pdf download - Software and Computer Technology - TU Delft

1.2 Automated Fault Diagnosis Introduction
Remote monitoring improves the diagnostic performance by providing a great amount of useful
data. However, experts still have to interpret this data manually. Also knowledgable experts need
to devote considerable time and effort to extract information and draw conclusions about faulty
components. Automating this process is called automated fault diagnosis, and is discussed in the
next section.
1.2 Automated Fault Diagnosis
Automated approaches provide many potential benefits compared to manual approaches. Automated
fault diagnosis approaches automate the interpretation of raw data, in order to produce diagnoses.
Such approaches are likely to be faster, less error prone, and less dependent on human
intervention. Therefore, automating the fault diagnosis process can be one of the solutions towards
improving dependability. Figure 1.1 shows how a diagnostic system could add to dependability. By
processing data that is (remotely) monitored, the diagnostic system produces diagnoses and supports
the service engineer in the diagnosing task. Then, the service engineer could focus on repair
(the supervisory controller in the figure), and does not have to bother about the interpretation of
log data. If the supervisory controller is also automated, a diagnosis that has been made within
milliseconds could add to safety and reliability. The exact benefits depend on the specific approach
and implementation of the diagnostic system. The remote monitoring technique produces data that
contains information about faulty components. This is a practical environment to examine various
automated approaches to fault diagnosis.
The possible automated approaches can be characterized by black box and white box approaches.
A black box approach uses externally observable behavior, but does not state anything about the
structure of the system or behavior of internal components of the system. An example of a black
box approach is a data mining technique that searches for correlations between log data and faulty
components. It is unable to explain why a certain correlation exists. A white box approach uses
internally non-observable behavior as well as externally observable behavior. The behavior of the
whole system is defined by the structure and the behavior of internal components. An example of
a white box approach to fault diagnosis are experts that manually define a mapping between symptoms
and diagnoses. The mapping is implemented by application-specific code or, more generally,
by using expert systems.
Another characterization of approaches is whether the approach uses cause-to-effect or effectto-cause
reasoning. The mapping of symptoms on diagnoses, as implemented by expert systems,
is effect-to-cause reasoning. The next section introduces an approach that defines cause-to-effect
relations.
1.3 Model-Based Fault Diagnosis
Model-Based fault Diagnosis (MBD) is a technique for doing fault diagnosis based on a model of
a system. The model specifies all relevant information for doing fault diagnosis. A separate tool,
a so-called diagnostic engine, operates on this model to pinpoint root causes of failures. It is first
suggested by Reiter [25] and continued by de Kleer, Mackworth and Reiter [10].
The idea behind the technique is shown in Figure 1.2. The cloud represents the real system,
and its runtime operation. The model formally defines nominal, and possibly faulty behavior, of
the system. The diagnostic engine uses this formal model for computing predictions of system
behavior. During system operation, live data is gathered and processed in order to obtain actual
4