all the scenarios evaluated.Item Impact of Joint Transmission (JT) Coordinated Multipoint (CoMP) on Mobile Users in 5G Heterogeneous Network(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Tandra, Tahmina Khanom; Tajrian, Fehima; Hossain, AfiaThe vision to ensure ubiquitous connectivity with ultra-reliable low latency, inconceivably high data rate, and support a myriad of data-hungry devices is foreseen with the widespread rollout of the 5G network. Ensuring seamless connectivity at the cell edge amidst the significant prevalence of Intercell Interference (ICI) and path loss proves to be complicated. In addition, the impact of mobility poses particular challenges to the wireless network and the high frequency of 5G networks limits the coverage area. With increase in UE mobility, Doppler effect becomes significant enough to impair the mean data rate and induce call drops. Joint transmission Coordinated (JT CoMP) is a promising ICI mitigation technique where several eNBs coordinate to create a virtual antenna array and transmit downlink (DL) data simultaneously to serve the UEs with strong radio signal links. The transmitted signals from the coordinated eNBs act as the desired signal for the UEs, reducing the interference of undesired signals. This paper examines the influence of JT CoMP technology on user velocities by incorporating closed loop spatial multiplexing (CLSM) into the Heterogenous Network (HetNet) with the aim of improving signal reception at cell edge and minimizing the effect of ICI for mobile users. To realize the effectiveness of inter-site and intra-site JT CoMP schemes in ICI mitigation and boosting cell edge throughput for mobile users in HetNet, the simulation was conducted for HetNet with CoMP and non-CoMP deployment. With the proliferation of UE velocity, the performance of CLSM degrades and less detailed feedback is reported. On the contrary, the simulation results reveal that CLSM integrated intrasite based JT CoMP offers better signal reception for low velocities and intersite based JT CoMP provides better throughput at high velocities.Item A Comprehensive Investigation of the Performances of Different Machine Learning Classifiers with SMOTE-ENN Oversampling Technique and Hyperparameter Optimization for Imbalanced Heart Failure Dataset(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Al-Monsur, Abdullah; Ratul, Ishrak Jahan; Ar-Rafi, Abrar MohammadThe chronic cardiac condition myocardial infarction (heart failure) is characterized by decreased blood supply to the body as a result of the heart muscles’ impaired contractile properties. Patients with heart failure, like those with any other cardiac disorder, have difficulty performing daily activities and have a shorter life expectancy, with the vast majority of cases resulting in death at some point during the patient’s lifetime. Treatment outcomes and patient quality of life improve significantly when patients with heart failure are identified early and are likely to survive. As a result, machine learning techniques can be extremely beneficial in this situation because they can be used to predict the survival of heart failure patients in advance, allowing patients to receive the most appropriate treat- ment at the earliest possible stage. As a result, six supervised machine learning algorithms were applied to a dataset of 299 people from the University of California, Irvine Machine Learning Repository in order to predict their chances of surviving heart failure. There were a variety of algorithms used in this study including Decision Tree Classifier, Logistic Regression, Gaussian Nave Bayes, Random Forest Classifier, K-Nearest Neighbors, and Support Vector Machine, among others. Prior to scaling the data, a preprocessing step was carried out, and both the standard and min-max scaling methods were employed. When it came to optimizing the hyperparameters, the techniques grid-search cross validation and random search cross validation were combined. Data resampling techniques such as the edited nearest neighbor (SMOTE-ENN) and synthetic minority oversampling (SMOTE) data resampling are also employed (SMOTE-ENN). It has been thoroughly compared and analyzed the outcomes of all of the different approaches. As a result of these findings, the Random Forest Classifier (RFC) outperforms all other approaches, achieving a test accuracy of 90 percent when compared to the other approaches when SMOTE-ENN and the standard scaling technique are employed. With the help of an imbalanced dataset, this comprehensive investigation vividly illustrates the application and compatibility of sev- eral machine learning algorithms. Among the methods for improving the performance of machine learning algorithms discussed in this investigation are the SMOTE-ENN algo- rithm and hyperparameter optimization.