Loading...
Publication
Motor overheating monitoring using thermal images and artificial neural networks
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.66 MB | Adobe PDF |
Authors
Abstract(s)
Motor overheating is a serious problem, which can be caused by overload, poormaintenance, age degradation, among other reasons. If necessary precautionarymeasures are not taken, it can result in premature damage of the motor or accidents.Monitoring the equipment condition, in this case the temperature, is a good way toprevent failures and enlarge its availability. The present work describes the outline of asystem to monitor running motors using a thermal camera. The camera takes images ofthe most sensitive parts of the motor in real time. The images are then processed toextract the region where the temperature is higher. The region of the image at highertemperature is analysed. The area (as number of pixels) and shape of the region are thenclassified using an artificial neural network. The network is trained to recognize shapeswhere the safety of the motor could be compromised. All the algorithms to processimages and artificial neural network development were written using Python tools. Themodel achieved an accuracy of 81.3% and 94.3% in the training and test sets, respectively.
Description
Keywords
Motor overheating Thermal image processing Artificial neural networks