PSI - Issue 38

Amaury Chabod et al. / Procedia Structural Integrity 38 (2022) 382–392 Amaury CHABOD / Structural Integrity Procedia 00 (2021) 000 – 000

388

7

Strain gage positions are optimized, as described in previous sections, seeking modal stresses as independent load cases. We can visualize the strain gage positions and orientations on the FE Model.

a)

b)

Clamping

Fig. 5. (a) First mode shape, (b) First mode shape stresses and position-optimized strain gages. Ten strain gages position are given, which is sufficient to represent ten modal stresses independent load cases. The result of the optimization process gives a large determinant value of 2.43e30 for [ ] [ ] strain gage matrix. The determinant being far from zero, inverting the matrix needed in equation (3) for load reconstruction presents no numerical difficulties. The mechanical behavior of this turbine blade is close to a clamped-free beam, where we can observe an alternated bending movement in the first mode shape (Fig. 5). 5.2. Load Reconstruction Prior to really measuring service loads, it is advisable to make FE/test correlation for validation of FE boundaries, and of the various assumptions, in particular the linear behavior. Global structure should behave fully elastically, and any local plasticity must be avoided. A Virtual Strain Gage Process may be done with known input loads. Furthermore, the dynamic behavior must be assessed with Experimental Modal Analysis to check FE resonances and mode shapes representativity. Once the FE/test correlation is achieved, instrumentation is released with previous requirements on strain gage position. Ten channels are measured during 600 seconds from the 10 strain gages defined previously, at 50 Hz sample rate. It will constitute our only information on the loads applied to the structure. This operation could be easily generalized to a set of different events, weighted with number of repeats, in a duty cycle representing the full component mission profile.

UNRESTRICTED