Applying parallel and distributed processing techniques to a tether dynamics simulation

Tether Dynamics has become an important area of space science and several accurate models which describe the behavior of tethered objects in low-earth orbit have been developed with their resulting simulations being validated in part with data collected from such unmanned flights as the SEDS-1 and SEDS-2. Safety concerns on manned flights, however, have demanded that these simulation models be expanded in scope and made more accurate with a much larger number of deployment scenarios being considered. This has resulted in the creation of simulations whose execution times on conventional single processor computing platforms are much too slow to adequately support the required safety analysis. The focus of this research is to demonstrate how parallel and distributed processing techniques can be used to accurately solve these types of problems in a timely manner. To accomplish this goal, a complex tethering dynamics simulation has been translated into a generic form and special partitioning and mapping software has been developed and used to generate a wide range of viable parallel representations for a ten processor Transputer System. The efficiency of these representations, as the number of processors and the number of tether nodal points is varied. is shown to support the feasibility of applying such techniques and software tools to expanded versions of this problem which in turn can be efficiently mapped for execution onto more advanced parallel computer configurations.