Reducing Delays for Unplanned Maintenance of Service Parts in MRO Workshops : A case study at an aerospace and defence company

Abstract: Service parts sometimes break down unexpectedly and require maintenance. The irregular nature of the need for this type of maintenance makes forecasting difficult and unreliable. Saab currently experiences problems with long delays when performing unplanned maintenance of service parts used in the two models of Gripen aircraft, Gripen C and Gripen D. These delays are source of monetary waste, as late delivery of maintained service parts results in Saab having to pay penalty fines to the customers. The purpose of this master thesis was to analyze data collected during a case study at Saab in Linköping, and suggest improvements for how to reduce these delays. This study focused on analyzing what caused the delays, and how the information provided by the customers can be used by the operative planners at the Maintenance, Repair \& Overhaul (MRO) workshops in order to be more efficient. The data was collected during the case study using semi-structured interviews of 16 people working with the current system, as well as by collecting historical data from an internal database at Saab. This data was analyzed in parallel with a literature study of relevant research related to service parts supply chains, MRO workshops, and unplanned maintenance operations. The analysis showed that there were four types of interruptions of maintenance; Internal stock-out of spare parts, internal stock-out of sub-units, external delays at the Original Equipment Manufacturer (OEM), and internal equipment breakdowns. A root cause analysis found that the four root causes of delays were: Saab does not have any contracts that incentivizes their OEM's to deliver on time. The  data from the technical report is not used to provide the operative planners with information about incoming orders. The MRO workshops do not have a standardized system for prioritizing maintenance of service parts. The MRO workshops currently lacks a method for predicting certain types of machine breakdowns.