pdf download - Software and Computer Technology - TU Delft
pdf download - Software and Computer Technology - TU Delft
pdf download - Software and Computer Technology - TU Delft
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Chapter 5<br />
Diagnosing the Beam Propeller<br />
Movement<br />
of the Frontal St<strong>and</strong><br />
The previous chapter presented the basic theory of model-based fault diagnosis. It already included<br />
an explanation of MBD on a real subsystem of the Philips Cardio-Vascular X-Ray System. This<br />
trivial example already shows benefits, but the main part of the Philips Cardio-Vascular X-Ray<br />
System has more complexity. Therefore, this chapter presents a case study of model-based diagnosis<br />
on a much more complex subsystem.<br />
In this case study, a MBD implementation for diagnosing failures of the beam propeller movement<br />
is developed. The LYDIA tool set is used as a diagnostic engine, <strong>and</strong> the LYDIA language is<br />
used for specifying the model. The first section of this chapter describes the experimental methodology<br />
<strong>and</strong> the approach to the case study. Section 5.2 describes the scope of the target system.<br />
The third section presents the fault diagnosis that experts are able to do using log data. Then, the<br />
three remaining sections deal with the model-based approach to fault diagnosis. Each of the sections<br />
presents a MBD implementation, <strong>and</strong> estimates its diagnostic performance in terms of entropy. Section<br />
5.4 presents a MBD implementation (MBD-1) that achieves equal entropy as experts are able to<br />
achieve. Section 5.5 presents a MBD implementation (MBD-2), that uses an improved model on the<br />
same system data, <strong>and</strong> calculates its entropy gain. The final section of this chapter presents which<br />
potential MBD implementation (MBD-3) could achieve the highest entropy gain, by considering<br />
various extra measurement points.<br />
5.1 Experimental Methodology<br />
Recall that the problem encountered in the work presented in this thesis is to improve fault diagnosis<br />
of Philips Cardio-Vascular X-Ray Systems with respect to higher dependability. This case study<br />
presents a proof-of-concept of the model-based approach to fault diagnosis, aimed at Philips Cardio-<br />
Vascular X-Ray Systems. The goal of this case study is to study the applicability of MBD on<br />
an example subsystem of the Philips Cardio-Vascular X-Ray System, namely the beam propeller<br />
movement of the frontal st<strong>and</strong>. The best way to do this is by using a metric for accuracy. Accuracy<br />
is generally considered as the most important criterion for diagnostic performance.<br />
MBD produces a list of diagnoses, in order of probability. If a supervisory user would like to<br />
use the result of MBD for its recovery actions, only one of the solutions can be accepted. The first<br />
49