Combined action of crack closure and residual stress under periodic overloads: A fractographic analysis. (January 2016)

• Record Type: Journal Article
• Authors: Sunder, R.; Andronik, A.; Biakov, A.; Eremin, A.; Panin, S.; Savkin, A.
• Is Part Of: International journal of fatigue. Volume 82:Part 3(2016)
• Page Start: 667
• Page End: 675
• Publication Date: 2016-01
• Subjects: Near threshold fatigue -- Near-tip residual stress -- Crack closure -- Variable-amplitude loading; Materials -- Fatigue -- Periodicals; Materials -- Fatigue; Periodicals; 620.1122
• DOI: 10.1016/j.ijfatigue.2015.09.025
• Languages: English
• ISSNs: 0142-1123

Abstract

Highlights: Near-threshold crack growth was fractographically studied under specially designed periodic overload sequences. Near-tip residual stress between overloads is determined by overload peak-valley sequence. Instantaneous Δ K th after overload is determined by near-tip residual stress. Growth rate is controlled by the independent action of crack closure and instantaneous Δ K th . High Δ K th combined with increasing closure can lead to crack arrest when overload peak follows valley. When overload valley follows peak, retardation is reduced primarily because of lower Δ K th . Abstract: The close relationship between sequence-sensitive near-tip residual stress and threshold stress intensity raises questions about load interaction models currently in use to estimate fatigue crack growth under variable amplitude loading. In an attempt to address them, experiments were performed on an Al–Cu alloy under specially designed load sequences with periodic overloads. Fractographic evidence from these tests confirms that fatigue crack closure, together with sequence sensitive variation in threshold stress intensity appear to explain all observed results. The fractographic data provide quantitative inputs for improved modeling of variable-amplitude fatigue, particularly at near-threshold crack growth rates. This study appears to suggest that conventional approaches based on the Wheeler and Willenborg residual stress models can provide reasonable estimates only by coincidence. They model the wrong parameter at lower fatigue crack growth rates and may simply not be valid at other growth rates.