There are many different approaches to fault detection and isolation suitable for different kind of problems. In this thesis the properties and effectiveness of six different algorithms for isolation are investigated. To be able to compare the algorithms, performance measures are developed and implementations of the algorithms on a pre-defined benchmark system are done. The benchmark consists of a system, which is assembled by two tanks connected with various components. The components can be simulated faulty in different ways. The isolation algorithms that are used are classic methods for isolation from the automatic control field and the AI field as well as more recent approaches, such as Bayesian methods. Based on the results of the benchmark, recommendations are made for which type of isolation problems the respective algorithm suits best.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:kth-106252 |
| Date | January 2007 |
| Creators | Torelm, Martin |
| Publisher | KTH, Reglerteknik |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0014 seconds