Turning point based fatigue testing: Combining multisines with turning point replication

Abstract Time waveform replication (TWR) is commonly used as control strategy to impose a given force (or displacement) target during fatigue testing. TWR belongs to the class of (linear) Iterative Learning Control strategies, which in general works well for systems with an approximately linear behavior. Larger errors occur when the nonlinear effects are more pronounced. In this contribution the TWR approach will be revisited taking into account the fact that in many applications it is not necessary to exactly replicate the target signal. Only the so-called “turning points” of the force signal need to be correctly controlled. Indeed, the controlled force signal should have the same turning points as the target force signal; they do not have to be exactly equal to each other for all time samples. This relaxation of the problem formulation allows generating a controlled force signal that better converges to the turning points of the target force signal. To do so, an easy-to-implement algorithm will be proposed in this contribution based on the classical TWR algorithm. The second part of this contribution consists of the introduction of the application of multisines in fatigue testing to better use the bandwidth of the actuator. The multisines will be combined with the proposed algorithm and compared to the classical TWR algorithm by means of experimental fatigue tests performed on the airplane component.