土木分野におけるAIの活用(6)

[CS15-45] Study on linerization analysis method using by Neural Network

〇Makito Nago¹, Takahiro Kondo¹ (1.Taisei)

Keywords:Neural Network, AI, Deep learning, linearinzation analysis, transfer function

The authors compared the results of linearization analysis of the relationship between crane load, boom angle and wire tension using the conventional linear analysis method and three-layer Neural Network. As a result, the linear analysis value by Neural Network learning approximated the model equation by the least squares method. The overall smooth interpolation curve can be obtained by setting an appropriate learning error. Linearization by Neural Network learning has practical significance as a countermeasure in cases where model equations cannot be derived.