Closed-Form EGN Model with Comprehensive Raman Support

Yanchao Jiang,Antonino Nespola,Stefano Straullu,Alberto Tanzi,Stefano Piciaccia,Fabrizio Forghieri,Dario Pilori,Pierluigi Poggiolini
2024-07-10
Abstract:We present a series of experiments testing the accuracy of a new closed-form multiband EGN model, carried out over a full-Raman 9-span C+L link. Transmission regimes ranged from linear to strongly non-linear with large ISRS. We found good correspondence between predicted and measured performance.
Signal Processing
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop and validate a new closed - form multi - band EGN model (CFM6) to accurately predict the nonlinear interference (NLI) in multi - band (C + L - band) fiber - optic communication systems in the presence of backward - pumped Raman amplification (BPRA). Specifically, the research aims to: 1. **Improve calculation speed and real - time management capabilities**: As the requirements for physical layer models (PLMs) continue to increase, faster calculation speeds are required to support iterative optimization and real - time management. 2. **Consider frequency - dependent parameters**: Expanding the available fiber bandwidth to multi - band (MB) requires that the model be able to handle frequency - dependent parameters. 3. **Include the effect of backward - pumped Raman amplification**: Existing models such as CFM5 are unable to handle the nonlinear interference (NLI) generated by backward - pumped Raman amplification, while CFM6 is able to model this. ### Research Background In optical networks, it has become crucial to incorporate physical - layer behavior into network design, management, and optimization processes. To this end, a variety of physical - layer models (PLMs) have been developed, among which the GN and EGN models are widely adopted. However, as requirements increase, these models need to have faster calculation speeds, be able to handle frequency - dependent parameters, and be extended to multi - band systems. To meet these challenges, researchers have developed approximate closed - form models (CFMs), such as CFM5 and CFM6. ### Characteristics of CFM6 CFM6 is an improved version based on the EGN model, and is special in that it can handle the nonlinear interference (NLI) caused by backward - pumped Raman amplification (BPRA). By introducing new formulas and parameters, this model can more accurately predict NLI in multi - band systems. ### Experimental Verification To verify the accuracy of CFM6, researchers designed a 9 - span, 540 - kilometer C + L - band experiment using backward - pumped Raman amplification. The experiment covered transmission conditions from linear to strongly nonlinear and tested performance under different conditions. The results show that the CFM6 model has good agreement with the experimental measurement results, especially under strongly nonlinear conditions, where the NLI predicted by the model is very close to the actual measured value. ### Conclusion This research shows that the CFM6 model has high accuracy in handling nonlinear interference in multi - band systems, especially in the presence of backward - pumped Raman amplification. This achievement provides strong support for future optical communication system design and optimization. ### Formula Summary 1. **Field Loss**: \[ \alpha_{n,i}(z)=\alpha_{0,n,i}+\alpha_{1,n,i}\cdot\exp(-\sigma_{n,i}\cdot z) \] where \(n\) is the channel index, \(i = 1\) represents the forward direction, and \(i = 2\) represents the backward direction. 2. **Effective Dispersion**: \[ \bar{\beta}_2^m=\beta_2+\pi\beta_3(f_n + f_m - 2f_0)+\frac{2\pi^2}{3}\beta_4[(f_n - f_0)^2+(f_n - f_0)(f_m - f_0)+(f_m - f_0)^2] \] 3. **Nonlinear Coefficient**: \[ \gamma_{n,m}=\frac{2\pi f_n}{c^2n^2}\left(\frac{1}{A_{\text{eff}}(f_n)}+\frac{1}{A_{\text{eff}}(f_m)}\right) \] 4. **NLI Power**: \[ P_{n,\text{NLI}}=\sum_{m = 1}^{N_c}\sum_{i = 1}^{2}\sum_{j = 0}^{1}\sum_{k=-1}^{1}\sum_{q = 0}^{Q_i}\frac{(\pi^2\bar{\beta}_2^mR_n\cdot(f_m - f_n+(-1)^j\cdot R_m/2)