Abstract:Compared with the rapid development of big data industry in recent years, the development of big data modeling technology in process industry is relatively slow. In this context, the big data modeling technology of thermal power units is selected as the research object. Firstly, the characteristics of big data of thermal power units and the difficulties faced in big data modeling are analyzed. Then, the big data modeling methods and existing problems of thermal power units at home and abroad are studied. Finally, an all-condition adaptive transfer function modeling method based on big data of thermal power units is proposed. This method uses data mining technology to select qualified modeling data under different operating conditions, and applies data-driven transfer function modeling method to build a large number of system models in the whole operating conditions through cloud computing platform. Then, linear parameter varying and mathematical interpolation algorithms are used to fuse these system models, and finally the adaptive transfer function model of the system under all operating conditions is obtained. This modeling method provides a good reference for big data modeling of thermal power plants and other process industries.