Production-Induced Variance of Guided Wave-Based SHM Systems – A Case Study

To show the effectiveness and sensitivity of SHM systems, proof of concept-experiments have been state of the art for quite some time. For the detailed analysis of SHM systems, frequently numerical modeling is used. Its quality is shown with a comparison to experimental data. Deviations are often called to be caused by the influence of environmental and operational conditions as well as simplifications in the setup of the model.Another aspect, often neglected in guided wave-based SHM, are inaccuracies during production leading to deviations from the nominal structure and SHM system setup. An intelligent structure using active guided waves for SHM consists of the structural component, piezoelectric transducers, data acquisition and data analysis units. Production processes may lead to small deviations in the structure itself, deviations of the bonding process for the transducers as well as geometric inaccuracies, i.e., production-induced variance.For guided wave-based SHM, a case study is presented, which shows the effect of production-induced variance for a simple experimental setup consisting of an aluminum plate with rectangular piezoelectric transducers. The influence highly depends on the type of data acquisition and feature extraction. Transducer signals as well as laser Doppler vibrometer signals and different extracted features are evaluated. Using a very simple artificial change of the structure by adding an acrylic additional mass, it is shown that the effect of this change is approximately the same scale as production-induced variance. It is a significant influence, which needs to be taken into account when aiming for model-based sensitivity analysis.