The increasing concerns surrounding demagnetization risk and fault tolerance in direct-drive in-wheel motor applications prompt a detailed investigation. This paper delves into the demagnetization risk of Permanent Magnets (PM) and fault tolerance for three motor topologies sharing similar characteristics: diameter, volume, output power, and the number of PMs. The three configurations, namely the Internal Single Teeth Rotor (ISTR), Internal Dual Yoke Rotor (IDYR), and External Dual Yoke Rotor (EDYR), are thoroughly analyzed under various operating conditions, providing a comprehensive comparison from the perspectives of PM demagnetization risk and fault tolerance. The analysis encompasses eccentricity rotor fault examinations to assess fault tolerance, thermal analysis to evaluate thermal performance, and an all-encompassing demagnetization assessment under diverse operating conditions for these three topologies. The findings are substantiated through rigorous 3-D finite element method (FEM) validation.
28 November 2024 in Computers and Electrical Engineering
Rece with t advancements in modern technology have significantly enhanced the transmission of information, particularly in image processing, utilizing deep learning algorithms. This study aims to propose a a robust deep-learning strategy for detecting and recognizing eye defects and diseases from medical images. We present a detailed practical simulation of hybrid deep learning techniques designed for medical image classification based on multi-descriptor algorithms. The focus is on the classification of eye diseases by applying an advanced deep-learning algorithm to a dataset comprising various pathological eye conditions. Training operations for the proposed algorithm were conducted following the initialization phase, which included the extraction of multi-specification features. This enables the deep learning model to effectively analyze input eye images and accurately diagnose conditions. Our results demonstrate a diagnostic efficiency of 99%, with an error rate not exceeding 0.015%. The findings underscore the high efficiency and accuracy of deep learning algorithms in classifying and analyzing image data, thereby significantly reducing the workload for healthcare professionals.
Electro surgical unit is a popular modern device. It has been used in operating rooms for cutting, fulguration and coagulation of human tissues. ESU generates high frequency alternating current to prevent the stimulation of nerves and muscles. The objective of this article was to improve the performance of an ESU by controlling its output power under the variation of tissue impedance using proportional integral derivative controller based on particle swarm optimization to achieve minimum overshoots and fast dynamic response. The controller was simulated in MATLAB/SIMULINK to demonstrate the superiority of the suggested method. A comparative analysis was presented with ESU utilizing manual tuning process. The results showed that the proposed controller offered best performance utilizing manual tuning method. Moreover, both of the tuning methods presented better results from open-loop controller as a result of which charring of tissues could be eliminated and clinical operations could be made more efficient.
Electrosurgical generators (ESGs) are currently the most widely used surgical technology for clinical operations. The main disadvantage of ESGs is their output power is irregular due to the variable tissue impedance. The heat dissipation caused by the high amount of thermal energy generated leads to medical complications for both patient and surgeon. In this research, various inverter topologies and power controllers are investigated to specify the best structure that ensures best performance. The type of inverter topologies investigated are three level and five level, while the PID structures investigated are integer order (IO-PID) and fractional order (FO-PID). The simulation results indicate that FO-PID with five level inverters is better than IO-PID with three level inverters in terms of minimum heat dissipation rate and THD of the output voltage and current.
1 February 2021 in IOP Conference Series: Materials Science and Engineering
The use of electroencephalogram (EEG) waves in the field of Brain Computer Interfaces (BCI) has recently attracted a lot of interest, with varied applications ranging from entertainment to medicine emerging, showing that EEG waves are suitable for controlling appliances through a microcontroller. This study presents an offline strategy for the extraction of digital control signals from brain EEG waves using MATLAB/SIMULINK. The principle is applied as follows: (i) the extraction of α- and β-bands from the original brain EEG wave; (ii) detection of the peaks of both bands; and (iii) calculation of the power density of both bands to classify the mental mode and to generate appropriate control signals. Simulation results including signal processing tracking for various EEG waves are demonstrated that verify the effectiveness of the proposed strategy in terms of using brain EEG waves for apparatus control.