Study Of Gain-spectrum-flattening Of Fiber Raman Amplifiers

As the Olympic motto of"Faster, Higher, Stronger"inspires, a broader transmission bandwidth, a higher transmission rate and a longer transmission distance have always been the target long haul optical fiber communication systems aspire after. Wavelength division multiplexing (WDM) and optical amplifier technologies are the crucial boosters for the evolutions of development. Among various optical amplifiers, fiber Raman amplifier (FRA) is an enable technology for WDM optical fiber communication systems due to their low noise figure, flexibility of gain band and broad bandwidth amplification. In this dissertation, the following progress was made in aiming to flatten the gain spectrum of a broadband FRA.(1) Flattening of gain spectrum of a distributed broadband FRA.A distributed broadband FRA uses conventional fibers with small and uneven Raman gain efficiency coefficient as gain media. Multiple pumps are usually needed to achieve a flat gain spectrum of an FRA. Regarding this issue, the following creative works are carried out.(1.1) Three shooting algorithms were proposed for stably and efficiently solving the Raman coupled equations with multiple counter pumps. In these shooting algorithms, the starting values of counter pumps are determined by the analyses of physical pictures of stimulated Raman scattering in fiber and some newly introduced parameters. In addition, the Newton-Raphson method for solving a set of nonlinear equations was modified to adjust the initial values of counter pumps. The simulations of multiple FRAs showed that the calculation efficiency and stability of these shooting algorithms are improved dramatically when compared with the conventional shooting algorithm.(1.2) An efficient method for extraction of input pump powers from pump power integrals was proposed. In this method, a set of coupled nonlinear equations of input pump powers for given target pump power integrals are constructed. Finding the input pump powers corresponding to the target pump power integrals is mapped to finding solution of this set of nonlinear equations. By combined application of Newton-Raphson root-finding method and the mean value theorem for integrals the set of nonlinear equations was solved efficiently. The calculation results showed that the proposed method reduces the computational effort by two orders of magnitude when compared with those previously reported techniques.(1.3) The particle swarm optimization (PSO) was modified and introduced to the design of FRA. The application to multiple benchmark functions showed that the modified PSO (MPSO) improves the efficiency and convergence of the program greatly, in the mean time keeps the simple and elegent feature of the basic PSO.(1.4) By organic combination of PSO, MPSO and the 3 shooting algorithms, average power analysis, three design methods for a multiple-wavelength-pumped broadband gain-flattened FRA were proposed. The design results of multiple FRAs indicated that the performances of these methods are all improved when compared with the old, traditional methods in literatures.(1.5) An intelligent implement method of FRA was proposed based on the thorough analyses of drawbacks of conventional implement methods. In this new method, an exact mathematical model of FRA to determine what happens and how it happens is not needed. Instead, a FRA is regarded as a black-box. The driving currents that determine the pump powers are treated as the input of the black box and the corresponding gain spectrum are treated as the output of the black box. The method achieves the flattening of gain spectrum of a FRA by adjusting the driving currents via an intelligent approach. In order to verify the feasibility of this intelligent implement method, detailed schemes of software and hardware were presented and design of multiple intelligent FRAs were carried out. The experimental results showed that the proposed method is an efficient and intelligent approach for the implementation of multiple-wavelength-pumped broadband gain-flattened FRA.(2) Flattening of gain spectra of a discrete broadband FRA.The above-mentioned methods of flat-spectrum design of a distributed FRA are also applicable to a discrete broadband FRA. On the other hand, a discrete FRA has its own methods to achieve flat gain spectrum because fibers of various special feature can be adopted as gain media in this case, in contrast with the distributed FRA where the transmission fiber is the gain media. On the flattening of a gain spectrum of a discrete broadband FRA, the following creative works were made in this dissertation.(2.1) A novel twin-core photonic crystal fiber (NPCF) was proposed. This NPCF possesses a higher and flatter Raman gain efficiency (RGE) spectrum over a specified band of wavelength than a conventional fiber. This feature is achieved by appropriate design of its cross-section. Therefore, it is a good candidate of gain medium for a flat, broad gain band FRA. It was numerically demonstrated that the relative fluctuations of RGE of the NPCF reduces to 3.3% in the C (1530 ~ 1565 nm) band, while that of a conventional single mode fiber (SMF) is 49.6% in this band. Moreover, its value of RGE is 1.3 times and 2.7 times larger than that of a SMF respectively at the long and short wavelength end.(2.2) By using the NPCF as gain medium a C-band FRA was designed. It was numerically demonstrated that when pumped with a single source, an average gain of 8.7 dB with a fluctuation of less than 0.9 dB is achievable. It is a remarkable improvement over a conventional FRA. The NPCF reduces the complexity and cost of an FRA greatly.