Thesis describes the analysis and establishing what is the cause of failure on existing Exhaust Gas Housing system in gas turbine. Based on current root cause analysis binary DSM of design parameters was created, from which eventually will be created numerical DSM. Tools for creating and analyzing DSM have been reviewed, then the most appropriate tool has been selected. Range of level numbers was defined accordingly to coverage of DSM matrix, this was carried because of easier way to add level numbers. Numerical DSM was created, the potential method (self ranking) analysis was carried out and the most influent parameters have been defined. At the end of thesis influent parameters have been highlighted in clustered numerical DSM, afterwards obtained results and conclusion have been interpreted. For the development of this thesis following tools have been used: Cambridge Advanced Modeller (CAM) in creating and clusters analyzing of numerical DSM, noncommerical tool for potential method analysis and MS Excel for interpreting results.