中文摘要：

本研究之宗旨在探討經三種時效熱處理（固溶處理Condition A、頂時效處理H900及過時效處理H1150）之17-4 PH不銹鋼其高溫疲勞裂縫成長行為。實驗結果顯示，在荷重比為0.1及負荷頻率2 Hz的條件下，於室溫至500oC之間，H1150的疲勞裂縫成長速率會隨著溫度的上升而增加，但Condition A與H900於500oC之疲勞裂縫成長速率反而比較低溫度之值為低。在300和400oC，比較三種熱處理之疲勞裂縫成長速率，均為H900最高，Condition A次之，H1150最低。在500oC，三種熱處理之疲勞裂縫成長速率則非常接近，Condition A及H900的疲勞裂縫成長速率於500oC時的異常變化，主要由即時之析出粒子粗化效應所造成。

在負荷頻率對高溫疲勞裂縫成長行為的影響方面，在500oC，三種熱處理在各頻率（0.002 ~ 2 Hz）下都有相近的疲勞裂縫成長速率，且在頻率低於2 Hz時其疲勞裂縫成長速率會隨頻率的下降而上升。此頻率效應是由裂縫尖端氧化與循環負載的加乘作用所造成，頻率越低，則氧化的加乘效果越明顯。在400oC，雖然三種熱處理的疲勞裂縫成長速率並不相同，但都會在較高頻率時隨頻率下降而上升，而在較低頻率時不受頻率效應影響。此三種熱處理於400oC時，呈現不同於一般材料的頻率效應，則是由動態應變時效（DSA）所造成。另外，在各個頻率下，H1150的疲勞裂縫成長速率會隨著溫度的上升而增加，而Condition A與H900於500oC之疲勞裂縫成長速率會較400oC之值為低，此現象如前述是由即時之析出粒子粗化效應所造成。

在荷重比對高溫疲勞裂縫成長行為的影響方面，在400oC、2 Hz的條件下，三種熱處理的疲勞裂縫成長速率在低ΔK區域皆會隨荷重比的增加而上升，且在高ΔK區域不受荷重比效應影響。唯H900在高ΔK區域時其疲勞裂縫成長速率仍會隨荷重比的增加而上升。在500oC，H1150的疲勞裂縫成長速率在較低荷重比時會隨荷重比的上升而增加，但在較高荷重比時不受荷重比效應所影響。若以統合分析方法（Unified Approach）對三種熱處理之17-4 PH不銹鋼進行分析，可發現其疲勞裂縫成長速率隨荷重比的變化，純粹是因材料之裂縫成長機制需同時滿足兩個裂縫成長門檻值所反應出來的結果。三種熱處理在不同溫度下都有共通的裂縫成長機制，並與一般疲勞裂縫成長之塑性鈍化成長機制（plastic blunting process）相似。唯Condition A和H900在400oC、高荷重比的條件下，其疲勞裂縫成長亦會受結晶面破裂（crystallographic-faceted）機制所影響。此外，各種熱處理於不同荷重比下的疲勞裂縫成長速率可用一荷重比整合參數（M*ΔK）加以整合，有相當好的效果。使用此整合參數配合最佳擬合曲線，可以用來預測各種條件下之17-4 PH不銹鋼在任意荷重比下的疲勞裂縫成長速率。

 

Abstract：

High-temperature fatigue crack growth (FCG) behavior was investigated for 17-4 PH stainless steels in three heat-treated conditions, namely unaged “Condition A,” peak-aged “Condition H900,” and overaged “Condition H1150.” Baseline FCG behavior was studied at a load ratio of 0.1, a frequency of 2 Hz and temperatures ranging from 300 to 500oC. The fatigue crack growth rates (FCGRs) of Condition H1150 were increased with an increase in temperature from 300 to 500oC. However, for Conditions A and H900 tested at 500oC, the FCGRs were lower than the lower-temperature ones. At 300 and 400oC, H1150 and H900 generally showed the lowest and highest FCGRs, respectively, with Condition A demonstrating behavior between the two. At 500oC, the FCGR curves for all heat-treated conditions merged together. The anomalous FCG behavior of 17-4 PH stainless steels at 500oC was mainly caused by an in-situ overaging and precipitate-coarsening effect during test.

The influence of frequency on FCG behavior was investigated at 400 and 500oC with a load ratio of 0.1 and frequencies ranging from 0.002 to 20 Hz. At 500oC, all heat-treated conditions exhibited similar FCG behavior in which no frequency effect was observed at frequencies higher than 2 Hz and the FCGRs increased with decreasing frequency below 2 Hz. The increase in FCGR at a lower frequency at 500oC was thought to be caused by a time-dependent, oxidation-assisted cracking mechanism. At 400oC, for a given heat-treated condition, the FCGRs increased with decreasing frequency at higher frequencies and leveled off at lower frequencies. Such an anomalous FCG behavior was attributable to a dynamic strain aging (DSA) effect. At a given frequency, when the temperature was increased from 400 to 500oC, the FCGR increased in Condition H1150, but decreased in Conditions A and H900 due to the aforementioned in-situ overaging and precipitate-coarsening effect during test.

The load ratio effect on FCG behavior was investigated at 400 and 500oC with a frequency of 2 Hz and load ratios ranging from 0.1 to 0.7. The FCGRs increased with load ratio at lower ΔK regime and merged together at higher ΔK regime for all given heat-treated conditions tested at 400oC, except for Condition H900, in which the FCGRs increased with load ratio for the whole ΔK regime. At 500oC, the FCGRs for a given heat-treated condition increased with load ratio at lower load ratios and became comparable at higher load ratios for the whole ΔK regime. The variations of FCGR with load ratio could be described as an intrinsic behavior of FCG when analyzed using a Unified Approach. Accordingly, the dominating crack growth mechanism was similar to the general plastic blunting process except for Conditions A and H900 at 400oC at the high load ratio regime, in which the dominating crack growth mechanism was considered as a crystallographic-faceted fracture mode. The variations of FCGR with load ratio for all given testing conditions at 2 Hz could be correlated with a modified fatigue crack driving force parameter, M*ΔK. Good correlation results were obtained and the FCGRs at any load ratio for all the given heat-treated 17-4 PH alloys could be predicted by such an M*ΔK parameter with the best-fitted relations.