Browsing by Author "Demirci, BA"

Now showing 1 - 5 of 5
  • No Thumbnail Available
    Item
    In vitro analysis of breast tumour detection using rotational infrared thermal imaging and machine learning techniques
    Demirci, BA; Engin, M
    Breast cancer is the most common cancer affecting women worldwide, and various methods, such as biopsy, mammography, 3D tomosynthesis, MRI, ultrasonography, and infrared thermal imaging (ITI), are utilized for its early detection. ITI is a technique that detects variations in thermal patterns on the breast surface, which are caused by the higher metabolic activity and vascularisation of cancerous cells. I As a radiation-free, non-invasive, and cost-effective screening method, ITI has been studied using in-silico, in-vivo, and in-vitro approaches to enhance its diagnostic performance and develop reliable imaging algorithms. Conventional ITI in in-vivo studies is limited by fixed imaging positions, making it difficult to detect deep or hidden tumours. To address these limitations, this study introduces a rotational ITI method integrated with machine learning algorithms in an in-vitro environment. The proposed method generates datasets with varying tumour depths for comprehensive algorithmic analysis. It captures thermal images from four distinct positions (0 degrees, 90 degrees, 180 degrees, and 270 degrees), enabling a more thorough evaluation of the phantom breast surface. Using the combined dataset, which incorporates information from all four positions, the Convolutional Autoencoder and Support Vector Machines methods achieved an accuracy of 98.28%, sensitivity of 97.75%, specificity of 98.82%, and an F1 score of 98.29%.
  • No Thumbnail Available
    Item
    Battery cell measurement and fault diagnosis system for detection of problem in automotive batteries
    Demirci, O; Demirci, BA; Taskin, S
    Increases in production volumes and higher quality product demands have led to the need of automated systems for continuously recurring tasks that are not directly dependent on people. This study is related to battery cell fault diagnosis system which is developed for passenger cars, light and heavy commercial vehicles. The battery used in vehicles commonly has a terminal voltage of 12 V. The battery consists of 6 cells connected to each other in series. Electrical and mechanical differences can be observed from time to time due to the production process parameters. These differences between cells cause performance problems including different internal resistances, capacity and cycle life. Moreover, these differences also require a time-consuming process in destructive battery problem analysis. So as to overcome such problems, a battery cell measurement and fault diagnosis system are required. In this study, batteries are tested according to the EN 50342-1 Battery Electrical Test Standard. The developed system user interface is designed with the LabVIEW (TM) program.
  • No Thumbnail Available
    Item
    Review of battery state estimation methods for electric vehicles-Part II: SOH estimation
    Demirci, O; Taskin, S; Schaltz, E; Demirci, BA
    State of Health (SOH) significantly determines the performance and durability of EV batteries, with Battery Management System (BMS) playing a crucial role in enhancing their efficiency and operational cycle life. This comprehensive review, the second part of our series on Battery State Estimation Methods for Electric Vehicles, provides an in-depth exploration of SOH estimation methods. SOH, which encompasses a battery's overall health, capacity, and aging characteristics, plays a fundamental role in making informed decisions, conducting proactive maintenance, and ensuring the safe and reliable operation of EVs. Diverse SOH estimation methods, ranging from data-driven to model-based approaches, address the multifaceted challenges associated with battery aging, including electrochemical processes, temperature variations, usage patterns, and external factors. In recent years, data-driven methods, especially those rooted in machine learning and artificial intelligence, have gained prominence. These methods facilitate the discovery of complex models and correlations, encompassing battery degradation and using datasets to train algorithms. Machine learning algorithms including Support Vector Machines (SVM), Artificial Neural Networks (ANN), and Deep Learning (DL), have shown significant promise in estimating SOH by learning from historical data and adapting to varying operational conditions. The studies highlighted in this review demonstrate significant advancements in SOH estimation techniques, leading to improved accuracy, efficiency, and adaptability. These advances contribute to the development of more reliable BMSs for EVs and battery energy storage systems.
  • No Thumbnail Available
    Item
    Comparative analysis of ANN performance of four feature extraction methods used in the detection of epileptic seizures
    Demirci, BA; Demirci, O; Engin, M
    Epilepsy, a prevalent neurological disorder characterized by disrupted brain activity, affects over 70 million individuals worldwide, as reported by the World Health Organization (WHO). The development of computeraided diagnosis systems has become vital in assessing epilepsy severity promptly and initiating timely treatment. These systems enable the detection of epileptic seizures by analyzing the electrical activity in the EEG recordings of the patients. In addition, it helps doctors to choose suitable treatment by quickly determining the type, duration, and characteristics of seizures and increases the patient's quality of life. The proposed computeraided diagnosis system in this study comprises three modules: preprocessing, feature extraction, and classification. The initial module employs a low-pass Chebyshev II filter to eliminate noise artifacts from signal recordings. The second module involves deriving feature vectors using Bispectrum Analysis, Empirical Mode Decomposition, Discrete Wavelet Transform, and Wavelet Packet Analysis. The third module employs the Artificial Neural Networks method for epileptic seizure detection. This study not only enables the comparison of feature extraction efficacy among Bispectrum Analysis, Empirical Mode Decomposition, Discrete Wavelet Transform, and Wavelet Packet Analysis techniques, but it also reveals that Bispectrum Analysis and Empirical Mode Decomposition yield the highest accuracy rate. The method achieves 100% accuracy in detecting epileptic seizures. Additionally, sensitivity analysis has been conducted to enhance the success of Discrete Wavelet Transform and Wavelet Packet Analysis methods and to identify significant features.
  • No Thumbnail Available
    Item
    Review of battery state estimation methods for electric vehicles - Part I: SOC estimation
    Demirci, O; Taskin, S; Schaltz, E; Demirci, BA
    This study presents a comprehensive review of State of Charge (SOC) estimation methods for Lithium -Ion (Li -Ion) batteries, with a specific focus on Electric Vehicles (EVs). The growing interest in EVs and the need for efficient battery management have driven advancements in SOC estimation techniques. Various approaches, including data -driven techniques, advanced filtering methods, and machine learning algorithms have been explored to enhance SOC estimation accuracy. The integration of artificial intelligence and hybrid models has shown promising results in improving SOC estimation performance. However, challenges remain in dealing with nonlinear battery behavior, temperature variations, and diverse operating conditions. Researchers are continuously studying to improve the robustness and adaptability of SOC estimation methods to address these challenges. The primary objective of this study is to provide an up-to-date summary of the latest advancements in SOC estimation, offering insights into innovative approaches and developments in this field. All existing SOC methods, their advantages, challenges, and usage rates have been comprehensively examined with a specific focus on EV battery management systems. As the EV market continues to expand, accurate SOC estimation will remain essential for optimal battery management and overall EV performance. Future research will focus on refining existing algorithms, exploring new data -driven techniques, and integrating advanced sensor technologies to achieve real-time and reliable SOC estimation in EVs.