| With more and more demands for high speed and large capacity of communication system, the technology of optical communication is improved greatly. DWDM has been applied widely. As one of the most important components of optical fiber communication systems, optical amplifier also needs new techniques. EDFA, which is widely used now, has been beyond its performance limit and can't meet the requirement of large capacity, high speed and long haul. Fiber Raman amplifier has two critical merits of the low noise level and the arbitrary gain bandwidth, which can not only compensate the fiber loss but also enlarge network capacity. So it has become a hot research topic in the field of optical communications.In the dissertation, we first introduce the current status of optical communication technique and optical amplifier, compare the characteristics and applications of semiconductor optical amplifier, rare earth ion-doped fiber amplifier and nonlinear fiber amplifier and analyze the history and the developing prospect of Raman amplifier. Then we study the basic theory, the configuration, the pumps and the noise characteristics of Raman amplifier as well as the improvement of system performances.Our research is emphasized on the mathematical model, numerical algorithms, gain characteristics and optimal design of multi-pumped Raman amplifier.The simulation model of multi-pumped Raman amplifier is complicated, which is based on a set of power coupling steady-state equations that include the effects of Raman interaction, fiber attenuation, Rayleigh backscattering, spontaneous Raman emission noise and thermal noise. In this paper, we use a simplified multi-pumped amplifier model to decrease the complexity of original model structure without impairing the simulation accuracy.It is a boundary value problem of ordinary different equation which we can solve using the method of direct integration. However, it takes exhaustive computational run-time to achieve well-behaved performance. People proposed some practicalmethods to reduce computational run-time. We analyze and compare some of them such as average power method, multi-steps method and PCM. Compared with the method of direct integration they can improve the simulation speed evidently. But their main drawback is that each of them uses fixed step-size. We expect a solution in which step-size follows the change of power. In this paper, we propose an adaptive step-size method based on Runge-Kutta-Fehlberg formulas. The proposed method can deal with adaptive step-size adjustment expediently by calculating the truncation error of each step and solve a two-point boundary value problem for Raman amplifier propagation equations appropriately.The efficiency of our adaptive step-size method and that of one-step method, four-step method and PCM is compared. Test results demonstrate that our adaptive step-size method can effectively improve the computing speed with lower error compared with the other three methods. As an application of the adaptive step-size method, we simulate a FRA where 100 forward WDM signals are amplified by five backward pumps by adaptive step-size method and draw some useful conclusions.It is also an important task to research the flat gain spectrum of Raman amplifier. In this paper, we simulate the factors of the gain spectrum of FRA and analyze the effects of the SRS among pumps and signals, pump number, pump power and pump wavelength on the flatness of gain spectrum.Raman gain profile can be adjusted by appropriately choosing the relative positions and powers of the pump waves. In the last part of the paper, an efficient algorithm for the optimal design of FRA—genetic algorithm is introduced. And as an illustration, we achieve a flat gain profile about 6dB with a maximum gain ripple less than ldB and about 90nm bandwidth. |
PDF Full Text Request