Study On Exact Evaluation Algorithm Of Four-wave Mixing Noise In Dwdm System

Recent years, fast development of the HDTV, video on demand and mobile broadband serviceputs forward an urgent demand on the requirement of bandwidth and capacity in opticalcommunication network. It promotes the rapid development of DWDM technology.As DWDM technology develops, many fiber nonlinearities have imposed more and morebadly influence on transmission performance, especially four-wave mixing (FWM) effect. Underthe influences of various effects, exact and computationally efficient evaluation method of thedeviation of four-wave mixing (FWM) noise is very important for optimizing design of DWDMsystem.Nonlinear Schrodinger equation (NLSE) is an important equation used to describe opticalpulse transmission. For most cases, NLSE cannot be solved analytically. Numerical methods areusually adopted to solve it. Among them, split-step Fourier method (SSF) is one of the mostcommonly used algorithms.In order to obtain exact numerical result, small step-size tends to be used for constant step-sizemethod. However, smaller is step-size, more is time consuming. Improper step-size selection canpossibly lead to severe spurious FWM tones. This paper discusses several step-size selectionmethods, and makes a more detailed analysis on dynamic split-step Fourier method.Ignoring inter-channel crosstalk but taking into account the interaction of dispersion and someintra-channel nonlinearities, this paper firstly obtains pulse evolution with distance by usingdynamic split-step Fourier method. Including the numerically derived pulse evolution with distance,channel overlap functions, which determine the efficiency of FWM effect, are further numericallysolved. In this way, this paper provides an effective way to numerically calculate the variance ofFWM noise.The main innovations of this paper are given as follows:(1) A new calculation model of FWM efficiency is introduced. Compared with the traditionalcalculation models, the new derived model mainly includes many effects, such as walk-off betweenchannels, random bit sequences and especially pulse shape and additional phase changing withdistance, which is induced by the interaction between dispersion and some nonlinear effects.(2) By using the model, this paper gains some relationships between the degenerate FWMefficiency and system parameters when launched pulse shapes are chirped Gaussian pulse and super-Gaussian pulse. When group-velocity dispersion is small and pulse width and peak-power ofincident pulse is larger, pulse changing with distance of each pump channel induced by theinteraction between dispersion and nonlinear effects is more important to the FWM efficiency.FWM efficiency with40Gb/s compared with10Gb/s, has a relatively large value with same systemparameters. In case of high bit rate per channel and a large and positive chirp of pulse shape, thepulse changing with distance induced by the interaction between dispersion and nonlinear effectsmust be taken into account especially.(3) By using the model, this paper also discusses the cases with new intensity modulationformats. Following some comparisons of calculation results in case of them, conclusions are given.Generally, under same system conditions, larger is duty cycle, larger is the FWM efficiency. It isobserved that modulation format with narrow spectrum has better dispersion tolerance, but its FWMtolerance is worse generally.