Accurate Component Load Prediction with Hybrid Test – FEA Approach.

Knowing accurate loads acting on the structure is still a main pain area for analyst. Accurate fatigue analysis through FEA is still a difficult task because of uncertainties associated with component load profile. Even for known duty cycle, approximation involved in measurements and load cascading techniques (MBD simulation models) put limits on accuracy of component load prediction. Small error in loads can lead to very large error in fatigue life calculations. Various researches proposed various innovative alternate means for accurate component load prediction and published in literature over past few decades. Author has explored and developed hybrid Test-FEA approach to estimate component loads based on robust strain-load transformation and pseudo inverse technique. This approach is based on effective use of strain transformations along with accurately measured strain data to calculate accurate load histories. The accuracy of load estimates depends upon number of factors such as strain measurement locations, orientations on component and number of strain gauges used. An optimization algorithm on MATLAB is created for finding apt strain gauge locations and orientations based on finite element results that leads to robust and stable transformation matrix. The key steps involved in the process are – 1. Identify interface points and associated DOFs for the component under consideration. Perform FE stress analysis as if each DOF experience external load in isolation. 2. Choose appropriate locations and orientations for strain gauge placement to build robust strain-load transformation matrix, using in-house algorithm. 3. Instrument the component with strain gauges and measure real life strain histories as per defined duty cycle. 4. Convert measured strain histories in to load histories using transformation matrix generated in step 2. 5. Perform detailed fatigue analysis of entire component using back calculated loads and FE analysis. A numerical as well as Experimental validation of the proposed method is presented through various case studies.