Essais & Simulations 151

MESURES MÉTHODE

MESURES MÉTHODE Operational replication of strain responses during MIMO random control tests Les essais de contrôle des vibrations aléatoires sont réalisés pour vérifier qu’un système et tous ses sous-composants peuvent résister à un environnement de vibrations aléatoires pendant la durée de vie opérationnelle. Ces essais visent à reproduire avec précision, par l’intermédiaire d’une excitation contrôlée, la réponse structurelle en service d’un dispositif testé (DUT) dans l’axe principal de vibration. Introduction Random vibration control tests are performed to verify that a system and all its sub-components can withstand a random vibration environment during the operational life. These tests aim to accurately replicate via controlled shaker excitation the inservice structural response of a device under test (DUT) in the main axis of vibration and in all the possible axes where the levels exceed the acceptance thresholds. In the recent years great visibility has been given to the problem of accurately replicating in the laboratory the operational conditions that the DUT will eventually experience in-service. Multiple-Input Multiple-Output (MIMO) Random Control testing allows for a close replication of the nature of the operational loads. Previous work on aerodynamically excited structures has shown how increasing the number of control channels and trying to match the operational mechanical impedance, on top of a successful random test, also allows to closely match the response in locations that are not controlled. These observations are at the basis of the so-called IMMAT (Impedance-Matched Multi-Axial Test). In this context, it was shown that the environmental replication further improves by increasing the number of shakers and adopting rectangular control strategies. The ultimate goal of these studies is to converge towards a solution that will lead to a laboratory test that will eventually lead to damage mechanisms closer to the one that the component would experience in service. The objective of this paper is to continue the investigations to the physical quantities that can be actively controlled. The research question to answer with this work is: since the failure of the unit under test is directly related to the stresses and hence the strains, could MIMO Random Control techniques applied directly to strain measurements improve the replication of the operational strain field? BACKGROUND AND METHODOLOGYS To address this question, a set of experiments was carried out on the setup shown in Figure 1. Figure 1 (a) and (b) show the setup adopted for definition of the “pseudo-operational” conditions, highlighting the locations of the sensors for recording strain and acceleration responses. The inputs for the definition of the pseudo-operational conditions were white-pink noise uncorrelated voltages sent from a Siemens Simcenter SCADAS system to Siemens Simcenter Q-Sources inertial shakers attached to the frame of the DUT. Lumped calibrated masses were also attached to the lower side of the DUT allowing for simulating mass loading differences during the “pseudo-test” conditions. Figure 1 (c) shows the setup adopted for definition of the “pseudo-test” conditions, The test complexity was gradually increased introducing differences between “pseudo-operational” and “pseudo-test” conditions and changing control strategy. Figure 1 - Test setup adopted for the “pseudo-operational” conditions: 20 I ESSAIS & SIMULATIONS • N°151 • Novembre - Décembre 2022 - Janvier 2023

MESURES ANALYSIS The first tests were simple Random Control tests performed without changing the test setup (“pseudo-test” and “pseudo-operational” setup were physically the same), but simply changing the control sensors and controlled quantities – accelerations or strains. The results are shown in Figure 2. The results illustrated in this Figure confirm the findings of previous work on random vibration control: a multi-input operational environments can be replicated using 1 exciter in exactly one control location. Measure sensors respond according to the dynamic of the system and are subjected to the difference in excitation mechanism (single- versus multi- input) and boundary conditions. This can impact to a certain extent the fatigue failure mechanism of the DUT, given that the strain response will also be impacted. Controlling directly on a strain sensor allows however for the exact replication of a single strain response, which can be crucial, for example in case of a single critical point identified during the design of the DUT. Figure 2 – Random control results without changing the pseudooperational system. (a) Strain response PSDs; solid green, pseudo-operational responses; solid blue, control on acceleration (FuseBack:+Z channel); solid magenta, control on strain (StrainStabilizerLeft channel). (b) and (c) Random Control results for the control on acceleration and on strain, respectively. ESSAIS & SIMULATIONS • N°151 • Novembre - Décembre 2022 - Janvier 2023 I21