In this time of energy crisis, windmill is a breath of air. However, its massive structure poses manufacturing, logistic and installation challenges. We cannot afford the cost of this scale. That’s why it is important to anticipate failure at the simulation stage.
Turbine engineer is given a right design using fatigue testing on breathing deformation.
Breathing deformation is characterized by
- Problems with the bond between the internal structural webs and the blade skin.
- Inter-laminar delamination in the composite material.
- Or the slow accumulation of cracks or fractures.
Fatigue testing unearths the cracks and fractures. To do so, one component of a prototype is subjected to repeated, cyclic loading that simulates decades of real-world operation.
For the turbine, a blade is bolted to a test bench at its root, and an excitation load is applied to force the blade into large oscillations that simulate long periods of wind and rotational loads. The goal is to verify that the blade’s structural design is sound and that the manufacturing process produced something that actually matches the design specifications. This tests the fatigue and patience as well.
To measure turbine blade’s interior deals with a dark, narrow and long cavity. Traditional and technological ways were explored:
- a person does not fit to move inside, no natural light, and no convenient mounting surfaces.
- Traditional contact sensors like accelerometers, strain gauges, and displacement transducers
- add weight, potentially alters the local stiffness of the material
- specific point information based on where it’s installed. An entire cross-section requires many sensors, but compromise the measurement.
- non-contact options like laser-based sensors addressed weight but limited to single-point measurements instead of a system that can track many points simultaneously, across the whole cross-section, without touching the blade at all.
Computer Vision to compute for the deformation inside the blade was studied. Their methodology involved feature detection and template matching system.
Feature Detection resolves geometry and computer vision problems thru feature extraction and calibration.
Photogrammetry triangulates three-dimensional position from camera’s flat and 2D output to reconstruct a 3D world. The distance between the two cameras is called the baseline. A longer baseline gives depth, a shorter baseline limits depth resolution necessary in constrained space. In the interior of these blades with constrained space creates some real engineering tradeoffs. But first solve geometric distortion: straight lines curve near the edge of the frame that can corrupt your data. Calibration characterizes
- how camera processes the light and space
- geometric relationship between cameras
Using the checkerboard for calibration’s standard tool by moving around the measurement space from many different poses.
Next, computer vision’s mechanism solves the problem of reliably tracking the target across thousands of image frames.
Lack of proper lighting replaced using reflective target with non-reflected target: flat, high-contrast markers that hold their appearance under ambient light. It works with poorer lighting but needs sophisticated tracking algorithm.
This is where this paper’s feature-based template matching is studied
More of this the next article…
Leave a Reply