Skip to Main Content
Local Controller Net orks (LCNs) provide nonlinear control by interpolating bet ee a set of locally valid, subcontrollers covering the operating range of the plant. Constructing such net orks typically requires kno ledge of valid local models. This paper describes a ne genetic learning approach to the constructio of LCNs directly from the dynamic equations of the plant, or from modelling data. The advantage is that a priori kno ledge about valid local models is not needed. I additio to allo ing simultaneous optimisatio of both the controller and validatio functio parameters, the approach aids transparency by ensuring that each local controller acts independently of the rest at its operating point. It thus is valuable for simultaneous desig of the LCNs and identificatio of the operating regimes of a unkno plant. Applicatio results from a highly nonlinear pH neutralisation process and its associated neural net ork representation are utilised to illustrate these issues.