From: On the use of neural networks to evaluate performances of shell models for composites
 | \({\mathbf {DOF}}\) | \({\mathbf {u}}_{\mathbf {x1}}\) | \({\mathbf {u}}_{\mathbf {y1}}\) | \({\mathbf {u}}_{\mathbf {z1}}\) | \({\mathbf {u}}_{\mathbf {x2}}\) | \({\mathbf {u}}_{\mathbf {y2}}\) | \({\mathbf {u}}_{\mathbf {z2}}\) | \({\mathbf {u}}_{\mathbf {x3}}\) | \({\mathbf {u}}_{\mathbf {y3}}\) | \({\mathbf {u}}_{\mathbf {z3}}\) | \({\mathbf {u}}_{\mathbf {x4}}\) | \({\mathbf {u}}_{\mathbf {y4}}\) | \({\mathbf {u}}_{\mathbf {z4}}\) | \({\mathbf {u}}_{\mathbf {x5}}\) | \({\mathbf {u}}_{\mathbf {y5}}\) | \({\mathbf {u}}_{\mathbf {z5}}\) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
N = 4 | 15 | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) |
TSDT | 9 | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\vartriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\vartriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) |
PTD | 7 | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\blacktriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) |
FSDT | 5 | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\blacktriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) | \(\vartriangle \) |