[home] [Personal Program] [Help]

---

Session: Session 6: Structural health / loads monitoring 
Session starts: Tuesday 27 June, 10:50
Presentation starts: 12:10
Room: Theatre café: parallel

---

Vesa Nieminen ()
Tomi Viitanen ()
Keijo Koski ()
Risto Laakso ()
Mikko Savolainen ()

Abstract:
Unmeasured quantities and responses at selected locations in structural monitoring can be estimated by different virtual sensing techniques. Virtual sensing enables estimation of operational response of the structure at any location based on a limited set of measurements and the numerical model of the structure. This paper presents the study for expansion of sparse response data by means of a component simulation model to enable fatigue analysis anywhere in the structural component or subassembly under question. The object of this case study is a high-performance F/A-18 Hornet fighter aircraft. It is extremely manoeuvrable thus capable of flying at high angles-of-attack. In those flight regimes the high-energy vortices generated by the inner wing leading edge extensions induce severe cyclic loading to the downstream structure by exciting the resonance frequencies of the empennage. The broad band dynamic loads, i.e. buffet loads, together with the static manoeuvring loads, contribute to the fatigue of the Vertical Tails (VT) of the aircraft. The Finnish Air Force (FINAF) has been running the Hornet Operational Loads Measurement (HOLM) program since 2006 to quantify the effects of operational usage on the structure of the F/A-18C/D aircraft. The HOLM program employs two intensively instrumented aircraft without operational restrictions. Both Vertical Tails of the HOLM aircraft include three strain sensors at the Vertical Tail Stubs and two acceleration transducers at the tip of the Vertical Tail, enabling to detect all the significant loading events locally in the VT. The virtual sensing applied in this study is based on bandpass filtering the measured acceleration data to get modal-specific responses for the primary VT modes. Component FE-model of the VT is then used to create modal-specific acceleration-to-strain conversion functions for the structural details of interest. For each mode individual virtual strain signal will be generated at the chosen structural location. All the signals of the primary modes at the same location are combined by the superposition principle to achieve the total virtual strain, which then can be used as an input in the typical fatigue analyses. Research focus of this study is applicability of the local/component FE-models for the virtual sensing. Research question is to investigate the limitations of the method when applying only a local/component FE-model and minimum number of sensors for the virtual sensing. The results demonstrated applicability of the local component FE-model and only a single acceleration measurement for virtual sensing in cases where dominant modes are well separated.