Investigation Of The Four-wave-mixing Effect In DWDM System Of Intensity Modulation

With the increase of the transmission capacity of dense wavelength division multiplexingsystem, the channel spacing becomes bigger; the incident power of light becomes higher. It hasbecome particularly important to estimate comprehensively the nonlinear effects in fiber, especiallyFour-Wave Mixing (FWM). As we know, the effect of FWM in Dense Wavelength DivisionMultiplex (DWDM) depends on many factors, such as modulation schemes of signals, dispersionmaps, channel spacing. Analysis of FWM in DWDM is based on two cases in this paper. One ofthem, we call―CASE A‖, is that the input polarizations of all channels are same and birefringencealong the whole fiber line does not exist. The other, called―CASE B‖, is that the input polarizationstates of all channels are independently random and birefringence strength and orientation alongfiber vary randomly.Firstly, under the condition that the launched pulses are chirped Gaussian pulses, the random ofthe pattern and initial phase of bit sequence in each channel is concluded, a semi-analyticcalculation model applied to evaluate the variance of degenerate four wave mixing noise indispersion managed wavelength-division-multiplexing (WDM) system for case A is derived. Thepulse amplitude expressions for the chirped Gaussian pulse in dispersion-managed fiber link arededuced in detail. The dispersion map is supposed to be consisted of two fiber segments. Thedispersion sign of the two fiber segments is opposite, but the absolute value of dispersion of them issupposed to be same. The length of each segment is variable. Some examples shows that thecharacteristic widths of pulse have an obviously impact on variance of FWM noise and the varianceof FWM noise increases with the characteristic widths of pulse; When the characteristic widths ofpulse is smaller, the chirp parameter values have a great influence on the FWM effect. Optimizingdispersion management scheme can make variance of FWM noise small.Secondly, the paper is based on case B, Four-wave-mixing (FWM) acts as noise due torandomness of bit sequences of all channels for wavelength-division-multiplexed (DWDM) system.The approximate direct ratios of variances of FWM noise between two cases are obtained.Including randomness of bit sequences and walk-off between channels, forintensity-modulated/direct-detection (IM/DD) dispersion-managed (DM) DWDM system, theperformance degradation due to FWM effect is calculated. The calculation model is based on the direct ratios between the above-mentioned two cases derived in this paper and a semi-analyticapproach for―CASE A‖having been introduced in previous works. The calculative results provideoptimized dispersion maps used to reduce the FWM noise. For the calculative that initialpolarizations of all channels are controllable, launched polarizations between neighboring channelsare suggested to be orthogonal in order to reduce the overall FWM noise from all channelcombinations.Finally, the parameter of FQP is defined. The example shows that the value of FQP dependsmore strongly on when the value of S dispersion map strength becomes smaller. A generaltrend also can be inferred from the example that larger values of S correspond to smaller values ofFQP.