Process Monitoring for Consistent AM Part Quality

The Challenge

Qualifying complex-shaped, rough-finish metal AM parts requires volumetric non-destructive testing methods. X-ray computed tomography (XCT) is currently favored but demands technical expertise, high equipment costs, and has limitations on part size and density. As metal AM moves from prototyping to production, manufacturers need faster, more accessible methods to verify that process parameters remain within specification, not just to detect defects, but to ensure consistency across production runs.

The Solution

This research demonstrated that resonant ultrasound spectroscopy can segregate metallic parts manufactured with different process parameters, not just sort defective from flawless parts. Eleven sets of parts were manufactured using metal laser powder bed fusion with varied wall thicknesses, laser powers, scanning speeds, and scanning strategies. Parts were tested using resonant frequency analysis and evaluated with Z-score statistical methods.

The resonance responses clearly distinguished parts made with different process parameters, enabling classification by manufacturing settings. This comparative approach, analyzing frequency spectra against reference samples from the same part family, provides a practical method for statistical process control in AM production.

Key takeaway: Z-score analysis of resonant frequency spectra classified all eleven manufacturing parameter variations, proving the method works for process control, not just defect detection.

Results

The research proved that resonant frequency methods can classify AM parts according to their process parameters, providing manufacturers a convenient tool for two critical functions: identifying defective parts and configuring machine parameters to achieve desired material properties. This enables practical quality control for production AM environments where CT scanning every part is impractical.