Unit Commitment And Economic Dispatch In The Combined Cycle Power Plant

With the improvement of the electrical industry step by step, independent power plants have to enhance their competitive capability to survive and be improved in the electrical market competition. Unit commitment (UC) and load economic dispatch (LED) is one of the most important methods to reduce the cost of producing electricity and increase the profit of the power plants for multi-unit plants.The UC and LED models for a electrical systems with Frame F combined cycle units under the environment of the modern electrical market are developed in this paper. Some new constrains are added in these models because of the method of classification combustion used in this kind of units. Comparison to the conventional thermal unit, combined cycle units are affected by more environmental conditions during the unit load variation process. The heat cost curve which only takes unit load to account can not reflect the real situation during this process. A model of unit load variation process based on feed forward artificial neural network which is trained by the unit historical data is constructed in this paper. This model can achieve the mapping from the affecting factor to the heat cost ratio and is use in the UC and LED models to replace the heat cost curve.Based on the comprehensive comparison of some sorts of algorithms available for solving UC and combining the characters of the models being proposed in this paper, genetic algorithm was chosen as numerical computing method. Some improvements concerning operator designing, constrain handling etc. were made so as to make the algorithm more suitable for the models constructed in the paper. Numerical analysis illustrates that the improved genetic algorithm designed in this paper is successful in solving high-dimension, multi-constrains, and nonlinear UC problem and the solving method with the load-variation model of the combined cycle unit based on the artificial neural network can acquire better result than the method with the heat cost curve of the unit.