Research on model reduction technique considering the actual stress state of structures
To mitigate the challenges associated with inaccuracy and lack of interactivity in existing model order reduc-tion techniques,a model reduction and correction technique is proposed that incorporates the actual stress state of the struc-ture.Utilizing the proper orthogonal decomposition method,a reduction model for a variable cross section beam structure is developed,which comprises a mapping of a load-modal coefficients and basis vectors.The modal coefficients undergo map-ping and updating procedures,which are based on the actual stress state of the structure as determined by response data ob-tained from testing points.An iterative process is employed to select the most suitable mapping technique and achieve cor-rection for model reduction.The corrected reduced model is validated through static and dynamic testing.Compared to con-ventional model reduction methods,the model reduction technique that considers the actual stress state achieved an accuracy improvement of 52.13％in bending static tests and 86.98％in wave loading dynamic tests,while increasing computational efficiency by 84.81％.This presents a novel approach to address the interaction of physical and virtual spaces in digital twin-ning,further meeting the requirements of real-time,accurate,and interactive digital twin structures for ships,thereby offer-ing technical assistance for the development of authentic ship structure digital twins.
digital twinreduced order modelproper orthogonal decomposition
NETL
NSTL
万方数据
