Identification of complex nonlinear processes

Abstract

The estimation of the unknown parameters in models described by algebraic equations, ordinary differential equations, and partial differential equations, as well as the initial state vectors in these models, was carried out by the maximum plausibility technique. As in problems of parametrical identification, quality functions have a complex form and contain a significant number of local minimums; to decrease their amount and simplify the form of the quality function in regions far from the absolute minimum, the method of asymptotic decomposition was suggested for problems of the identification and estimation of the parameters. For the application of the suggested procedure, the technology of the parametrical identification of the physical-chemical processes in a furnace plant working in the mode of gas production with a fluid bed was introduced. A measurement model was developed by the experimental data obtained during the combustion of brown coals in primary furnaces with a fluid bed at HES-2 in Chelyabinsk.