A steam turbine is one of the most important parts of a power plant, whose main pressure influences the system thermo-economy a lot. With the upgrading of industry structure, the electricity structure in our country has changed greatly, which means the peak-valley gap of power grid is getting bigger. What’s more, many large-scale power plants have to participate in the peak regulation. As a result, the operating efficiency decreases. In this paper, the initial pressure is adjusted by a newly proposed optimization algorithm, which increases the system thermo-economy. The main contents are as follows:Firstly, Echo State Networks is studied and some relevant thermal parameters are analyzed in detail. Then the parameters which are closely related to the heat rate are chosen to be the input parameters. And the Echo State Network is used to build the prediction model of the heat rate. The simulation result proves that the model has a good predicting accuracy and shows a high generalization capability.Secondly, for the shortages Krill Herd(KH) has, an improved Krill Herd algorithm(I-KH) is proposed. Specifically, the foraging behavior of onlookers from Artificial Bee Colony algorithm, which has the dominance of quickness and accuracy, is introduced to the algorithm. This behavior is combined with the foraging motion and motion induced by other krill individuals. After a few iterations, performance of the new algorithm is improved. Then the algorithm is applied to 8 benchmark optimization problems.Experiments show it has very good global search ability and convergence speed.Finally, after the prediction model of heat rate is built, I-KH is adopted to search the optimal initial steam pressure in the range of permitted pressure. The simulation experiments show that the heat rate reduce. Then a new sliding pressure operation curve,which is similar to the trend of the design one, is obtained. The new curve can reduce the heat rate and can also guide safety and economic operation of steam turbine. |