Abstract
An increasing number of structures including offshore jackets, wind turbines and pipelines, composed of axisymmetric section profiles, are approaching or have already exceeded their design service lives. Such structures are prone to damage from age-related issues such as corrosion and fatigue; therefore, condition monitoring is crucial to reveal their damage state and hence facilitate decision-making. For instance, proof that a late-life structure is undamaged may justify its continued usage; whereas identification of damage could trigger retrofitting or destruction in order to prevent failure or collapse.
This research commenced with an investigation of various traditional vibration-based damage indicators which were employed in identifying corrosion-like damage in pipes through a series of numerical and experimental test cases. It was confirmed, from both numerical and experimental studies, that traditional indicators, such as Frequency Shifts and Modal Assurance Criterion, are capable of Level I identification (detection). Whereas indicators such as Co-ordinate Modal Assurance Criterion, Modeshape Curvature and Modal Strain Energy could detect and locate damage (i.e., achieve Level II identification) in numerical models but were limited to Level I identification (detection only) experimentally owing to the relatively small difference between undamaged and damaged modal properties smeared by signal noise.
The above shortcomings prompted the search for a more robust damage indicator applicable to tubular structures which unravelled the phenomenon of local vibration mode pairs (LVMP). The LVMP occurrence, characteristics and properties are discussed in detail based on an experimental investigation of various types of corrosion-like damage profiles artificially induced on aluminium pipes. It was found that each real (valid) global vibration mode of a locally damaged tubular structure consists of two components – hereby called “active” and “passive”. The active component is sensitive to damage while the passive component is relatively insensitive to damage. Therefore, the passive component can be used to estimate baseline modal properties while the active component confirms damage existence. In other words, the existence of LVMPs within any real vibration mode of a structure is characteristic of damage.
The LVMP phenomenon was exploited in developing a novel method of level III identification (i.e., capable of detecting, locating and quantifying damage) in pipes. The accuracy of the proposed method was verified through a numerical study of 500 test cases with varying corrosion-like damage profiles. Damage was detected in all (i.e. 100%) of the investigated cases and was successfully located in 93% of the cases. Quantification of the located damage (in terms of material volume loss) was achieved to an accuracy of 92%. Furthermore, the method was successfully assessed using experimental modal data from an artificially corroded pipe. One of the main advantages of the proposed method over traditional vibration-based indicators is its non-reliance on baseline modal properties of the structure underscoring a novel concept in vibration-based damage identification for axisymmetric structures.