The overall fuel consumption and the associated carbon dioxide emission of mobile machinery, including construction machines, agricultural machines, forestry machinery and municipal utility vehicles is comparable to that of passenger cars, even though the latter outnumber the former. This is due to the process-related very high engine power and the partly very poor efficiency of integrated hydraulic drive systems.


At the moment, the optimization of the efficiency of hydraulic drive systems is achieved by using real prototypes. However, this method is time consuming as well as expensive. In addition to that, it requires the inclusion of all components as well as an elaborate test environment in advance. In a test rig, it is hardly possible to reproduce the complex properties of actual machines. At the same time, we are dealing with increasing complexity due to the growing influence of electronics. The requirements are even higher if the test has to consider an interaction with an operator. As a consequence, the development of virtual prototypes increasingly replaces experiments on actual machines via model-supported analyses. For example, the Bosch-Rexroth company computerized the development of new hydraulic components for work machines. However, solving the emerging optimization problems involves numerous simulations and the associated processing load is so extensive, it exceeds the capacities of standard computers.


A quantitative disadvantage common to all recent approaches, is the computational cost conditioned by the high number of equations. The computational cost for lumped element Multi Body Systems amounts to several hours, for mesh based FEM approaches it can add up to days. So far, there is no parallelization concept with promising potential for optimization of the computational efficiency.


On top of that, in order to achieve an optimized calculation of the operational performance of working machinery (total machine management) it is crucial to factor in the so far unknown load cases and driving situation. Consequently, the computational costs increase enormously. The current technology does not measure up to the demands of corresponding simulations essential to the development cycle. Nonetheless, if we succeeded in optimizing whole machine systems with regard to their drive technology down to each single component and its functionality within the process, the above-mentioned challenges could be managed, e.g. a significant enhancement of energy efficiency. To achieve that, we would have to be able to obtain information from virtual test facilities comparable to those derived from real test rigs.


This puts high requirements on machine simulations with respect to the size and complexity of models, the level of detail of multi-physical sub models as well as the calculation stability and speed.