Depot-Level Simulation and Multivariate Analysis on B-1 High Velocity Maintenance

The objective of this thesis is to gain insights on the B-1B depot maintenance operations, with a focus on direct maintenance hours or burn rate, under the implementation of High Velocity Maintenance (HVM). Based on historical depot maintenance data and the current B-1 depot HVM prototype data, a discrete-event simulation model is developed using Arena 12.0. Some United States Air Force supply chain influences, such as manning levels and kitting characteristics of the B-1 depot operations, are incorporated in our models as design factors. The model captures the stochastic nature of 30 HVM tasks performed on one B-1 aircraft in a representative HVM cycle at the B-1 depot located in Oklahoma City Air Logistics Center, Tinker Air Force Base. To examine the impact of HVM, we vary the levels of the design factors and conduct a design of experiment (DOE). The DOE analysis reveals that manning levels and kitting characteristics have statistically significant impact on some HVM task completion times, which are used collectively as a surrogate measure for burn rate. In particular, manning schedule with a centric focus on direct maintenance, high task kit availability, and small kit deficiency produce the highest burn rate. Additionally, by performing multivariate analysis, we are able to reduce the dimensionality of the output statistics and conclude that kit deficiency is the main driver for HVM task duration with our simulation.