Accurate and data-lite EDFA model for dynamic pump configurations.

Area of Science:

• Optical Engineering
• Telecommunications Systems
• Fiber Optics

Background:

• Erbium-doped fiber amplifiers (EDFAs) are crucial for high-capacity long-haul transmission.
• Existing analytical EDFA spectral gain models suffer accuracy degradation due to pump power variations, impacting output power and OSNR estimations.
• Dynamic pump control in long-haul systems necessitates accurate and data-efficient EDFA models.

Purpose of the Study:

• To propose an accurate and data-efficient EDFA spectral gain model for dynamic pump scenarios.
• To enhance the predictive capability of EDFA models by characterizing their operational state based on physical mechanisms.
• To improve autonomous network operation through enhanced dynamic pump control adaptability.

Main Methods:

• Singular value decomposition (SVD) was employed to analyze cross-sections and erbium population inversion, reducing data requirements.
• The model leverages underlying physical mechanisms to characterize the EDFA's operational state.
• The model incorporates the spectral hole burning (SHB) effect to account for inhomogeneous gain.

Main Results:

• The proposed model achieved a root mean square error (RMSE) of 0.11 dB with only 10 flat input measurements, outperforming the affine-law-based model.
• Incorporating the SHB effect further improved accuracy, achieving an RMSE of 0.07 dB.
• High prediction accuracy was maintained across a dynamic pump current range (360 mA to 880 mA).

Conclusions:

• The developed EDFA spectral gain model offers superior prediction accuracy and data efficiency in dynamic pump scenarios.
• The model's ability to account for spectral hole burning enhances its performance in long-haul transmission systems.
• This advancement facilitates more autonomous network operation by improving dynamic pump control adaptability.