Characterizing Pores and Surface Roughness to Build StructureProperty
Characterizing Pores and Surface Roughness to Build Structure-Property Linkages for AM 17 -4 PH By Sven Voigt Advisors: Dr. Kalidindi, Dr. Yabansu In collaboration with Sandia National Laboratories
Background of AM • Builds parts by adding material • 17 -4 PH and H 900 HT • Metal laser powder bed fusion • Laser • Powder Bed • Variability in structure • Variability in properties
Goal Predict performance from NDE of structure How • Statistical representation of structure • Link to properties Benefit • ID parts that do not meet design specification • Predict how the properties of AM parts can be improved by modifying the structure
Structure • Tries to build flat surfaces with homogeneous structure • Variability in process and powder leads to variability in the structure • Structure is imaged using Computed Tomography (CT) • Otsu’s method is used for segmentation issues • Larger pores and merging with edge
Properties • Mechanical Properties- tensile • Obtained from stress-strain data • Stress calculated with constant cross-section • Strain obtained from crosshead displacement • Variability shown in yield strength, ultimate tensile strength, strain to failure
Structure Measurements • Attempts were made to characterize the structure using measurements • Includes: Average pore volume, number of pores, maximum pore size, total pore volume, and equivalent spherical diameter
Project Outline 1. 2. 3. 4. 5. 6. 7. Acquire Structure Data Segment Images Identify Phases Represent phases using n-point statistics Acquire Property Data Link Structure and Properties Make predictions about improving structure for performance
Completed • Acquire Structure Data • Received from Sandia as CT images • Acquire Property Data • Received as single scalar values from Sandia • Might need raw stress-strain curves • Segment Images • Might need improvement
Upcoming Soon • Identifying phases • 3 phases: material, pore, edge • Use algorithm for identifying enclosed volumes • N-point statistics • Representing the pore and edge as different values • If edge is included, the structure is non-periodic
Future • Build linkages • Connect n-point statistics to the as-calculated material properties • Predict structural improvements • Understand the impact of each of the phases
- Slides: 10