Abstract
dc:descriptionBandwidth requirements double up every three years. The main amount of this traffic is carried out by optical networks which form the backbone of communication systems. Accordingly, optical fiber has been designated as the ideal telecommunication channel allowing to achieve rates of tens of Terabits per second transmitted over thousands of kilometers. Such enormous data rates are mandated in order to meet the exponentially increasing data traffic. Unfortunately, by stressing the current technology to its limits, we have found that the experimental achieved capacity has reached a plateau of 100 Tb/s for a single fiber. To overcome this “capacity crunch”, novel optical transmission technologies have to be introduced such as new modulation formats, new transmission fibers and new multiplexing technologies at the coming years. In this thesis we study the coherent optical systems and in particular we derive closed-form expressions which calculate the power of the dominant nonlinear effect which is Four Wave Mixing (FWM). Assuming that the signal approaches Gaussian distribution when FWM dominates, we can simply add the calculated FWM crosstalk along with ASE noise in a performance metric similar to OSNR and readily calculate BER. This approach eases performance estimation ensuring very good accuracy over a wide range of critical system parameters. Moreover, the derived expressions can be used for optimization in terms of BER, transmission length, system capacity, amplifier spacing etc. Depending on the installed modules (i.e. transmitters, receivers, amplifiers etc.) the performance of optical systems can vary greatly and as a consequence performance estimation mandates analytical solutions adapted in each particular case. Accordingly, the proposed simplified expressions of this thesis can model various transmission systems such as systems with and without guard band, uncompensated and dispersion managed systems, fixed-grid and flex-grid systems, homogeneous and heterogeneous systems depending on if identical or non-identical fiber spans are assumed etc. The proposed closed-form expression for the flex-grid case can be considered as a “universal” solution which can be applied in all the aforementioned systems, demonstrating very good accuracy. For the derivation of the proposed expressions we start from the exact FWM model which is approximated for different scenarios. The accuracy of the resulting expressions is verified by comparing them against 1) the exact FWM model and 2) numerical results using a commercial simulation tool, where the fiber is modeled using Split Step Fourier Method (SSFM). The simplified models are extensively tested in a wide range of system parameters and configurations, modeling more practical systems than the existing models of the literature. It is worth mentioning that the most general expression, developed in this thesis, can calculate the performance of any channel or superchannel of the spectrum, accounting for any channel distribution at the spectrum as well as considering that the channels can have different power levels. Consequently, this model can be used as an integral part of Routing and Spectrum Assignment Algorithms (RSA) which take into account physical layer constraints and can be used for dynamic resource allocation and optimization of various system parameters such as channel power, bandwidth, modulation format etc.Performance estimation is an inevitable process for optical engineers who aim to upgrade or optimize transmission systems. Accordingly, the proposed models of this thesis are compact and can be used for fast and accurate performance prediction whilst they can be easily used to optimize or explore the limits of the current installed systems in terms of capacity, performance, transmission length etc.
Degree
thesis:*- Grantor dc:publisher
- University of Peloponesse
- Year dc:date
- 2016
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Uzunidis, Dimitris
- Ουζουνίδης, Δημήτριος
Subjects
dc:subject × 6Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/41140