theses_web.jpg

Thesis for the Degree of Master of Science

 

Year 2015
Lanre Balogun

Visualization of fault propagation paths in process monitoring systems


Modern industrial processes typically have complex interconnections among process equipment and units. A fault can easily propagate far beyond its origin, making it difficult to locate the root cause and to identify its propagation paths. Thus, the latest approach for identifying fault propagation paths uses qualitative models, e.g., topology-based models, for improving the results acquired from the data-driven methods of causal analysis. However, the current approach includes no display of the fault propagation paths in process monitoring systems, which could enable the visualisation of the spread of fault effects.

In this thesis, the aim is to develop and to test a new technique which enables a process monitoring system to display fault propagation paths, based on the causality matrix obtained from a process. A causality matrix is created using the Granger causality method, which contains the details of the fault propagation path. Subsequently, the causality matrix is refined using a connectivity matrix, which was obtained from a piping and instrumentation diagram (P&ID). Furthermore, the process monitoring system and the refined causality matrix are linked together, to transfer the fault propagation paths details to the process monitoring interface. A number of algorithms in the automation system are developed and implemented to accomplish the link.

A flotation pilot process is the case study in this thesis, used for testing this technique. The experiment data acquired from the flotation pilot was analysed to generate a causality matrix, which is subsequently refined and stored in the database of the automation system. Hence, the fault propagation paths is visualised in the process monitoring system. Finally, this procedure confirmed that the propagation of the effects of a fault is better understood if the propagation path is visualised in the process monitoring system.

Thesis electronical version can be downloaded from here


This info last modified 22 Nov 2017 by Jukka Kortela