Self-adaptive design of a nonlinear temperature control system
The paper describes the design and implementation of an unusual self-adaptive temperature control system for a nonlinear heated bar system. The design is based on a proportional-integral-plus (PIP) control scheme, combined with a special recursive algorithm which continually updates the estimates of the rapidly variable parameters which characterise a piecewise-linear model of the heated bar system. Amongst the novel features of this true digital control (TDC) system design are: the nonminimum states-space (NMSS) formulation of the control problem, which allows for state-variable feedback pole assignment using only sampled input and output signals in the control law; the incorporation of a priori information on the physical nature of the system into the recursive estimation algorithm in order to enhance its parameter tracking capabilities; the exploitation of multirate sampling; and the implementation of the linear PIP control law in a bang-bang mode using pulsewidth modulation.