Table 2 Prediction results of topcoat thickness measurement (\(\:{T}_{topcoat}\)) and refractive index (\(\:n\)) for real-world TBC samples.
Training data | Model | Sample | Thickness, \(\:{\varvec{T}}_{\varvec{t}\varvec{o}\varvec{p}\varvec{c}\varvec{o}\varvec{a}\varvec{t}}\) (µm) | Refractive index, \(\:\varvec{n}\) | ||||
Mean | SD | MAPE | Mean | SD | MAPE | |||
Simulation | 1D-CNN | 1 | 30.03 | 5.01 | 25.13 | 4.72 | 0.15 | 3.06 |
2 | 40.43 | 4.97 | 26.35 | 4.64 | 0.11 | 3.31 | ||
3 | 70.09 | 3.19 | 7.84 | 4.78 | 0.13 | 2.59 | ||
4 | 126.56 | 2.34 | 5.47 | 4.82 | 0.05 | 1.00 | ||
LSTM + Attention | 1 | 27.96 | 3.01 | 16.5 | 4.84 | 0.03 | 1.12 | |
2 | 37.11 | 2.97 | 15.96 | 4.87 | 0.02 | 1.47 | ||
3 | 68.24 | 2.19 | 4.98 | 4.79 | 0.05 | 0.90 | ||
4 | 124.06 | 1.86 | 3.38 | 4.72 | 0.06 | 1.67 | ||
Simulation + Experiment | 1D-CNN | 1 | 29.06 | 3.99 | 21.08 | 4.69 | 0.16 | 3.28 |
2 | 39.49 | 3.78 | 23.40 | 4.62 | 0.10 | 3.67 | ||
3 | 68.05 | 2.09 | 4.69 | 4.63 | 0.10 | 3.49 | ||
4 | 122.98 | 1.45 | 2.48 | 4.83 | 0.09 | 1.78 | ||
LSTM + Attention | 1 | 26.76 | 2.67 | 11.53 | 4.8 | 0.01 | 0.27 | |
2 | 35.89 | 2.56 | 12.16 | 4.82 | 0.03 | 0.70 | ||
3 | 67.93 | 2.02 | 4.50 | 4.78 | 0.02 | 0.51 | ||
4 | 122.04 | 1.07 | 1.75 | 4.77 | 0.05 | 1.05 | ||
