ABSTRACT
This chapter describes some of the special features and problems, encountered when identification is used to establish the plant model for the detection and isolation of additive faults. It is concerned with the effect of parameter bias in the identified model. The chapter deals with the problem of identifying the model from linearly dependent data. It describes the powerful and practically important technique of direct identification of structured parity relations. The chapter addresses the issue from the point of view of the interaction between model identification and the diagnosis of additive faults. It presents the phenomenon of gain shifting and shows how it affects the identification procedure. The chapter examines the residual errors which arise from shifted gains under various conditions. The effects of modeling error on the residuals will be reviewed, in the framework of the generic residual generator structure.
