The application of mixed genetic algorithm in parameter identification of circulating fluidized bed units

The maximum likelihood algorithm has some requirements regarding the initial values, or else there would be no guarantee for the convergence of the parameters. The Genetic Algorithm (GA) was introduced in this paper to cope with the problem of local convergence in maximum likelihood algorithm. The typical function test showed that the method had manifested both the global convergence as in GA and the high accuracy rate as in maximum likelihood algorithm. Besides, the special software for general model identification was designed for identification in typical thermal systems. The results showed that it was a readable and valuable method in the realm of identification.