Arz/talep belirsizliği ve emisyon sınırlamaları altında geleneksel-ve-rüzgar entegre hibrit enerji sistemlerinde birim yüklenme problemi için çözüm yöntemleri

Unit commitment problem (UCP) is one of the essential problems in operations planning of power generation systems. The objective is to minimize total operating cost while meeting the forecasted load requirements and satisfying several operational and technical constraints. Nevertheless, the UCP is a mixed integer, non-linear, combinatorial and NP-hard problem, making it difficult to develop any rigorous optimization method for a real-size system. In this thesis, we address two variants of the UCP: (1) the deterministic UCP in conventional power systems, (2) the stochastic UCP in wind integrated hybrid power systems. For the first one, an effective and efficient Genetic Algorithm-based approach is developed. For the second one, Mixed-Integer Quadratic Programming-based approaches are developed. In these approaches for the stochastic UCP, novel expected energy not served (EENS) approximation methods are proposed to model both load demand uncertainties and supply uncertainties due to intermittent nature of wind power generation and outages in conventional generation. Furthermore, the proposed approaches are extended to consider: (i) the Valve Point Loading Effect in efficiencies of conventional generating units by proposing efficient multi-area piecewise linear approximation, (ii) the impacts of Emission Control Technologies and Emission Trading and Taxing Mechanisms in mitigating greenhouse gas and air pollutant emissions caused by conventional generating units. According to numerical experiments and sensitivity analysis results, both Genetic Algorithm-based and Quadratic Programming-based approaches are proven to be valid and effective, and they can provide satisfactorily good power generation schedules for large scale power systems in a reasonable computational time.