本研究在於探討如何利用類神經演算法以配合材料數值模型分析之需求，來分析出未來產品設計所需之關鍵材料參數或製程參數，如超薄晶圓(25 μm – 75 μm)於針測時之可靠度設計準則，以及發泡材料於中高速率下中高應變率(100 and 102 s?1)之應力-應變參數。 在晶圓針測上，超薄晶圓針測可靠度一直只能依靠實際針測測試，才能評估出可靠之製程參數，本研究運用舊有之實際針測行程，搭配發展之超薄晶圓強度機械性質測試，進而使用數值分析模型獲取於針測行程中相關參數對晶圓破裂強度所帶來的影響，之後並針對不同超薄厚度晶圓強度與多層針型參數，使用類神經演算法以得到一針測製程參數預測模式，使其加快針測工程師快速設計針測參數與針型之依據。 在發泡材料中高速率衝擊測試上多是使用落下測試(Drop Test)，然而落下測試僅僅只能獲取材料之回彈高度等參數，尤其當材料外型幾何變得複雜時，便無法確認材料之緩衝能力，故本研究基於舊有落下測試下運用高速攝影機搭配低密度發泡材料數值模型，以及類神經演算法，建立起新的一套能夠測試出發泡材料在中高速應變率下之各應力-應變參數的方法。 由兩項研究實驗與模擬驗證結果可知，基於舊實驗方式再配合開發出之新式實驗技術，以及數值模型搭配類神經演算法的分析下，可得到以往實驗中無法獲取之關鍵材料與製程可靠度設計參數，進而提供設計者或工程師有效之產品設計。 In this paper, the finite element method is employed in conjunction with the neural network to predict the thin wafer probing needle shape parameter and EVA foam material stress-strain curve at different specific strain rates.Mechanical contact caused by using excessive probe force produces an oversized scrubbing mark that may result in damage to the die pad and the silicon chip breaking for thin wafer. Therefore, investigating the relationship between the wafer thickness and the limited breaking stress of the wafer, and applying this relationship as a basis for establishing suitable design rules for a multilayer needle layout are crucial. In this paper, two experimental techniques, Three-Point-Bending (TPB) test and Ball-On-Ring (BOR), were set up and carried out to measure the force-displacement relation of various wafer thickness (100 ?m – 25 ?m). The results from the testing then coupled with finite element analysis to reverse finding the breaking stress/strain as a function of wafer thickness. In addition, experimental set up with a single tungsten needle probe contact with Al pad were employed to investigate the relations between the overdrive, beam length and scrub mark length. An Neural Network inverse analysis technique combined with a finite element model for probing process parameter identification was developed to determine the multilayer needle shape.EVA foams, like all other polymers, also exhibit strain-rate effects and hysteresis. However, currently available approaches for predicting the mechanical response of polymeric foam subjected to an arbitrarily imposed loading history and strain-rate effect are highly limited. Especially, the strain rates in the intermediate rate domain (between 100 and 102 s?1) are extremely dif?cult to study. The use of data generated through the drop tower technique for implementation in constitutive equations or numerical models has not been considered in past studies. In this study, an experiment including a quasi-static compression test and drop impact tests with a high speed camera was conducted. An Neural Network inverse analysis technique combined with a finite element model for material parameter identification was developed to determine the stress–strain behavior of foam at different specific strain rates.