Skip to Main Content
In this paper, a new technique for adaptive control of nonlinear chemical processes based on feedback linearization method is presented. The method is based on semi-mechanistic or grey-box modeling. By this method, only the difficult-to-model part or complicated part of the plant is identified by an adaptive neural network and the remaining parts of the system are obtained online by using first-principle model of the process and special measuring techniques. The method is applied to a well-known CSTR benchmark process and its advantages and drawbacks are discussed.