Electric Motor Winding Inspection
Prevent inefficient motors by detecting potential winding errors with deep learning solutions
Graphical programming environment for deep learning-based industrial image analysis
In an electric motor, insulated copper wire is wound around a core to create or receive electromagnetic energy, transferring that energy by induction to another coil. Such coils are also found in converters. These coils are rapidly wound by a machine.
The windings in electric vehicle (EV) motors are extremely dense. Any inaccuracies in how they were wound can have a negative effect on the motor’s efficiency. Given the vast number of windings crammed into a narrow space, even small winding errors can be significant, but hard to identify. The winding error may be subtle and can occur anywhere among the many visible wires.
There is no efficient way to code a rule-based machine vision system to cover all the winding error possibilities anywhere on the coil. Human inspection is also not suited for identifying such subtle errors in a complex image.
Cognex Deep Learning using a color camera accurately verifies that the winding process has been accomplished without error. The defect detection tool learns from a set of training images consisting of error-free windings and labeled images featuring a wide range of overlaps, mispositionings, crossings, and other potential errors in various locations.