Session: 03-03-01: Annual Congress-Wide Symposium on Additive Manufacturing I

Paper Number: 165355

Scaling of Fatigue Curves Derived From Surface Roughness 

For additively manufactured (AM) parts, different printing conditions, part orientations, and post-processing steps can result in substantially different surface roughness. The as-built surface roughness inherent to AM processes can be difficult to remove when it is internal, such as on the internal surface of a lattice heat exchanger. Other times it is kept as-built due to post-processing cost and time constraints. Several previous studies have shown significant reduction of fatigue life for parts with as-built surface roughness. One of the challenges in adopting AM hardware in the space industry is to fly parts with as-built surface roughness and predict their fatigue life.

Fatigue test data collected using coupons with one surface roughness condition can be difficult to apply to real parts with different geometry and roughness condition. A method to scale fatigue S-N curves based on surface roughness characteristics is proposed in this paper. The ratio of fatigue life between two surface roughness conditions is derived using fracture mechanics equations based on the largest pit area on the surface. In the full paper, a numerical example of scaling S-N curves from one surface condition to another is provided. This method will be compared against S-N curves from fatigue tests of coupons with varying levels of surface roughness.

The state of the art in studying fatigue life with surface roughness is reviewed in the full paper and summarized here: X-ray computed tomography (CT) data are used to characterize as-built surface roughness and identify the surface pit with highest stress intensity factor, i.e., killer notch, where fatigue failure initiates. The fatigue endurance limit of AM fatigue specimens with as-built surface roughness is evaluated using fracture mechanics-based analysis of maximum pit depth. This approach finds the maximum stress a part can be cycled at indefinitely without causing fatigue failure.

To design parts with surface roughness for limited-life applications where the cyclic stress amplitude exceeds the fatigue endurance limit, a common approach in the space industry is:

1.                  Print fatigue test coupons with surface roughness at various build orientations

2.                  Perform fatigue tests to obtain the worst S-N curve among various orientations

3.                  Apply a knock down factor to the worst S-N curve from test to account for dis-similitude between the coupon roughness and real part roughness, and use this as the lower bound S-N curve

The challenge of this approach is how to choose proper knock down factor. This work attempts to answer the following question: If fatigue testing is performed on specimens with one surface roughness condition, how can the resulting fatigue S-N curve be applied to specimens with different surface roughness?

In the full paper: A fracture-based fatigue life assessment is used to derive scaling factors for S-N curves with respect to roughness measurements. The derivation of equations to scale SN curve is summarized. A numerical example compares scaled S-N curve with curve obtained from fracture-based fatigue assessment. It was found that fatigue curve scaling by number of cycles and stress level are both promising approaches to determining fatigue knockdowns based on surface roughness pit depth. Overall, scaling by stress level resulted in a more accurate result over a larger portion of the S-N curve.

For the final paper, the authors intend to use rotating bending fatigue data for Inconel 718 available in NIST dataset to evaluate this method. The S-N curve shifts predicted by the scaling method in this paper will be compared to the shifts observed in the real fatigue data.

Presenting Author: Leland Shimizu The Aerospace Corporation

Presenting Author Biography: Leland Shimizu is an Engineering Specialist at The Aerospace Corporation. He specializes in space structures and systems, fatigue and fracture, and uncertainty quantification. He earned a MS and Bachelor degree in Aerospace Engineering from UCLA.

Authors:

Leland Shimizu The Aerospace Corporation
Xueyong Qu The Aerospace Corporation