|
|
|
Fitting Methods based on Custom Neural Network for Relaxation Modulus of Viscoelastic Materials
|
|
He Yun,Li Haibin,Du Juan
|
|
| |
Table 6 Simulation results of training sample points
|
|
| t (s) | Custom neural network (kPa) | BP neural network (kPa) | Prony series (kPa) | Experimental data (kPa) | | 2 | 2954.8 | 2965.4 | 2928.9 | 2965.5 | | 6 | 2909.1 | 2897.6 | 2886.6 | 2922.5 | | 10 | 2881.1 | 2877.0 | 2855.7 | 2891.1 | | 40 | 2803.5 | 2805.5 | 2795.9 | 2798.1 | | 80 | 2745.0 | 2743.2 | 2731.2 | 2750.6 | | 200 | 2643.2 | 2643.1 | 2640.0 | 2643.1 | | Max relative error | 0.46% | 0.85% | 1.24% | 0 |
|
|
|