Thesis for the Degree of Master of Science


Year 1999
Tomi Salmi

Fault diagnosis system for copper smelting process

The aim of this thesis was to develop a fault diagnosis system for the copper smelting process in Outokumpu’s Harjavalta plant.

Automatic fault diagnosis systems for processes and process equipment have been studied very actively during recent years. Decrease of fault situations and ease of maintenance are the main reasons for the use of a fault diagnosis system. With expert systems using heuristic and analytic information, faults can be detected on an early stage.

In the literature part, the structures of fault diagnosis systems and the use of self-organizing maps in fault diagnosis were studied. Also industrial applications of the fault diagnosis systems were considered.

In the experimental part Outokumpu’s copper smelting process was studied first. The structure and implementation of the fault diagnosis system were described next. A prototype was designed and programmed for testing the structure of the planned fault diagnosis system.

Using data collected from the copper smelting process, self-organizing maps were trained to classify different states in the sub-systems. The same data were also used for testing the system. In testing, the system responded correctly to faults and normal conditions.

Self-organizing maps used in the test system succeed in classifying the faults in the copper smelting process. Maps are deployed as part of the decision module, which will be implemented later.

This info last modified 24 Sep 2017 - contact webmaster