Scheduling Power Production of Hybrid Power Systems with Data Mining and Computational Intelligence

Description

Hybrid power systems are developed to tackle the intermittency of wind power output through integrating wind turbines with other generation sources. Knowledge of operating such systems is in its infancy. This project aims to offer a pioneering study of scheduling the power production of hybrid power systems. A scheduling model produces schedules for minimizing the total cost in power production will be developed. The schedules include two types of decisions, the on/off states of power generation units and the control settings to determine generation capacity of operated units.Hybrid power systems are deployed in various configurations. A wind-thermal-solar system configuration is considered in this project. The wind-thermal-solar system includes a set of wind, thermal, and solar power generation units. A cost model will be developed to estimate the total cost of operating such a system. Data-driven approaches will be applied to build wind and solar power generation models to accurately estimate the power output and capture characteristics of power generation efficiency. The scheduling problem is a mixed-integer nonlinear programming (MINLP) model with binary and fractional variables. The cost model will serve as the objective function and the data-driven wind and solar power generation models will be incorporated in the constraints. To solve this MINLP model, a novel computational intelligence algorithm which is composed of computational intelligence algorithms for discrete and continuous optimization will be developed. A comparative analysis will be performed to assess the effectiveness of the proposed scheduling model and its computed schedules. Different allocation of the installed power generation capacity of generation sources in the wind-thermal-solar system will be analyzed. Techniques from design and analysis of computer experiments will be deployed.