PSI - Issue 57

Kalle Lipiäinen et al. / Procedia Structural Integrity 57 (2024) 32–41 Lipiäinen et al. / Structural Integrity Procedia 00 (2019) 000 – 000

35

4

using the local stress ratio:

k 1 R   −

,

(8)

k,ref  = 

local

where Δ σ k,ref is the SWT mean stress-corrected stress applied in the S - N - curve. Δ σ k,ref is computed for each test result based on the applied input parameters (applied load range and stress ratio, defect size, residual stress, and material model at the notch). In a workflow of the 4R method combining research and design of components, the fatigue tests and fracture surfaces are analysed in a first step (Fig. 2a). The initial data is the processed numerically with FEA including quality parameter , a nd reference data is produced using mean stress correction (Fig. 2b). When the proposed method is applied to assess a fatigue life of engineering component, expected quality is modelled in most critical locations of component geometry (Fig. 2c). High-quality products, typically benefit from verification testing which is considered in Fig. 2d with following a and b steps to compare performance against 4R reference stress data and further verify design procedure validity.

a

b

Material testing and characterization

Experimental result analysis – creating and updating database

Unnotched and notched specimens in as built and post-treated conditions

Database for fatigue capacity with universal mean stress – level, surface condition and build quality

Input parameter extraction from fractography, and metallurgical analysis

Local stress range with extracted equivalent crack length

Mean stress correction

σ 1

2000

2D FEA

Cyclic loading

0 2 4 6 8 10

Surface roughness

Stress

σ norm.

R local = σ 2 / σ 1

(MPa)

Internal Crack length

Surface

0 200 400

Equivalent crack length Hardness 350 HV

Load range

x ( μ m)

Crack based on fractography

Stress averaging distance

200

10 000 100 000 1 000 000 Reference stress [MPa] Fatigue life

Material performance focused Scientific publications

Strain

σ 2

Loop to failure analysis

Product design procedure

Verification and prototyping

c

d

FEA extracted stress range including quality effects

Validating method with prototype fatigue testing and updating database

2000

Shape from CAD-software (topology optimization) Including quality parameter as numerical value from database

Equivalent crack

Load

Mean stress corrected value from FEA

Database values

Component fatigue life estimation against database values

FE-model including simplified: - Build orientation - Build quality - Surface finish

200 10 000 100 000 1 000 000 Reference stress [MPa] Fatigue life

Fixed

Fig 2. Schematic 4R fatigue strength database creation and fatigue strength assessment workflow shing (a) fatigue test and fractography, (b) transferring data to 4R reference stresses via local stresses and mean stress correction, (c) component and product design phase and (d) experimental verification test and small-scale test comparison. 3. Results L-PBF specimens tested in (Afkhami,Dabiri, Lipiäinen, et al., 2021; Afkhami,Dabiri, Piili, et al., 2021; Afkhami et al., 2022) were re-evaluated with TCD-based 4R method (see Table 1 for 4R material parameters and Table 2 for more detailed specimen identification). The data includes both notched and un-notched specimens. MS1 (18Ni300) is a high strength tool steel (17-19 wt-% Ni, 8.5-9.5 wt-% Co and 4.5-5.2 wt-% Mo as main alloying elements) that was tested in as-built condition yield and tensile strength around 1000 MPa. CX-material (13Cr10Ni1.7Mo2Al0.4Mn0.4Si) is a high strength stainless tool steel that was tested in as-built and heat-treated (HT) conditions. Heat treatment for CX material was conducted heat-treated specimens were subjected to annealing at 850 ° C for 30 min and subsequent aging at 525 ° C for 120 min followed by cooling in air. For reference low-strength 316L was evaluated in as-built condition. CX-material provide interesting results due to heat-treatment and increase in material strength. Original data is illustrated in Figure 3 utilizing nominal stress method.