Preventing Trapped Air and Lost Strength: Vision AI for Wind Blade Infusion

One of the most consequential calls in wind blade manufacturing is still made by eye. During infusion, someone stands over the vacuum bag, watches resin spread through the shell, and decides when to open the next feed line. Open it too early, and the resin can seal off a patch of dry fiber or trap air beneath the surface. While the resin is still liquid, crews can rework a dry patch through the bag. Once it cures, the flaw is locked into the laminate.

The tools to make that call more precisely are scarce. One study of vacuum infusion found monitoring and active control almost absent in industry. AI-powered visibility from WorkWatch is helping to close that gap.

A Few Percent of Trapped Air Costs Real Strength

When Sandia National Laboratories, a Department of Energy research lab, set out to study what makes wind blades unreliable, porosity — the voids left behind by trapped air — was one of two defects its Blade Reliability Collaborative singled out, alongside fiber waves. Testing for that program set 2% porosity as the acceptable ceiling for blades. Even at that threshold, test samples kept just 84% of their compressive strength, because voids soften the resin that holds fibers in line under load.

That strength loss adds up in a structure built to run for decades. Sandia has long noted that early blade failures typically start at manufacturing defects. Some infusion defects can still be caught at post-cure inspection, before the blade halves are bonded. By then, though, fixing them means cutting into cured laminate. During infusion, the fix is still a matter of moving resin.

Infusion Timing Decides Where Defects Form

Resin doesn't go in all at once. It enters through a main line down the middle of the shell, and crews open extra feed lines one at a time as the resin spreads past them. A Department of Energy–funded study from Wichita State University shows why that timing is delicate. Resin doesn't spread evenly through the thickness of the shell: it moves faster across the top layers than through the layers underneath. When the researchers simulated opening an extra line too soon, the resin on top surged ahead, cut off the slower resin below, and left a pocket of dry fiber that resin could no longer reach.

That's the trouble with watching from above. A crew can see how far resin has spread across the top of the shell, but not whether the layers underneath have caught up. Research on vacuum infusion notes that this gap is often difficult or impossible to spot by eye.

The Check Has to Happen While the Resin Is Still Moving

Verifying infusion means watching the whole shell — not a handful of points on it — for as long as the fill runs. The ideal check tracks how far the resin has spread and signals when it has cleared each feed line, so the crew can open the next one on time. It flags any region that's falling behind, so a line doesn't open early and cut it off. And it records how fast each fill moved, shell after shell, so process engineers can see which molds and layups give them trouble.

That's the kind of visibility WorkWatch is built to provide. It alerts crews on the floor through a line display while there's still time to act, and gives quality teams off the floor a dashboard view of every stage. The crew still opens the line. They just know when.

Using cameras combined with AI analysis, WorkWatch also creates an audit trail of the blade production process. When a question comes up in the field, quality teams can pull the record instead of reconstructing it from memory. For contract manufacturers, it's documentation that every stage ran to spec, which limits liability down the line.

Infusion is the last point where a defect in the shell is still soft. AI-powered visibility gives manufacturers a way to see it there, on every blade, with a record that outlasts the shift.

To see what Vision AI could mean for your operation, schedule a demo of WorkWatch today.

‍