Abstract:
Presently most of the world’s electricity demand is met by thermal power generating stations which purely growth on conventional fossil fuels. Every thermal generating unit is characterized by its distinct incremental heat rate curve which directs the production cost. An optimal allocation of generation among the different power generating units can thus save significant fuel input and cost. However, a larger amount of fuel cost can potentially be saved by spreading this optimization process to decide which of these thermal units would participate in the optimal allocation. It is vital to determine whether the current unit must be ON/OFF and committing thermal generating units among the available ones in this mode is popular as Unit Commitment (UC). In general, UC schedules are determined a day ahead. On the other hand, optimally allocating the generation among the set of units already committed/scheduled for operation is familiar as Economic Dispatch (ED).
In power generation, the operational planning is a major activity. UC is a significant economic problem in thermal generation systems with main condition is meeting the load demand under a specific time horizon while all the unit operational constraints are taken in account. The UC is highly composite problem in power generation system operational planning. The deterministic (conventional) techniques comprise priority listing (PL), DP (dynamic programming), simulated annealing, IP and mixed-integer programming, lagrangian relaxation, tabu search etc. The non-conventional techniques include expert system, genetic algorithms, fuzzy system, artificial neural network, particle swarm optimization, evolutionary programming etc. The conventional and nonconventional optimization techniques are not capable to effectively solve this complex system.
In this thesis, a Hybrid Algorithm is used in solving the UC problem. Hybrid algorithms are developed by combining both conventional and nonconventional optimization methods. Hybrid techniques are reduced number of variable and problem size. The implementation time of the hybrid methods growths approximately linearly according to problem dimension, hence takes less overall execution time. In this research, Hybrid Algorithm has been tested on different standard test systems and its effectiveness has been confirmed by comparing the simulation results with those of other algorithms in literature. Simulation results have been calculated by using MATLAB.
