Study On The Performance Optimization Theory And Design Method Of Fiber Raman Amplifiers

Theoretical models for fiber Raman amplifier (FRA), including power propagation coupled equations and complicated full phenomena nonlinear Schrodinger coupled equations, are developed along with the fast numerical simulation algorithms. Based on these models, a comprehensive theoretical study on the characteristics of the gain, noise, nonlinear effects and polarization dependence of FRA is carried out. First, the relation between the Raman gain and the operation parameters is investigated. The method of increasing the gain, the power conversion efficiency and the saturation output power are presented. Secondly, the polarization mode dispersion induced Raman gain fluctuations are studied. The method of obtaining the probability density function of Raman gain is developed, giving a quantitatively description of the gain fluctuations along with the method of restraining it. Thirdly, a new analytical expression for the noise figure of FRA is derived. Compared with exiting ones the proposed expression has an obvious improvement in the accuracy. Using this expression, the impacts of ASE and DRS noise on the NF are investigated. The results show that in the low gain domain the performance is mainly limited by ASE, while in the high gain domain the performance is mainly limited by DRS. And there exits an optimal pump power and media fiber length with the lowest NF. In a further discussing, the characteristics of the noise and nonlinear effects of the bidirectionally pumped FRA is investigated. By optimizing the ratio of forward pump power to total pump power, the performance has an obvious improvement compared with backward only pumped FRA. Fourthly, an analysis of the performance of FRA cooperating with dispersion managed fibers and EDFA is performed yielding an optimum configuration for the ultrahigh-speed and ultralong-haul optical transmission systems. For a given transmission distance and performance the optimal configuration has the fewest repeaters. While for a given transmission performance the optimal configuration has the longest transmission distance. Eye diagrams and bit error rates (BER) are numerically obtained for the validation of the results. Fifthly, a novel optimalalgorithm for gain-flattened multi-wavelength pumped FRA is proposed. The algorithm based on genetic algorithm and neural networks has a dramatic improvement in the accuracy and efficiency.