الفهرس 
SummaryOnly 14 pages are availabe for public view
 |  | from | 159 |  |  |
| | | from | 159 | | |
Abstract
In the present investigation an experimental apparatus is designed and constructed to simulate the cooling process through the channel in the core of Egypt Test Research Reactor No.2 (ET-RR-2). The measurements included the effect of operating conditions such as the pool temperature, heat flux and geometrical parameters, in particular the height of the chimney (extension ratio), on the coolant velocity, heat transfer coefficient, coolant and surface temperatures in the natural convection-cooling mode where the flow was upward. The cooling channel is made from two vertical parallel plates aluminum alloy 6061 used as heating surfaces through a vertical rectangular channel. The channel surfaces are provided with 54 thermocouples to 11 thermocouples distributed inside the coolant flowing through the coolant channel.
Experiments are carried out under the atmospheric conditions (open pool). Experimental measurements were done for free convection from vertical isoflux, and parallel-walled channels. To explore the heat transfer enhancement is due to adding adiabatic extension (chimney) of various sizes (15, 35 and 55 cm) to the coolant channel. The corresponding extension ratios are 1.2, 1.41 and 1.62. The measurements indicated that for the same height of the chimney, the coolant velocity and the heat transfer coefficient increases with the increase in the heat flux at constant coolant inlet temperature. At constant heat flux the coolant velocity increases and the heat transfer coefficient is enhanced by the increase of coolant inlet temperature, i.e. the reactor operates more efficiently in the summer conditions in natural circulation cooling mode. A computer program, is written in Fortran to calculate the flow velocity, local temperatures of coolant and surface as well as the local and average heat transfer coefficient and Nusselt number as a function of the operating conditions.
The present experimental measurements are also compared with model results, the RELAP5code results and the published data. 147
…………………………………………………………………………
149
This dissertation presents the result of an experimental/computational study of the combined effects of external environment and opening in natural ventilation and thermal comfort in buildings. The investigation was carried out on three different stages where two different building models were investigated. One model (denoted M1) represented an open space in a typical commercial building (i.g. an office building) with multiple and identical front and rear openings. The other model (denoted M2) represented a room in a typical residential building with a signal opening in each of the front and rear walls. In the first stage, both models were subjected to a series of tests using a smoke tunnel test facility in with a view to evaluate the effect of wind orientation and opening location and combination on the flow pattern inside the ventilated space. Although pictures of such flow patterns gave only a qualitative evaluation, they were quite useful in understanding and interpreting of the experimental findings from wind tunnel measurements.
In the second stage, measurements were taken in a wind tunnel inside which turbulent wind was simulated (wind simulator). This tunnel test facility was designed and constructed with local materials and know how. Suitability of this wind simulator for the purpose of the present work was first checked through extensive testing before carrying out the main investigation on building models.This investigation consisted of two parts. The first part was concerned with tests on model M1 for the effect of the window combination and arrangement on thermal comfort in the ventilated space under different external environmental conditions and fixed thermal load. In the second part, model M2 was subjected to a series of tests in which pressure distributions on exposed surfaces (front, rear, side walls, and window zone) were obtained for different external environmental conditions.
Here also thermal comfort characteristics of the ventilated space were examined for different window sizes and locations. For both models, single-sided and cross ventilation schemes were considered. Results indicated substantial effects of the external environmental conditions, window size and combination on thermal comfort level.
In the third stage of the work, a CFD-based software package (ANSYS FLOTRAN version)was applied to the present problem, where the above effects were predicted and compared with experiment. Comparisons showed good agreements, including that CFD techniques such used by ANSYS (finite element technique) can be reliably used in this type of problems. Also, comparison indicated that the experimentally-obtained results can be3e accepted with confidence, leading to solid conclusions. The present work gives important recommendations to architects and designers of new buildings with regard to combination, location and size of the openings in building front and rear facades for the best possible thermal comfort, depending on prevailing wind conditions and building orientation. Thermal comfort in an existing building, subject to analysis, may be improved through certain combinations of openings or by changing window size and location of allowed. Such improvement would be limited by the permissible changes and the prevailing wind characteristics. A thermal comfort chart derived in the present work should be quite useful both for designers and.
Finally, CFD techniques can be prove economical in problems of building thermal comfort, saving time and cost spent in experimental efforts, especially if all possible conditions are to be studied.
…………………………………………………………………………………………………………………..
131
The starting point of this thesis is a mathematical model of the vibration measurements and an overview of its use in different pumps bearing monitoring methods. This is one of several factors which determine the performance of the different methods.
. If the model is reliable then a (numerically stable) method which make full use of the model will most likely be among the best methods. Otherwise methods which depend on the model can be more robust and useful.
. The method with best overall performance is one which is able to detect single bearing fault impacts (if they are large enough), but also looks for some kind of cyclic behavior of potential impacts. A method designed for finding strictly periodic impacts. Will have problems finding inner ring faults in loaded bearings (because the load suppresses the periodicity of the impacts). On the other hand, if one aims for detection of individual impacts, then they must be much larger than all naturally occurring vibrations.
Chapter 5 is a more detailed overview of different mathematical tools which can be combined into an immense number of different bearing monitoring methods. from these, we have chosen reasonable combinations, implemented them in MATLAB and compared using test signals from both laboratory and industrial environments (see chapter 6).
Most promising results is presented in chapter 7 . We give examples of both
. plots of individual analyzed signals, which require expert knowledge for correct diagnosis,
. Simplified plots where every measured signal corresponds to one point in the plane and
. a simple way to inter prêt these plots is to split in two a classification line and give the diagnosis ”functional” or ”faulty” machine (pump) bearing, depending on which side of this line the corresponding point appears.
The last two steps were suggested to improved for better performance in an actual implementation.
The best performing method so far is one based on analysis of a continuous wavelet transform (CWT) of the input signal. This is not entirely surprising since an advantage of CWT (compared to the classical Fourier transform) is that it allows both to find potential individual impact and to look for some kind of cyclic behavior in these.
The tested methods were developed using a mainly different set of test signals and not optimized for those used in this thesis.
The classification line was, however chosen for these particular signals. This and the number of test signals (103) makes it difficult to estimate the precision of obtained misclassification rates. Still, it was a positive surprise that the best method gave wrong diagnosis for only 10% of the signals. Especially. Some especially important examples, are:
*Obtained results only from analysis of the shape of the measurement signals. Historic data is usually available, and should considerably improve the performance of any method in this report.
*A more careful adaptation of the measurement equipment to the measurement environment. For example, the sampling frequency should be as low as possible, for maximal resolution in the resulting frequency range, but high enough for this range to include the frequencies which are most for bearing monitoring bearing (thus the choice will be different for different machines).
*Adapting the methods to important properties of the input signals and environment, such as rotational frequency of the axis, sampling frequency and signal length.
*Further refinement of the signal model. An exact knowledge of the impulse response of the bearing-axis system would be especially useful for the statistical methods.