Issue 78

C. Vendittozzi et alii, Fracture and Structural Integrity, 78 (2026) 519-540; DOI: 10.3221/IGF-ESIS.78.30

operational sources, damage states, or landing-severity classes. Per-event power spectra complement the time-resolved spectrograms by providing compact event-level spectral summaries that facilitate comparison across Events 1–3 and between the two representative datasets. Differences in peak prominence across events are expected and are consistent with variability in excitation intensity, interaction duration, local strain transfer, and redistribution of dominant spectral content within the same event-centered interval. In this sense, the periodograms do not replace ridge-based tracking and should not be interpreted as standalone event “fingerprints” for classification. Rather, they provide compact frequency-domain summaries that help compare the spectral organization of FBG-derived event-centered windows. From an application perspective, the recurrence of comparable frequency families and their compact representation through ridge trajectories, band-limited descriptors, and per-event spectra are relevant to future digital-engineering workflows because they provide traceable, reduced-order descriptors of field-acquired structural response. However, in the present study these descriptors remain response-based indicators and are not yet decision-oriented HUMS or structural-integrity metrics. Their use for maintenance screening, hard-landing assessment, damage detection, or health-state evaluation would require synchronized operational references, larger repeated datasets, temperature compensation, and validation against independent measurements or models. With a 1 kHz acquisition rate and the selected window/hop configuration, descriptor updates occur at approximately 2 Hz. This update rate is compatible with future near-real-time implementations, but the present work should be interpreted as a field demonstration of descriptor extraction rather than as a validated real-time HUMS implementation. Limitations and next steps The present study was intentionally restricted to FBG-derived measurements acquired under realistic installation and operational constraints, without synchronized avionics or other onboard reference systems. As a result, the interpretation remains response-driven and does not permit unambiguous attribution of each event candidate to a specific flight or ground maneuver. The FBG-derived events should therefore be considered high-energy response events, not independently validated touchdown instants, braking phases, hard-landing events, or operational-state labels. The installation was also subject to practical airworthiness-compatible constraints, including external routing, limited strain-relief options, and the need to avoid interference with flight-critical or moving assemblies. No dedicated strain-transfer calibration was performed for each bonded grating; consequently, absolute strain amplitudes should be interpreted as local sensor responses, while the analysis mainly emphasizes relative response evolution and spectral descriptors. Temperature represents an additional limitation. No independent temperature channel or thermally isolated reference grating was available during the campaign. Detrending and high-pass filtering reduce slow drift, but they do not replace physical temperature compensation; therefore, quasi-static and long-duration strain trends were not interpreted quantitatively. Future work should include synchronized operational references, such as weight-on-wheels signals, airspeed, vertical speed, braking command, accelerometers, video, or flight-log markers. These measurements would allow event windows to be associated more rigorously with specific operational phases and would support validation of the FBG-derived descriptors. Additional priorities include explicit temperature compensation, extension to a larger set of labeled flights and repeated maneuvers, and comparison with conventional sensors, modal data, or numerical models. Only after this validation step would it be appropriate to define decision-oriented indicators for landing severity, asymmetry, usage monitoring, or maintenance screening. Overall, the present work should be regarded as a field-oriented feasibility study showing that imperfect operational FBG data can be screened, converted, and reduced to interpretable event-level time–frequency descriptors. It does not yet provide validated damage detection, hard-landing classification, health-state assessment, or structural-integrity decision support. processing workflow was presented to extract monitoring-oriented time–frequency descriptors from field-acquired FBG measurements on a helicopter landing-gear structure, including recordings affected by dropouts and saturation artifacts in the raw Bragg-wavelength histories. The workflow combines integrity screening and short-gap handling, baseline normalization and wavelength-to-microstrain conversion, detrending and high-pass filtering to reduce slow drift, and time–frequency analysis performed on FBG-derived event-centered windows identified from a short-time multichannel strain-energy proxy. Two representative datasets were processed using the same procedure. In both cases, three dominant high-energy event candidates were detected, each associated with increased oscillation amplitude, inter-channel divergence, and distinct time–frequency signatures. In Dataset A, the post-event response reconverges after each excitation and resumes the underlying trend observed during non-event intervals. Dataset B exhibits a different post-event evolution, with more persistent activity after the final event candidate. These differences show that the workflow can capture both event-centered A C ONCLUSIONS

538