Modeling and control of industrial processes using linear parameter varying method

Abstract

Modeling the dynamic characteristics of a process is an important step to understand the process behaviour better Most of the industrial processes are nonlinear in nature The identification of accurate models for a nonlinear system is very difficult In nonlinear system identification nonlinear AR MAX models and neural network models are often used These are complex in structure and difficult to compute numerically Block oriented nonlinear models such as Hammerstein and Wiener models are simpler but they can only model nonlinearity in static gains which is often too limiting in process control applications Most of the existing works on process simulation are based on Linear Time Invariant LTI models which is satisfactory for a number of systems However system identification based on the LTI model appears to be of limited value when the plant operating conditions vary significantly and may lead to lower the control performance Thus instead of a global nonlinear description of the plant often an intermediate description is searched for that preserves the advantageous properties of the LTI models and is still able to represent a wide range of nonlinear systems newlineIn model identification for control it is sufficient to have a model that can approximately represent the process behaviour in a thin envelop covering its operating trajectory newline newline