Analysis And Design Of Multi-Pumped Broadband Fiber Raman Amplifiers

Fiber Raman amplifier(FRA) can amplify the whole wave band, get a wideband gain spectrum using multiple pumps, and has low noise、low cost. So FRA is more suited for the wavelength division multiplexing system, and maybe will replace the traditional erbium-doped fiber amplifier, being the key device in optical fiber communication system. This paper mainly studied the optimal algorithm for multi-backward-pumped Raman amplifier which obtaining flat gain spectrum, proposed a new method-combinatorial genetic algorithm for optimizing pump power and wavelength to flatten the Raman gain. The specific content includes:Firstly, considering fiber loss、Rayleigh scattering、ASE noise、the complex stimulated Raman scattering between pump-pump、signal-signal and signal-pump, this paper analyzed the power coupled equation of multi-pumped Raman amplifier and introduced two numerical algorithm, Runge-Kutta method、 average power method. In order to solve backward-pumped Raman amplifier’s specific boundary value problem, a new method average power+shooting method is used to solve multi-backward-pumped Raman amplifier power coupled equation. The MATLAB simulation result shows that the convergence speed of this algorithm is very fast, iteration is about four or five with high accuracy. Through comparing the results of other reference, it proved that this method is effective in solving Raman coupled equation.Secondly, based on the introduction of three optimization algorithms:simulated annealing method, LTI method and genetic algorithm, a new optimization method-combination genetic algorithm is used to optimize pump wavelength and power of multi-backward-pump FRA. Through stimulating in MATLAB, we can get80nm wide, less than1dB gain fluctuation for three backward pump and100nm wide, less than0.5dB gain ripple for five-backward-pump, Thus the algorithm can optimize Raman amplifier with any pumps, and obtain flat gain spectrum with high running efficiency, the whole process takes only a few minutes.Thirdly, through comparing the results of combined genetic algorithm and computing result which using optimal parameters, it proved that this algorithm is effective in optimization design of multi-pumped Raman amplifier, and has better accuracy, high efficiency, wide application range, is more practical.