A novel method proposed detects and identifies faulty sensors in dynamic sy
stems using a subspace identification model. A consistent estimate of this
subspace model was obtained from noisy input and output measurements by usi
ng errors-in-variables subspace identification algorithms. A parity vector
was generated, which was decoupled from the system state, leading to a mode
l residual for fault detection. An exponentially weighted moving average (E
WMA) filter was applied to the residual to reduce false alarms due to noise
. To identify, faulty sensors, a dynamic structured residual approach with
maximized sensitivity is proposed which generates a set of structured resid
uals, each decoupled from one subset of faults but most sensitive to others
. All the structured residuals ale also subject to an EWMA filtering to red
uce the noise effect. Confidence limits for filtered structrued residuals w
ere determined using statistical inferential techniques. Other indices like
generalized likelihood ratio and cumulative variance were compared to iden
tify different types of faulty sensors. The fault magnitude was then estima
ted based on the model and faulty data. Data from a simulated 4 x 4 process
and an industrial waste-water reactor were used to test the effectiveness
of this method, where four types of sensor faults, including bias, precisio
n degradation, drift, and complete failure were tested.