| Summary: | 碩士 === 遠東科技大學 === 電腦應用工程研究所 === 101 === This research used the CFD simulation technology to replace the actual experiment measurements and applied tow design methods, Taguchi robust design method and ANNs-GA Optimal method, to get the blade shape parameters of the single stage axial fans which have the best performance with high static pressure characteristics under a small flow rate. In the Taguchi method, the orthogonal array is applied and five control parameters with four levels which including the pretwist angle, the forward angle, the height of blade, the ideal lift coefficient and the number of blades. Perform the CFD simulation and analyze the results by the ANOVA method to find interaction effects between the control parameters. The optimal design parameters will recommend by the ratio of the bigger then better request after the ANOVA analysis. The optimal blade performance also will confirmed by the AMCA wind tunnel measurement and it shown that upgrade 92.87% with the static pressure and 31% with the large mass flow rate.
In the robust design results, there were shown that the forward angle of the blade rotor and stator is more important design factor for the high static pressure of the single stage axial fan. In the rotor blade ratio analysis, it was found that have the greater value will cause the lower static pressure performance and also found that decreasing the setting angle of blade and the pre-twist angle at the tip will get higher pressure performance relatively. On the other hand, the stator forward angle of the blade shown the opposite trend with rotor performance features, the higher angle value will improve the static pressure. In this Taguchi robust analysis, the recommend control parameters will improve the static pressure 92.87%. For the large mass flow rate goal, the pre-twist angle at the tip and the forward angle is more important design factor for the single stage axial fan. In the rotor blade ratio analysis, it was found that have the greater value will cause better maximum mass flow rate performance. As the Taguchi robust results shown, the recommend control parameters will raising 31% flow rate. For those two different blade design goals in Taguchi method that will cause different blade shape of axial fan to meet the aerodynamic characteristics requirements.
We were also used turbo machinery optimum tool Fine/Design3D to get best performance blade shape for the single-stage axial fan. The Fine/Design3D is using the Artificial Neural Networks (ANN) and the Genetic Algorithms (GA) combining method to guarantee to get the global optimum result. We defined seven blade profile parameters to construct a blade shape and set the operating conditions in flow solver as same as the Taguchi method. We assigned ten geometric control parameters, the torque value as constrain, the target value for the performance requirements, ANNs samples for the database generation and GA search iterations in Fine/Design3D system for the optimization procedure. This optimization result is shown the setting angle and pre-twist angle will promote the fan performance that is same design trend of the blade shape as Taguchi method. This optimal design was confirmed by AMCA wind tunnel which improved efficiency 3%, maximum flow 7.5% and torque 16% in measurement
|
|---|
