Lifecycle Management of Optical Networks with Dynamic-Updating Digital Twin: A Hybrid Data-Driven and Physics-Informed Approach

TL;DR

The paper tackles the need for accurate, dynamically updated lifecycle models of next-generation optical networks operating in low-margin regimes. It introduces a dynamic-updating digital twin built on a hybrid data-driven and physics-informed neural operator framework (DeepONet) to model multi-channel fiber-channel power evolution and QoT, with a physics-informed regularization of the SRS-ODE and the ability to update key physical parameters $r$, $ ext{δ}_{in}$, $ ext{δ}_{out}$, and $g_n$. The approach achieves fast, physically consistent forward predictions and enables inverse updating of critical parameters, demonstrated through large-scale simulations on COST 239 and a field-deployed C+L-band link, with notable improvements in accuracy (up to 1.4 dB QoT) and substantial speedups (up to 100×). The work lays groundwork for lifecycle autonomous management of optical networks and points to integration with SDN/NOS for practical deployment, while highlighting open challenges such as longitudinal anomaly detection and transceiver/EDFA-noise updates. Overall, the dynamic-updating DT provides a promising pathway to continuously reflect evolving network states and optimize operation across deployment and maintenance phases.

Abstract

Digital twin (DT) techniques have been proposed for the autonomous operation and lifecycle management of next-generation optical networks. To fully utilize potential capacity and accommodate dynamic services, the DT must dynamically update in sync with deployed optical networks throughout their lifecycle, ensuring low-margin operation. This paper proposes a dynamic-updating DT for the lifecycle management of optical networks, employing a hybrid approach that integrates data-driven and physics-informed techniques for fiber channel modeling. This integration ensures both rapid calculation speed and high physics consistency in optical performance prediction while enabling the dynamic updating of critical physical parameters for DT. The lifecycle management of optical networks, covering accurate performance prediction at the network deployment and dynamic updating during network operation, is demonstrated through simulation in a large-scale network. Up to 100 times speedup in prediction is observed compared to classical numerical methods. In addition, the fiber Raman gain strength, amplifier frequency-dependent gain profile, and connector loss between fiber and amplifier on C and L bands can be simultaneously updated. Moreover, the dynamic-updating DT is verified on a field-trial C+L-band transmission link, achieving a maximum accuracy improvement of 1.4 dB for performance estimation post-device replacement. Overall, the dynamic-updating DT holds promise for driving the next-generation optical networks towards lifecycle autonomous management.

Figures (11)

• Figure 1: Lifecycle management of optical networks using dynamic-updating DT with different types of monitoring data collected from optical networks.
• Figure 2: Schematic of (a) hybrid data-driven and physics-informed DeepONet for fiber channel modeling, (b) trained DT for forward prediction in multi-span link, and (c) dynamic updating of the trained DT with changing networks (shifts on physical parameters) during lifecycle. Parameters to be updated are circled by yellow square.
• Figure 3: Training workflow of using hybrid data-driven and physics-informed neural operator network in (a) forward channel power prediction, and (b) dynamic parameter updating.
• Figure 4: Simulation networks with COST 239 topology. Coupler is used for coupling or splitting C and L band. Link between node 8 and 11 is shown as an example.
• Figure 5: Training loss of PEO in the forward power prediction.