Frequency domain load reconstruction noise sensitivity investigation using a cantilever beam test model and experimental results

This dataset was obtained by simulation and experimental testing on a simple cantilever beam. Firstly, a modal test was simulated and performed, the results were used to estimate the frequency response function of the beam system. After the FRF was obtained, a number of known forces were applied to the beam and the resulting accelerations at fixed points recorded. The accelerations due to the applied loads were then used with the FRFs obtained in the first step to reconstruct the applied loads.

The reconstructed loads were then compared to the applied (known) loads in order to evaluate the accuracy of 5 different load reconstruction methods. These methods are:

1. Inverse
2. Least Squares
3. Singular Value Decomposition
4. Tikhonov with L-Curve Parameter Optimisation
5. Tikhonov with GCV Parameter optimisation.

Once the baseline accuracies were established, noise was introduced artificially to the response signals from 0.5% noise up to 10% noise and the loads were reconstructed at each noise level to determine the accuracy of each method under noisy conditions. Finite element beam of a basic cantilever beam used to in research to the effects of noise on frequency domain load reconstruction.