Hüseyin Demirel

Prediction of the brain temperature from other body temperatures in hypothermia induced rats by using artificial neural networks

In this study, a microcontroller based temperature controlled hypothermia induction system is realized. The novelty of the experimental system is in using a thermoelectric Peltier cooler to bring the body temperature to hypothermic levels. The right and left ear, rectum, and brain temperatures of a rat are measured with this system. Then, an artificial neural network (ANN) is trained using temperature data to predict brain temperature from other body temperatures. It is shown that the system has the capability to predict the brain temperature with low error from rectum temperature measurements. The results from this empirical prototype also show that it is possible to use the system on humans to induce local hypothermia safely, where an interior temperature probe in the scalp is eliminated.

Performance and cooling efficiency of thermoelectric modules on server central processing unit and Northbridge

Display Omitted The cooling efficiency of the specific cooling systems on critical server hardware is investigated.The performance of the cooling systems on server CPU is outlined.The applied cooling systems in the context of temperature and performance are compared.The superior performance rating and cooling efficiency is obtained by the thermoelectric cooler. Computer systems are required to process data more rapidly than ever, due to recent software and internet technology developments. The server computers work continuously and provide services to many clients simultaneously, which results in greater heat production and high temperature that must be managed in order to avoid malfunction and failure of critical hardware. In this study, three cooling systems were used comparatively to examine the temperature and performance of the CPU and motherboard. The temperature characteristics and performance of the CPU were tested with a heat sink, water cooling system, and thermoelectric cooler. According to the test results, the thermoelectric cooling system has better cooling performance than the other two systems under continuous operating conditions. Additionally, the performance rating of the CPU was the best with a thermoelectric cooler under varying workloads.

Design, Implementation and Evaluation of Fuzzy Logic and PID Controllers for Fuel Cell Systems

In this paper, fuel cell control is investigated in addition to the use of fuzzy logic to control fuel cells. For fuzzy rules, the maximum power point tracking algorithm is used. Additionally, PID control is used and tested in this paper. As simulation results show, the performance of fuzzy logic is better than PID control. In general, for fuel cell systems, humidification is required for the air or the hydrogen, or both the air and hydrogen at the fuel cell inlets. Moreover, water content is very important for the protonic conductivity in the proton exchange membranes. If membrane dehydration or drying occurs, electrical performance decreases due to significant ohmic losses. 

Multi-criterion design and 2D cosimulation model of 4 kW PM synchronous generator for standalone run-of-the-river stations

This study reports the analytical computation including performance characteristics such like load line voltage and output power in combination with coupled-field circuit analysis of a 4kW direct drive permanent magnet synchronous generator (PMSG) to be used in a micro-scale run-of-the-river-station application. The specifications such like slot opening, pole embrace and magnet length of PMSG are optimized by using parametric approach including multi-criterion design optimization. Based on the optimized design, the model has been exposed to some transient coupled-field circuit analyses based on variable river flow speed and variable ohmic load conditions. The analytical studies related to finite element methods and conducted parametric approaches verified the effectiveness of the employed dynamic co-simulations.

A VHDL application for kinematic equation solutions of multi-degree-of-freedom systems

As kinematic calculations are complicated, it takes a long time and is difficult to get the desired accurate result with a single processor in real-time motion control of multi-degree-of-freedom (MDOF) systems. Another calculation unit is needed, especially with the increase in the degree of freedom. The main central processing unit (CPU) has additional loads because of numerous motion elements which move independently from each other and their closed-loop controls. The system designed is also complicated because there are many parts and cabling. This paper presents the design and implementation of a hardware that will provide solutions to these problems. It is realized using the Very High Speed Integrated Circuit Hardware Description Language (VHDL) and field-programmable gate array (FPGA). This hardware is designed for a six-legged robot and has been working with servo motors controlled via the serial port. The hardware on FPGA calculates the required joint angles for the feet positions received from the serial port and sends the calculated angels to the servo motors via the serial port. This hardware has a co-processor for the calculation of kinematic equations and can be used together with the equipment that would reduce the electromechanical mess. It is intended to be used as a tool which will accelerate the transition from design to application for robots.