Unmeasured strain estimation on bottom-fixed offshore structures by multimetric data fusion and kalman filtering with data buffer

Offshore structures for tidal current energy converters and offshore wind turbines are always exposed to the much harsher environment with strong tidal current and wind loadings. While monitoring structural soundness and integrity is considered to be crucial to prevent catastrophic collapses and prolong lifetime, it is intrinsically challenging due to the difficulties in accessing to the critical structural members located under water for installing and repairing monitoring sensors and data acquisition systems. Virtual sensing technologies have the potential to alleviate such difficulties by estimating unmeasured structural responses at desired locations with other measured responses. This approach is particularly advantageous when (1) some sensors are malfunctioning and (2) sensor installation at critical members is difficult. Despite the usefulness of the virtual sensing, its performance and applicability for structural health monitoring of the offshore structures under non-stationary excitations has not been fully studied to date. This paper investigates the use of the virtual sensing for structural health monitoring of offshore structures for marine energy facilities rather for oil and gas exploitation. The Kalman filter is introduced for data fusion of different types of measured responses to produce estimated responses at structural members of interest. This study discusses how the non-stationary tidal current loading in the real-time response estimation is appropriately handled. Numerical analysis and laboratory experiment are conducted to verify the performance of the virtual sensing strategy using a bottom-fixed offshore structure model, showing that the unmeasured responses can be reasonably recovered from the measured responses.