Probabilistic Performance Analysis of Fault Diagnosis Schemes
Probabilistic Performance Analysis of Fault Diagnosis Schemes
Probabilistic Performance Analysis of Fault Diagnosis Schemes
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
malfunction — an intermittent irregularity in the fulfilment <strong>of</strong> a system’s desired<br />
function.<br />
disturbance — an unknown (and uncontrolled) input acting on a system.<br />
residual — a fault indicator, based on a deviation between measurements and<br />
model-equation-based computations.<br />
fault detection — determination <strong>of</strong> the faults present in a system and the time<br />
<strong>of</strong> detection.<br />
fault isolation — determination <strong>of</strong> the kind, location and time <strong>of</strong> detection <strong>of</strong> a<br />
fault. Follows fault detection.<br />
fault identification — determination <strong>of</strong> the size and time-variant behaviour <strong>of</strong> a<br />
fault. Follows fault isolation.<br />
fault diagnosis — determination <strong>of</strong> the kind, size, location and time <strong>of</strong> detection<br />
<strong>of</strong> a fault. Follows fault detection. Includes fault identification.<br />
reliability — ability <strong>of</strong> a system to perform a required function under stated<br />
conditions, within a given scope, during a given period <strong>of</strong> time.<br />
safety — ability <strong>of</strong> a system not to cause danger to persons or equipment or the<br />
environment.<br />
availability — probability that a system or equipment will operate satisfactorily<br />
and effectively at any point <strong>of</strong> time.<br />
2.4.2 Brief Survey <strong>of</strong> <strong>Fault</strong> <strong>Diagnosis</strong><br />
In this section, we present a brief survey <strong>of</strong> the vast field <strong>of</strong> fault diagnosis. For a thorough<br />
treatment, see Chen and Patton [9], Ding [24], or Isermann [48]. Consider the general fault<br />
diagnosis problem in Figure 2.2. The system G is affected by known inputs u, stochastic<br />
noises v, unknown deterministic disturbances w, and an exogenous signal f representing<br />
a fault. The fault diagnosis scheme is comprised <strong>of</strong> two parts: a residual generator F and<br />
a decision function δ. The residual generator F uses the known input u and the measured<br />
output y to produce a residual r , which carries information about the occurrence <strong>of</strong> faults.<br />
The decision function δ evaluates the residual and determines what type <strong>of</strong> fault, if any, has<br />
occurred. The output <strong>of</strong> the residual generator, d, is called the decision issued by the fdi<br />
scheme. Typically, d takes values in some finite set <strong>of</strong> decisions D. This separation <strong>of</strong> a fault<br />
diagnosis scheme into two stages was first proposed in [13].<br />
There are a number <strong>of</strong> approaches to constructing meaningful residual signals. In a<br />
structured residual set, the residual r is a vector such that each component r i is sensitive to<br />
a subset <strong>of</strong> faults. If each residual component r i is sensitive to a single component f i <strong>of</strong> the<br />
fault vector, then r is said to be a dedicated residual set. Another approach is to make each<br />
14