A maximum likelihood identification of block industrial systems with input nonlinearity and applying to the chemical process

According to the maximum likelihood principle, a recursive maximum likelihood identification method is proposed for the identification of block industrial systems with nonlinear inputs. The system uses the Hammerstein structural model to fit the dynamic input and output data. The parameters of the nonlinear block structure and the static parameters of the system are coupled together, which makes it difficult to separate the parameters and makes the estimation model parameters not unique. The proposed identification method is a kind of parametric method, which facilitates analyzing the dynamic characteristics of the system and designing the system controller. The consistency of the proposed algorithm is also proven. A numerical example and a practical chemical process are used to demonstrate the superiority and usefulness of the proposed algorithm. The algorithm can be extended to multi-variable systems or other real industrial systems.