Practical Approaches for Inverse Calculations of Drive Signals in a Virtual Test Rig with Regard to Agricultural Machines

Due to high demands in the farming industry agricultural machines become more and more complex. Therefore it is necessary to know exactly the dynamic loads, which affect the machine. The basis for strength calculations are outer excitations. In the case of agricultural machines it is either not possible or financially not affordable to measure forces and torques at the wheel hub, the point where loads are introduced into the structure. In addition, nonlinearities like complex tires, bearing slackness or elasticity of hydraulic components characterize such agricultural machines. Synthetic road profiles cannot be used because typi-cal farm tracks show stochastic trends. Hence, the load introduction is complex and stochas-tic as well. In contrast to the input variables, interior parameters of a machine can be measured without problems by using standard techniques. Such measurement data of output functions are defined as target signals. Usually accelerations, strains or forces are used for the physical description of these targets. For that reason it is of great interest to find methods to calculate input signals in an inverse way. Basis of the inverse calculation are the model parameters and the measurements (targets). In this paper different approaches are consid-ered. The first approach is the method of virtual iteration, which is a well known method in the automotive industry. The model is linearised at the specific state and the transfer matrix of the system is calculated. By using the inverse of the transfer matrix and the vector of target signals, the vector of inputs (drives) can be calculated. Due to nonlinearities in the system this procedure has to be repeated in several iterations. Another mathematical method is an optimal control approach. By using optimization methods, such input variables are calculated which bring the simulation outputs and the targets at highest possible accordance. The opti-mization of the nonlinear system is carried out for individual time intervals. Efficiency, com-putational effort and practical convenience of both methods are compared in this paper.