In this paper, the deterministic criterion is used for spinning reserve allocation. Uit defines the ON/OFF state of unit i as 0/1. The objective of the UC problem is to minimize the operation cost, subject to the following constraints: Where i, i, i are the startup cost coefficients of unit i and T OFFi t is the off-period of this unit before time t. In this equation ai, bi, ci are cost coefficients of the i th unit.įCit (Pit ) = ai P2i t + bi Pit + ci $/h (2)In (1), STit is the startup cost of units, which can be formulated as: Fuel cost function is generally formulated ina quadratic form shown in (2). For N generating units, the operation cost ina horizon time T can be calculated as:įCit () is the fuel cost function and P MWit is the generated power. The operation cost isthe summation of fuel costs and startup costs of generating units.
The object of the UC problem is to minimize the operation cost in the planning period. It is shown that the operationcost reduced more than $8,000 and $18,000, respectively.
The results are shown for10- and 26-unit power systems in comparison with GA and binary PSO methods. In this paper, a new method is proposed using particle swarm optimization in composition with a proposedsearch method which can be used in order to solve the unit commitment problem. Successful PSO applications indifferent optimization problems, such as function minimization and neural networks parameter tuning, showthe capability and usefulness of this method. Incomparison with GA, PSO is easier to implement and has less parameters. The system isinitialized by a set of random values and then looks for the optimal solution through updating these candidates.In PSO, potential solutions or particles move in the solution space while tracking the optimal particles. Like genetic algorithm, PSO is an optimization method based on the Swarm Intelligence.
Intelligent methods do not limited by restricting assumptions about search space, such as continuity orexistence of derivatives. Khorasani (B)Department of Electrical Engineering, Bojnourd Branch, Islamic Azad University, Bojnourd, IranE-mail: Arab J Sci Eng (2012) 37:10331042 Thesemethods can find an appropriate answer in an acceptable time, but they cannot guarantee finding the optimalsolution. Heuristic and meta-heuristic methods look for an answer in a subspace of the total search space. The conditions and characteristics of each method make it useful for specific problems. Enumeration priority listing, dynamic programming, linear andinteger programming, Lagrangian Relaxation, interior-point optimization are theoretical methodsand Tabu search, simulated annealing, expert systems, fuzzy optimization, neuralnetworks, genetic algorithm (GA) and particle swarm optimization (PSO) are some of theintelligent methods. There are different methods used to solve unit commitment problem, from very complicated and theoret-ical through heuristic methods. The main objective of the unit commitment problem is tominimize the generation cost subject to the technical constraints of the generating units and the power systemsecurity constraints. This is called unit commitment (UC) problem. An important issue in power systemoperation is to supply demand using an optimized combination of generating units at minimum operationcost. In almost all integrated power systems, thermal units supply the base load. Keywords Unit commitment Heuristic methods Particle swarm optimization The results are more accurate and get loweroperation costs in comparison with other traditional methods. An expert system is utilized to generatethe search space combinations of the generating units and a particle swarm optimization algorithm is usedto optimize generation levels according to each combination. This method limits the search space considering demand information and the generation historyof the generating units to solve the UC problem in an hourly basis.
Received: 10 February 2010 / Accepted: 3 November 2010 / Published online: 12 April 2012 King Fahd University of Petroleum and Minerals 2012Ībstract In this paper, a new method is proposed to solve unit commitment problem using particle swarmoptimization. RESEARCH ARTICLE - ELECTRICAL ENGINEERINGĪ New Heuristic Approach for Unit Commitment ProblemUsing Particle Swarm Optimization