Diagnosis of instrument fault

Abstract

The diagnosis of faults in instrumentation equipments can often be confused with faults in the system. The correct diagnosis of instrument faults is of importance.Here it is described how to detect instrument faults in non-linearity. Time-varying processes that include uncertainties such as modelling error, parameter ambiguity, and input and output noise. The design of state estimation filters with zero sensitivity to the uncertainties and maximum sensitivity to the instrument faults is described together with the conditions for the existence of such filters.The idea was applied to the fault diagnosis of a heat exchanger. The heat exchanger can be described by a bilinear model with modelling uncertainties. The cause of the fault is estimated from a fault dictionary which was compiled.

Probabilistic model for sensor fault detection and identification

“…Apart from the inadequacies that may be caused by the linear approximation of process model used in EKF, one Ž . needs a set of ''matched filters'' dedicated observers for each Ž sensor to perform fault isolation task Mehra and Peshon, 1971;Watanabe and Himmelblau, 1982;Watanabe et al, . 1994 .…”

Precision loss diagnosis about dynamic measurement system

“…From the analysis result, it can be known that this method is feasible for the research of the dynamic measuring system, and provide the theoretical basis for the further research of the practical system, which has significance in the design, maintenance and measurement of the system. [7,8] …”