Optimizing Launch‐on‐Time Probability

Future space projects such as space stations, large unmanned space structures, interplanetary spacecraft, and lunar and Mars bases will require highly reliable launch systems for their construction and logistic support. It may also be necessary to provide “guaranteed” commercial launches. These needs will place increased emphasis on mastering every aspect of launch‐vehicle operations, particularly a launch‐on‐time capability. Causes for delay during launch, i.e., unplanned “holds,” are attributable to several sources: weather, range activities, vehicle conditions, human performance, and so forth. In this paper, actual launch data on unscheduled hold distributions of various launch vehicles are presented. The data were supplied by industrial associate companies of the Center for Space Construction of the University of Colorado, Boulder. Probability models that can describe these historical data and be used for several purposes are determined: (1) As inputs to broader simulations of launch‐vehicle logistic space construction support processes; (2) to determine which launch operations sources cause the majority of the unscheduled holds; and (3) to suggest changes that might improve launch on time. The paper demonstrates the ability of a compound distribution probability model to fit the actual data. This model is then manipulated to develop a method for optimally improving launch‐on‐time probability.