To calculate the damage distribution for a component, e.g. an axle, the time signals of the loads are required. In many cases these forces can only be measured with an enormous effort, while strains can be measured rather easily.

Based on measured time series of local strains at arbitrary spots of the axle the section forces of the component have to be identified. These can be used in a subsequent step to calculate the durability of the component and the critical points.

To solve this inverse problem, the method of »iterative learning control« is chosen. This method is usually applied on test rigs to determine time signals for the valve input in such a way, that predefined outputs (targets) are reached as close as possible. In the context here, the input signals are the unknown section forces and the output signals are the measured strains. The main advantage of the method is the possibility to use it for arbitrary systems or components.

The iteratively determined time signals of the section forces can be used to calculate the fatigue life of the whole component. The critical locations need not be known prior to the measurement but are identified in a subsequent step.