Adaptive Genetic Algorithm For Parameters Identification Of Centrifugal Compressor Model

Centrifugal compressor is the key equipment in the industrial production. They have the advantages of exhausting high pressure and conveying small flow, and they have been playing a crucial role in aerospace, energy, chemical, metallurgy and other industries. In order to meet the needs of recycled economy development, making full use of surplus coal gas in the process of iron and steel smelting, reducing the consumption of energy, while reducing the pollution due to the coal gas discharging, Baosteel built gas-steam combined cycle power project, which has both economic and social benefits. Centrifugal compressor supplies pressured gas to the gas turbine, which is one of the most important equipment of the whole generation process. For performance analysis and study of centrifugal compressor system in the actual production process, establishing the mechanism model the compressor system is the main technical method, however determining parameter identification combining with actual operation of data of the system is a problem to be solved.The paper considers the centrifugal compressor of Baosteel combined cycle power plant as a research object, and does research on the problems of parameters of multistage centrifugal compressor model. By analyzing the working mechanism of centrifugal compressor, and using the energy conservation in the process of compressor doing work for gas and the gas mass conservation relationship, the paper establishes the mechanism model of centrifugal compressor. Combining with the large quantities of measured datum, and identifying the four nonlinear parameters of mechanism model by SAGA, the paper obtains the accurate model of centrifugal compressor. This paper also compares SGA, CAGA with SAGA, and the simulation results show that SAGA has faster convergence speed, better stability. SAGA is suitable for processing the complex and nonlinear problems and possesses global optimization ability. Moreover, the relative error between the actual measured value and the model output after parameters identification is smaller. Model validation results show that the parameters identification can reflect the operating characteristics of centrifugal compressors accurately and the precision of the model is improved.