ISSN: 2319-8753 International Journal of Innovative Research in Science, Engineering and Technology Vol. 2, Issue 4, April 2013 Copyright to IJIRSET www.ijirset.com 988 Multi Parameter Gain Optimization of Raman Fiber Amplifier for Dense Wavelength Division Multiplexed Systems Sonak Saini 1 , Simranjit Singh 2 M. Tech. Student, Department of ECE, UCoE, Punjabi University, Patiala, Punjab, India 1 Assistant Professor, Department of ECE, UCoE, Punjabi University, Patiala, Punjab, India 2 Abstract: In this paper, various parameters of the Raman Fiber Amplifier are optimized for DWDM system at 0.2 nm channel spacing. The various system parameters attained optimum value to give gain value targeted at 20dB for each of the 100 channels. The multi stage approach adopted in this work results in providing maximum gain (above 15dB) for the optimized system parameters i.e. for pump power nearly equal to 600mW, signal power lies between -4 to -5dBm, pump wavelength equals to 1460nm and Raman fiber length is approximately 25Km. Keywords: Raman fiber amplifier (RFA), Multi parameter optimization (MPO), gain optimization, DWDM, Raman length. I. INTRODUCTION Raman Fiber Amplifiers (RFAs) have been known for providing a simple single platform for long-haul Dense Wavelength Division Multiplexed (DWDM) systems [1]. Raman amplification mechanism in optical fibers is of great importance. In multiple wavelength telecom system, Raman amplification offers the ability to achieve gain flatness without inserting any wavelength-dependent lossy elements [2]. RFA comprises several advantages which have attracted much attention during the recent years [3]-6]. The key merits of Raman amplifiers along with flexibility of gain of amplifiers are: (i) Gain is non-resonant, i.e. gain is available over the entire transparency region of the fiber, (ii) Gain spectrum can be tailored by adjusting the pump wavelengths, (iii) the medium for propagation and amplification are the same, (iv) gain is independent of relative direction of a pump and a signal, (v) broad gain-bandwidth. One of these distinctive merit is the broad gain-bandwidth associated with RFAs which is an attraction for application to the DWDM systems [7]. A large number of research papers have been dedicated to the optimization of the RFA gain by considering multiple pumps. B. Neto et. al [3] used Hybrid Genetic Algorithm for gain optimization of a system covering entire C-band by optimizing pump powers. Farzin Emami et. al [4] employed Fuzzy Adaptive Modified Particle Swarm Optimization (FAMPSO) to obtain pump powers and pump wavelengths with minimum gain ripple for S-band. Junhe Zhou et. al [5] presented a Novel robust, compact and flexible neural network model for a Fiber Raman Amplifier to obtain optimal pump power and wavelength configuration to attain desired gain spectrum. Pencheng Xiao et. al [6] proposed a new optimal algorithm scheme based on neural networks model for multipump sources of distributed RFA. Zhi Tong et. al [8]analyzed broadband Distributed Raman Amplifier (DRA) for its OSNR and net gain and optimized a bi-directionally multi-wavelength pumped RA. B. Neto et. al [9] employed a Hybrid genetic algorithm in the gain optimization of multipump DRAs by efficiently combining Genetic algorithm with the Nelder- Mead method. According to the above reported work, RFAs have been analyzed and optimized for all the three bands (S, C and L-bands) [3], [8], [9] but most of the work has been done for Wavelength Division Multiplexed (WDM) systems with 5nm [4], 0.8nm (100 GHz) [5], [8] or even more channel spacing. Moreover, all the optimization criteria have been devoted towards pump power and wavelengths. There are also many other parameters for RFA which need to be optimized for improving the gain spectrum. Hence, multi-objective or multi-parameter optimization is required to be introduced. Recent work presented by [10-16], has deployed numerous numerical methods and analytical approaches. Junhe Zhou et. al [10] proposed an analytical design for RFAs with Time Division Multiplexed (TDM) pumps but for the channel spacing equal to 100 GHz (0.8nm). Hai Ming Jiang et. al [11] introduced Ant Colony Optimization (ACO) but having the system channel spacing equal to 1nm.Carmelo J. A. et. al [12] presented a Multi- objective Particle Swarm Optimizer (MOPSO) to define the number of pump lasers and their wavelengths and powers with a 40 WDM signal channels at 100 GHz of separation. Gustavo C. M. et. al [13] also used a Hybrid Genetic algorithm with Geometric Compensation Technique for simultaneous analysis of multiple parameters and multi- objective problem but few of the important parameters have been presented. Javeria Yasmin et. al [14] optimized RFA parameters like pump power and wavelength for WDM chaotic communication but presented the work for a channel spacing of 0.8nm only. Ferreira et. al [15] proposed a low complexity computational model for the gain of the RFA, to be used suitably with multiple pumps and a large number of signals, but used a tunable laser operating at 1mW instead of considering the whole broadband. Despite of a numerous work on the optimization of RFA, still there is a need to