在採用田口實驗計畫法前，設計參數必須採用六個合適的設計參數，此研究設計分別是導光管與反射面的距離、LED與Lens相對的距離、導光管的表面特性(AR-coating)、導光管的材質、導光管兩側切角及Lens的曲度等６個可控制因子，每個因子分別有５個水準做參數設計，以提升照明燈具之輻射照度。 選用田口實驗計畫法中的Ｌ25(56) 直交表來設定各項參數，利用Tracepro光學模擬軟體實際模擬，然後會取得２５組模擬結果，再將模擬結果放入直交表中，接著利用直交表的設計和變異數的分析，會找出實驗中主要效應分子為何，以及各可控參數間彼此關係對目標值的最佳化，將可有效減少模擬次數，將原本6參數5水準需做15625次模擬簡化為２５次，較短的時間內設計者即可獲得較佳的設計參數。 第一組初始模擬可得狀況之輻射照度(Irradiance)為3.6065 W/m2；利用田口實驗計畫法的直交表之望大特性分析比對後，由信號雜訊比可預測最佳設計參數狀況之輻射照度(Irradiance)為3. 9122W/m2，輻射照度獲得明顯提升。本研究確實減少對於非序列式光學元件的開發時間與設計成本，希望對光電產業有所幫助。 Before executing Taguchi method, we must estab¬lish the six parameters such as the gap y between the tube and reflective surface, the relative distance x between lens and LED, the characteristics of the tube surfaces (AR-coating), the material of the lighting tube, the tilt angle ψ of the end surface of the tube, and the Curvature of the lens. First, the above six parameters must be determined. Next, these parameters are divided into five levels separately. This research using Taguchi Method makes the optimization of non-sequence type optical element easier whereas the simulation times of the 6 parameters and the 5 levels of each parameter will be in total 15,625 times. We establish a product that can effectively re¬duce simulation times to 25. The luminous efficiency of the observation plane is larger than all of probable combination exactly. The original design of lighting tube without optimization by Taguchi method, the observation plane is only 262.68lm. With the optimal optimization design we can reach an output power of 287.81 lumens. Moreover the optimization by Taguchi method design can get the best construction in a short time. Therefore, this research for the non-sequence type optical ele¬ment decreases development time and the cost of production. We believe this research will be helpful for the photoelec¬tric indus¬try.
