Validating Ply Tolerances with AI: A Smarter Approach to Wind Turbine Blade Quality Control

Why Ply Tolerance Matters

In composite wind turbine blade manufacturing, it’s not just whether a ply is placed—it’s how precisely it’s placed. Each ply in a blade's laminate stack has a designated location and orientation, defined with tolerances measured in millimeters. These tolerances aren’t arbitrary; they’re engineered to ensure structural performance, fatigue resistance, and longevity in some of the most challenging operating environments in renewable energy.

Wind turbine blades must withstand millions of load cycles, including flapwise bending, torsional stress, and shear forces, especially in large offshore models where blades exceed 100 meters in length. Ply misplacement—even if only by a few centimeters—can shift the distribution of stresses through the laminate. Over time, that deviation can:

• Initiate delamination at ply interfaces
• Create resin-rich pockets or fiber wrinkles that reduce stiffness and fatigue life
• Alter aerodynamic performance by changing the geometry of the blade shell
• Compromise the bond line or spar cap interface, which are structurally critical regions

A well-placed ply contributes to a strong, predictable structure. A misaligned one introduces variability that’s difficult to catch after cure—and expensive to correct in the field.

How Manufacturers Manage Ply Tolerance Today

To meet these tight tolerances, manufacturers rely heavily on a combination of skilled labor and tooling aids. Most commonly:

• Laser projection systems are mounted above the mold to project outlines of each ply onto the surface. These outlines help guide operators in positioning plies according to the design.
• Manual review is performed by technicians or quality personnel who visually inspect placement and sometimes measure edge offsets or orientation using rulers, gauges, or angle templates.
• Ply books or digital work instructions are used in tandem to help workers confirm they’re placing the correct ply, in the correct location, in the correct order.

These methods, while effective to a point, come with trade-offs:

• They’re labor-intensive: Operators must interpret visual instructions and confirm placement manually.
• They depend on training and attention to detail: Even experienced technicians can make small mistakes under pressure.
• They offer limited traceability: If an issue arises later in the process, it’s difficult to verify exactly how a ply was placed without interrupting production.

Why AFP Isn’t (Yet) the Solution

In aerospace and automotive composite manufacturing, automated fiber placement systems have become the gold standard for high-precision layup. AFP machines use robotic heads to place narrow tows of fiber with sub-millimeter accuracy, achieving exceptional control over fiber orientation and ply thickness.

The wind industry has explored AFP for blade layup, but progress has been slow, primarily due to a mismatch between the technology’s strengths and the realities of blade production:

• AFP is slow for large areas: The deposition rate of traditional AFP systems isn’t sufficient for wind blades, which require thousands of square meters of layup.
• Blades require large-scale robots or gantries: The size of the molds demands highly customized, expensive infrastructure.
• Cost and ROI concerns: In many factories, the capital investment in AFP equipment doesn’t currently justify the production gains, especially when labor costs remain competitive.

As a result, while AFP research continues—particularly for hybrid or modular blade sections—it’s not yet a practical solution for mainstream ply placement in wind turbine blade manufacturing.

AI-Powered Ply Tolerance Validation

Computer vision, however, offers a promising middle ground. Rather than replacing human operators or reengineering the layup process entirely, AI-based vision systems can enhance current workflows by passively validating ply placement accuracy.

Here’s how it works:

• Cameras mounted above the mold capture images during the layup process.
• AI models analyze each ply placement in real time, comparing it to expected outlines.
• The system can detect whether the ply:
  • Was placed within the allowed tolerance
  • Was the correct ply altogether (by reading printed labels)

This process requires no action from the operator, meaning there’s no interruption to production. The system passively validates that the ply was placed and placed correctly.

Benefits of AI-Based Validation

• Reduces human error: Provides a second set of eyes without adding inspection steps.
• Improves traceability: Automatically captures and stores images of each layer for audit or root-cause analysis.
• Supports high-throughput manufacturing: Scales across multiple stations or shifts without needing additional inspectors.
• Adds dual-layer verification: Presence and tolerance validation can work in tandem.

Making Precision Practical at Scale

Precision in ply placement is non-negotiable in wind turbine blade manufacturing—but maintaining that precision at industrial scale is a constant challenge. Existing tools like laser projection and manual inspection do their job but rely heavily on labor, experience, and time.

Computer vision introduces a more scalable way to validate ply placement without disrupting production. By passively confirming both presence and tolerance using AI, manufacturers gain a new level of confidence and traceability—supporting structural performance and reducing the risk of costly rework or in-field failure.

At Leverege, we’ve developed WorkWatch, a computer vision tool designed to meet this need. WorkWatch uses overhead cameras and trained models to help teams monitor each ply for accuracy and completeness. It’s a practical step forward for manufacturers who want to enhance quality assurance without overhauling their process.

As blades grow and margins tighten, precision at every layer becomes even more critical. With the right tools, that precision becomes easier to achieve.

‍