South Ural State University (SUSU) has successfully applied a new algorithm for the automatic diagnostics of bearings using a neural network for the first time in Russia. This was announced by Vladimir Sinitsyn, Deputy Head of the Research Laboratory of Technical Self-Diagnostics and Self-Control of Instruments and Systems.
The scientific and technical problem was that a hybrid model, Hybrid MLP-CNN, is used for automatic diagnostics of bearings involving neural networks. This model requires computational and time costs, as well as a large amount of data for training, otherwise it works with errors.
To solve this problem, scientists used the LPC algorithm, which requires 15 times less time to train the neural network. Previously, this algorithm was used for human speech recognition and synthesis. Now it has been adapted for bearing diagnostics.
Earlier, www1.ru reported that a platform for collaborative robots with quality control function was presented in Skolkovo.
Read more on the topic:
Microelectronics manufacturer "Element" launches industrial robotics division
Collaborative robots from Kazan are being tested in KAMAZ workshops