방적

The Use of Neural Nets to Simulate the Spinning Process.

  • 출판일1999.01
  • 저자
  • 서지사항
  • 등록일 2016.11.02
  • 조회수 202
Researchers developed methods to calculate yarn properties based on fiber properties and spinning machine settings. A neural network simulates the spinning process to predict yarn properties for rotor and ring spinning machines at an accuracy of 95 percent. Researchers employed coupled self training systems in which the output of the first neural network was fed into several other neural networks whose output was fed into other neural networks. The method thereby reduces the number of inputs via prefiltering and intercorrelates the output by means of postfiltering. The number of neural networks increases if the number of neurons per network decreases and if the number of hidden layers per network remains low. Researchers measured the yarn properties of 1,382 samples to verify the accuracy of the system's specification behavior and generalization behavior. 7 refs.
이전글 다음글 목록