ABSTRACT
This chapter deals with the estimation of changes in the underlying parameters of the plant. It reviews a two-step procedure in which the model parameters are estimated first. The chapter shows how the underlying parameter changes can be identified in a single step. It comments on the excitation requirements of the two procedures. The chapter shows that the parity relation approach is a limiting case of parameter change estimation, which uses the minimal dataset under which estimates can still be obtained. It shows an alternative way which utilizes the parity equation residuals obtained with the nominal model. It explores the transient behavior of the estimation algorithm following a parameter jump. The chapter considers how parameter bias and uncertainty, caused by the presence of noise, affect the change estimation. The relationship between the physical parameters of the continuous-time plant and the parameters of its discrete-time model are usually complex and nonlinear.
