Identification methods in nonlinear heat conduction. I: Model reduction (Q2488658)

Although computers are more and more powerful, the reduction of CPU time remains a crucial issue, especially for 3D large scale industrial problems. The modal identification model previously developed for linear systems is extended to nonlinear systems in heat conduction. As expected, the reduced model has a much smaller size than the detailed model. It is claimed that the loss of accuracy is insignificant, but no mathematical quantification for this is provided.