Machine Learning Predicting Mode Properties for Multi-Layer Active Fibers

Yi An; Min Jiang; Xiao Chen; Jun Li; Rongtao Su; Liangjin Huang; Zhiyong Pan; Jinyong Leng; Zongfu Jiang; Pu Zhou

doi:10.3788/CJL230476

Chinese Journal of Lasers, Volume. 50, Issue 11, 1101013(2023)

Machine Learning Predicting Mode Properties for Multi-Layer Active Fibers

Yi An¹, Min Jiang^1,2, Xiao Chen¹, Jun Li¹, Rongtao Su^1,3,4, Liangjin Huang^1,3,4、*, Zhiyong Pan^1,3,4, Jinyong Leng^1,3,4, Zongfu Jiang^1,3,4, and Pu Zhou^1、**

Author Affiliations

¹College of Advanced Interdisciplinary Studies, National University of Defense Technology, Changsha 410073, Hunan, China

²Test Center, National University of Defense Technology, Xi an 710106, Shaanxi, China

³Nanhu Laser Laboratory, National University of Defense Technology, Changsha 410073, Hunan, China

⁴Hunan Provincial Key Laboratory of High Energy Laser Technology, National University of Defense Technology, Changsha 410073, Hunan, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

$Schematics of the cross-section structure and refractive index distribution for conventional and several typical multi-layer active fibers. (a) Conventional fiber; (b) confine doped fiber; (c) M-type fiber; (d) pedestal fiber; (e) single trench fiber$

Fig. 1. Schematics of the cross-section structure and refractive index distribution for conventional and several typical multi-layer active fibers. (a) Conventional fiber; (b) confine doped fiber; (c) M-type fiber; (d) pedestal fiber; (e) single trench fiber

Download full size

Fig. 2. Schematic of machine learning approach to predict mode properties of multi-layer active fibers

Download full size

Fig. 3. Variation of MSE during training process for NN and NN′

Download full size

Fig. 4. Predicted n_eff results of fundamental mode (FM) and high-order-mode (HOM) for the testing samples through NN and NN′. (a) FM; (b) HOM

Download full size

Fig. 5. Intensity distributions of FM and HOM for the three testing samples A, B, and C

Download full size

Fig. 6. Comparison between NN predicted values and ground truths for the three testing samples A, B, and C

Download full size

Fig. 7. Comparison between NN predicted values and ground truths for all the testing samples

Download full size

Fig. 8. Comparison between NN predicted mode properties and ground truths for the testing sample under different working wavelengths

Download full size

Table 1. Varying range and step of structural parameters for multi-layer active fibers
View table
Table 1. Varying range and step of structural parameters for multi-layer active fibers
Fiber parameter Range Step
∆n 0.0008-0.0016 0.00002
t₁ /μm 0.25-10 0.25
t₂ /μm 0.25-10 0.25
λ /μm 0.8-1.2 0.01

Table 2. Averaged prediction error for the testing samples
View table
Table 2. Averaged prediction error for the testing samples
Mode P_E（n_eff）/% P_E（A_eff）/% P_E（Г）/%
FM 0.05 0.27 0.30
HOM 0.08 0.56 0.47

Tools

Get Citation

Copy Citation Text

Yi An, Min Jiang, Xiao Chen, Jun Li, Rongtao Su, Liangjin Huang, Zhiyong Pan, Jinyong Leng, Zongfu Jiang, Pu Zhou. Machine Learning Predicting Mode Properties for Multi-Layer Active Fibers[J]. Chinese Journal of Lasers, 2023, 50(11): 1101013

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites