Rammer compactor machines perform dynamic soil compaction. The complexity of this machine type makes design optimisation through traditional prototype testing impractical. This has pointed to the need for a theoretical model and simulation procedure for prediction of the dynamic behaviour of the machine and a procedure for optimisation as design parameters are changed during product development. In this paper a theoretical model of the rammer machine in combination with a soil model is described. This multi-body dynamics system is solved numerically. The system is non-linear and chaotic behaviour is possible. This parameter sensitivity emphasises the need for this kind of simulations in the product development process. A fairly regular behaviour is necessary for a predictable and safe operation. Parameter combinations giving too irregular behaviour are non-feasible. The energy transfer rate from the rammer machine into the soil is used as the objective function for optimisation. Multi-start Sequential Quadratic Programming for optimum search is used. To cover the design space well a Uniform Experimental Design is used for selection of starting points. This procedure proves to work well for the problem of this introductory study. The study shows a significant potential for improved compaction capacity although considering only the three design parameters that are most easily changed in practice. Approximately the same optimum is obtained both for operation on soft soil and hard soil, so a good all-round design seems possible. Including this theoretical support in the product development process should make it much more effective in finding optimum designs, also for other machines of similar type.

En inledande optimeringstudie av en jordkompakteringsmaskin har gjorts. Studien visar att potentialen för förbättringar är stor.