| Summary: | 碩士 === 國立成功大學 === 工程科學系碩博士班 === 96 === I n this thesis, design of experiments is used to plan the experimental evaluations of heat-pipe performance under different quality factors. The statistical software Design-Expert is utilized to validate the influences of quality factors. From these evaluations, the experiment model can be built to find out the quality factors of significant influence.
In this research, the D-Optimality of response surface methodology is used as the design method of quality problem and then the experimental regression formula can be obtained from the construction model. Afterwards, the experimental model is evaluated by the analysis of variance (ANOVA). From this analysis, the significant factors on the maximum heat transfer rate (Q’max) and then the optimal parameters can be obtained, which can promote the yield and quality of heat pipe.
From the analysis of the experimental data, the effects of different quality factors on heat-pipe performance can be evaluated. The pipe length has the highest influence on Q’max. As the length is increased from 18 to 25cm, the Q’max can be changed from 31.85 to 49.75 Watts. The pipe diameter has the second highest influence. As the diameter is increased from 6 to 8mm, the Q'max can be changed from 34.09 to 47.51Watts. The next important factor is the sintered-layer thickness. As the thickness is increased from 0.5 mm to 0.7 mm, the Q'max can be changed from 36.90 to 44.70Watts. The final factor is powder size. The Q'max only changes from 39.73 to 41.87 Watts when the powder size is increased from 45 to 100 mesh.
Finally, according to the prediction by using the construction model, the optimization conditions are the pipe length of 18 cm, diameter of 8 mm, sintered-layer thickness of 0.5 mm and powder size of 45 mesh. Under these conditions, the predicted value of Q'max is 72 Watts, the experimental one is 69.34 Watts and their relative error is 3.69%. From the results described above, it can be concluded that design of experiments is an effective way to plan the heat-pipe experiment, which can help to find the significant factors and give the working parameters of the optimal performance.
|
|---|
