The Establishment of Repeated Loading and Unloading Constitutive Model Based on CNN Network

With the rapid development of industry, deep learning tools represented by neural networks have been widely applied in various fields of life due to their powerful processing capabilities for complex multidimensional data. The stress-strain changes caused by repeated loading and unloading during the sheet metal forming process are a typical complex multidimensional data problem. In order to better study the stress-strain changes of materials under complex load conditions, this study will use CNN neural network as the main modeling tool to theoretically model unidirectional repeated loading and unloading and bidirectional repeated loading and unloading. By comparing experimental and network prediction results, the feasibility of the proposed method is demonstrated.