A team from the University of Waterloo's Multi-Scale Additive Manufacturing lab used Dyndrite LPBF Pro to develop laser paths whose settings can be changed at vector level to reduce pore formation in laser powder bed fusion. The project by Jigar Patel, Sagar Patel, Rene Lam and Mihaela Vlasea won the $10,000 Innovation Award in Dyndrite's inaugural student competition at ASTM ICAM 2026 in Orlando.

Dyndrite's 1 October 2026 release says the optimised strategy reduced the number of detected pores by 75 per cent and maximum pore size by about 40 per cent. The figures are notable, but the public release does not provide the complete experimental matrix, alloy, replicate count, measurement uncertainty or raw tomography data. They should therefore be treated as project results reported by the research group and software vendor.

Why does porosity form in LPBF?

In laser powder bed fusion the beam creates a melt pool whose shape depends on power, travel speed, spot size, material, powder bed and the thermal state left by previous tracks. Too little energy or poor overlap can leave lack-of-fusion defects. Excess local energy can create an unstable deep keyhole and trap gas. Spatter, contamination and gas flow add further failure routes.

A conventional parameter set often applies the same power, speed, hatch spacing and contour rule across a layer or region. Real geometry mixes corners, short vectors, thin walls, bulk infill and boundaries. The same setting may be stable on one vector but overheated or incompletely fused on another.

What is vector-level control?

The machine divides each layer cross-section into laser vectors. When software can assign power, speed, order, delay or strategy to individual vectors or groups, the process can respond to local geometry. The Waterloo project used geometry segmentation and custom toolpath programming so critical regions could be exposed differently from general infill.

This is not equivalent to optimising one energy-density number. Two settings with equal nominal volumetric energy density can create different melt pools when power and speed are combined differently. Vector order and time between neighbouring tracks also change heat accumulation.

What do 75 and 40 per cent prove?

The 75 per cent result refers to detected pore count and the roughly 40 per cent result to the largest observed pore in the reported comparison. Neither automatically means 75 per cent longer fatigue life, 40 per cent higher strength or a fully dense part. Mechanical effect depends on pore shape, position, orientation and loading.

A percentage also needs a baseline. A reduction from ten pores to two or three and from ten thousand to two thousand five hundred share the same relative scale while describing different quality states. Maximum pore linear dimension, area and volume are different quantities. The release does not supply every definition, so an absolute acceptance threshold cannot be derived from it.

What is strong about the project?

Its strength is process-programming logic. Instead of seeking one recipe for an entire part, the workflow segments geometry into regions with different physical risks and laser strategies. That reflects a broader evolution in manufacturing software: the build processor becomes a rule-based process description rather than merely a file generator.

The award was presented at ASTM ICAM, but it is not an ASTM standard or independent certification. Judges assessed innovation and project merit. Dyndrite is both the vendor of the software and organiser of the competition, making its release an interested primary source. That does not invalidate the result, but it defines the evidence boundary.

What is needed for industrial generalisation?

First comes a complete experimental description: alloy and powder condition, machine, layer thickness, atmosphere, optics, specimen geometry, replicate count and reference scan strategy. Second is the pore-measurement method—metallographic section, X-ray CT or another technique—with resolution, segmentation threshold and uncertainty. Third is repetition across builds, plate positions and ideally another machine.

Industrial value also requires mechanical and productivity results. A strategy may reduce pores while increasing scan time, residual stress or roughness. The correct decision is multivariable: porosity, dimensions, surface, stress, tensile and fatigue properties, and cycle time all matter.

A practical adoption route

A manufacturer should first compare current and vector-level strategies on a familiar reference geometry. Use one powder lot, matched plate positions and multiple repeats. Add dimensional, surface and mechanical tests to CT or metallography. Evaluate distributions and uncertainty rather than only an average density.

If the improvement repeats, release the rule for one limited part family. Software version, segmentation logic and vector parameters need the same change control as machine settings. A later automatic update must not silently replace the qualified strategy.

Verified fact, developer claim and editorial conclusion

The Waterloo team, project title, use of Dyndrite LPBF Pro and $10,000 award at ASTM ICAM 2026 are verified. The 75 per cent pore-count and roughly 40 per cent maximum-pore reductions are research-team results carried in Dyndrite's release; the public material does not provide a complete independent reproduction.

LAYERTIMES concludes that the project's central message is local process control, not the two headline percentages. LPBF software becomes a more valuable production tool when regions with different risks receive documented, measurable and versioned laser strategies. The specific gain still requires material-, machine- and part-specific qualification.

Sources

- Dyndrite: University of Waterloo researchers win inaugural Student Competition — 1 October 2026. - University of Waterloo Multi-Scale Additive Manufacturing Lab — institutional research context. - 3Druck.com Hungarian edition — topic discovery and Hungarian technical context, 2 October 2026.