球面滾子軸承具有具有自調心、承載能力高、摩擦小、壽命長等特點。球面滾子軸承可以承受徑向負載和軸向負載，自調心功能可以適應平衡，但不適用於高速應用。本研究使用SolidWorks繪圖軟體建構3D實體圓錐滾子軸承，並進行有限元素模擬分析。使用不同軸承材質分析，結果顯示安全係數大小順序為：不鏽鋼>軸承鋼>混合型，模態頻率大小順序為：軸承鋼>混合型>不鏽鋼。當軸承滾子數目及軸承寬高比增加時，接觸應力和變形量皆為下降趨勢，安全係數、模態頻率皆為上升。期望本研究能提高球面滾子軸承精度、品質及使用壽命，提升業界對軸承設計與製造能力。

In the recent years wind turbines have gained much attention because of the depletion of nonrenewable resources. Many countries today use wind turbines due to its sustainability, cost efficiency and is also less harmful to the environment. Wind turbines are thus an important device to produce energy. These wind turbines however can get affected during stall induced vibration produced by strong winds with varying wind speed. To study this effect, we researched on the effect of wind turbine airfoil during stall induced vibration. In this research we consider the pitching mode of oscillation of a wind turbine airfoil. The type of airfoil used for analysis is the NACA 0012 airfoil. This type of airfoil is one of the most commonly used airfoil in wind turbine blades. The type of stall which is being analyzed in this paper is the dynamic stall. Dynamic stall occurs when the airfoils rapidly change the angle of attack. With dynamic stall we get an increase in dynamic loads, particularly in case of horizontal axis wind turbines. Dynamic stall effects the effective power output of wind turbines. Dynamic stall results in fatigue failure of the airfoil. This paper explores the nonlinear behaviour present in the system by constructing the bifurcation diagram, phase diagram, frequency spectrum plot Poincare map and Lyapunov exponent curve. The analysis is first performed with parametric variations and the resulting bifurcation behaviour is noted. The bifurcation diagram is analysed further by using different tools such as phase plot, Poincare map and Lyapunov exponent maximum Lyapunov exponent is the key index for the occurrence of chaos and can be applied for prediction using Gaussian processes regression (GPR) and the back propagation neural network (BPNN). The system shows the presence of a significant nonlinear behaviour with the introduction of structural nonlinearity parameter. The presence of structural nonlinearity has a greater effect in the bifurcation pattern of the system.

This study focuses on the nonlinear dynamic vibration analysis of cutting machines. Inclusive of the following nonlinear effects for cutting process, i.e. nonlinear chatter effect, nonlinear stiffness of structure, nonlinear damping effect and nonlinear cutting forces are taken into consideration in this study. The stability analysis and the modeling for multi- degree of freedoms coupling systems are also taken into consideration in this study. After finding out the non-periodic dynamic responses (e.g. chaotic motions), those parameters will be useful information for scientists and engineers how to control those nonlinear systems and then even get rid of the undesired dynamic responses. The dynamic systems in our nature are actually nonlinear but we often assume or simplify those systems to be linear systems. The error of linearization assumption in weakly nonlinear systems may be tiny, but it may causes significant error in strong nonlinear systems. All the dynamic equations, the related interaction forces and nonlinear effect have strong nonlinearity in the cutting machines (i.e. turning, milling and drilling). Therefore it will help future development in the precision manufacturing by observing and studying the nonlinear dynamics. Besides, the above-mentioned research is not investigated or studied in the past. The study will develop the theoretical nonlinear vibration model and simulate it. It can provide a theoretical foundation and build dynamic parameters to help the development on precision manufacturing in the future.

本研究使用聚乳酸（Polylactic Acid，PLA）並利用3D列印的方式來製作螺旋槳。藉由改變PLA材料在3D列印時的填充率（10%~100%）來驗證在不同填充比例下，PLA材料經3D列印後拉伸強度的變化並進一步將不同填充比例來製備3D列印螺旋槳，藉由探討螺旋槳對軟質塑膠管的高速切割後的損傷情形來瞭解不同塑料填充比例於螺旋槳葉片的實際應用。希望未來以3D列印製備葉片在執行工業巡檢任務時，可以更加的安全可靠。

In micro-drawing process, when designing and making small-sized molds, the gap between the punch and the mold is extremely small, and the processing technology often uses ultra-precision micro-discharge machining to meet the accuracy requirements. In order to obtain an accurate finished product size and shape, it is necessary to rely on the control of various variables in the process, including the problem of spring back after load removal, control of the accuracy of the finished product size and shape, the occurrence of cracks in the stretching process and the estimation of the punch load. Therefore, in this paper, SUS304 stainless steel plate is taken as the research object, consider the influence of scale effect on the thin plate through the ULF theory and construction of finite elements, and simulate the behavior of the plate with a new micro-elastoplastic material model, and compare this model with traditional materials. The difference of the model. Then use the LS-DYNA solver for simulation analysis, and perform pre-processing and post-processing through Dynaform to obtain material deformation history, thickness change distribution, and material stress and strain distribution maps. Finally, the simulation results are compared with the experimental results to confirm the correctness of the elastoplastic deformation three-dimensional finite element analysis. The results show that when the sheet is at the 0º section position, the thickness of the material changes during the stretching process. The theoretical and experimental results The error is not large, about 0.005~0.014mm. Compared with the 0º section, at the 45º section position, the plastic flow of the material in the peripheral area of the blank is worse, and the blank will have a larger thickness when it enters the fillet area of the lower die. The overall theoretical and experimental errors are also in Within the range of 0.005~0.01mm. This result can be used as an important reference for the design and processing of micro-sized molds.

本研究第一部份為物件偵測之級聯分類器，使用OpenCV影像函式庫與哈爾特徵建立級聯分類器之銑削刀具偵測模型，進行銑削刀具偵測，在即時影像中，將偵測到的銑削刀具框出；第二部分為銑削刀具分類預測模型，使用卷積神經網路建立銑削刀具分類預測模型，透過卷積神經網路架構中各種參數之調整以提高模型準確率，將第一部分之銑削刀具偵測模型結合銑削刀具分類預測模型，即得到本研究最終產出之AI即時刀具辨識系統，並利用混淆矩陣評估此系統的各項指標。