| 摘要: | 本研究以常用的積層製造17-4 PH不鏽鋼為例，進行破壞韌性、固定及變動負荷振幅疲勞裂縫成長、超載負荷及裂縫閉合等實驗。量測多種疲勞理論預測模式中之材料參數及提出新的預測模式。利用多種不同特性的負荷歷程，其疲勞裂縫成長實驗值與各種理論模式之預測壽命值進行比較，歸納出在不同負荷歷程特性下之最佳的疲勞裂縫成長預測模式。此外，亦探討積層製造17-4 PH不鏽鋼之金相組織及疲勞破壞機制。
研究結果顯示，積層製造的17-4 PH不鏽鋼的破壞韌性能達到80.37 MPa√m，比相同熱處理條件下的鑄造件更高。固定負荷振幅疲勞裂縫成長方面，本研究新提出的裂縫閉合方程式，比Elber模式及Schijve 模式對於平均應力之修正更佳。超載負荷比在2.0及2.5時，在每個施加超載負荷之位置，都能觀察到明顯的裂縫成長遲滯現象。Wheeler 模式之材料參數p會受到超載負荷應力強度因子影響，關係式為p = 0.145KOL - 1.757。當超載負荷應力強度因子在22至47 MPa√m區間時，p之平均值為3.04，可更方便用於超載負荷下之疲勞裂縫成長預測。在裂縫閉合量測方面，裂縫閉合因子會隨著應力比R及應力強度因子∆K提高而增加，方程式為∆Keff/∆K = 0.023∆K + 0.758R + 0.186。在預測SAE三種變動負荷歷程作用下之壽命時，Schijve 模式及本研究提出的裂縫閉合模式有較佳的疲勞壽命預測結果。本研究提出的裂縫閉合模式，對於在傳動軸TRN歷程下之壽命預測最佳，誤差為6.92%。Schijve 模式對於在支架BRK及懸吊SUS歷程下之壽命預測最佳，誤差分別為2.25%及2.53%。負荷次序效應方面，利用傳動軸TRN歷程重新排列成應力振幅由大到小(Hi-Lo)及由小到大(Lo-Hi)兩種歷程。在Hi-Lo歷程作用下有較長之疲勞壽命，可利用疲勞裂縫閉合效應解釋，但三種歷程的疲勞壽命只差13%，歸因於本研究三種歷程引起的次序效應差異不大。在金相、成分分析及疲勞破斷面的觀察上，可以得知積層製造17-4PH不鏽鋼經過CA-H900熱處理後，99%為麻田散鐵組織，大部分呈現板狀結構。固定負荷振幅歷程下之疲勞裂縫穩定成長區有明顯的疲勞紋。在變動負荷振幅歷程的破斷面，當∆Kmax在24~40 MPa√m之間，能區分出每個負荷區塊所造成的裂縫成長。在超載負荷試片的破斷面上，當超載負荷比相同，則超載負荷應力強度因子越高其暗帶也越明顯。在高負荷及裂縫較長的情況下，超載負荷造成的暫時加速區有高度塑性區的撕裂及韌窩狀組織。
;This study employs the commonly used additive manufacturing method with 17-4 PH stainless steel to conduct experiments on fracture toughness, fatigue crack growth under both constant and variable load amplitudes, overload, and crack closure. Various fatigue theoretical prediction models are utilized, and material parameters are measured to propose new predictive models. By comparing fatigue crack growth experimental values under different load cycle characteristics with predicted life values from various theoretical models, the study aims to identify the optimal fatigue crack growth prediction model for different load cycle characteristics. Additionally, the study investigates the microstructure and fatigue failure mechanisms of additively manufactured 17-4 PH stainless steel.
The results show that the fracture toughness of additive manufacturing 17-4 PH stainless steel can reach 80.37 MPa√m, which is higher than that of cast parts under the same heat treatment conditions. In terms of constant amplitude load fatigue crack growth, the new crack closure equation proposed in this study provides better correction for mean stress compared to the Elber model and the Schijve model. When the overload ratio is 2.0 and 2.5, significant crack growth retardation can be observed at each overload location. The material parameter p in the Wheeler model is affected by the overload stress intensity factor, with the relationship being p = 0.145KOL - 1.757. When the overload stress intensity factor is in the range of 22 to 47 MPa√m, the average value of p is 3.04, which can be more conveniently used for fatigue crack growth prediction under overload conditions. Regarding crack closure measurements, the crack closure factor increases with the stress ratio R and the stress intensity factor ΔK, with the equation being ∆Keff/∆K = 0.023∆K + 0.758R + 0.186. When predicting the life under three SAE variable loading histories, the Schijve model and the crack closure model proposed in this study provide better fatigue life prediction results. The crack closure model proposed in this study offers the best life prediction for the transmission shaft (TRN) history with an error of 6.92%. The Schijve model provides the best life prediction for the bracket (BRK) and suspension (SUS) histories, with errors of 2.25% and 2.53%, respectively. For load sequence effects, the transmission shaft (TRN) history was rearranged into two sequences: from high to low stress amplitude (Hi-Lo) and from low to high stress amplitude (Lo-Hi). Under the Hi-Lo sequence, a longer fatigue life is observed, which can be explained by the fatigue crack closure effect. However, the fatigue life difference among the three histories is only 13%, attributed to the minor sequence effect differences caused by the three histories in this study. Observations of metallography, composition analysis, and fatigue fracture surfaces reveal that 99% of the additive manufacturing 17-4 PH stainless steel after CA-H900 heat treated consists of martensitic structure, mostly in a plate-like form. There are obvious fatigue striations in the stable growth region of the fatigue crack under constant amplitude loading history. On the fracture surface of variable amplitude loading history, when ∆Kmax is between 24 and 40 MPa√m, the crack growth caused by each load block can be distinguished. On the fracture surface of the overload specimens, the higher the overload stress intensity factor under the same overload ratio, the more distinct the dark bands. Under high load and long crack conditions, the temporary acceleration zone caused by overload has highly plastic tear ridge and dimples structures. |
