Dual Scale Fatigue Crack Monitoring Scheme Considering Random Material, Geometric and Load Characteristics

Abstract

Effectiveness of the existing fracture control methodology depends on the non-destructive evaluation of defects arising from repeated service of structures. Damage detection is by far being automatic since not all defects are dangerous. The correct diagnostics require not only a knowledge of the size of the prevailing defects but also their likely locations of occurrence from the initial design and/or past experience. What is not known is the time when the defects will become critical since the service conditions change as a rule and the material age in time. These inherent variables can affect the accuracy of non-destructive evaluation which in its strict sense implies certain cut-off scale of small defects that are assumed to be not harmful. Material microstructural effects have been known to affect the structural integrity. The obvious implication is multiscaling, the least consideration of which involves the dual scale of micro/macro. Hence, defect monitoring presents over whelming difficulties because of the entanglement of multi-variables.

Starting with the basic quantities of the local force and displacement that can be detected from the commercially available transducers, continuous records can be made available for the local compliance C and its time rate history, say dC/dt or dC/dN in fatigue with N being the number of cycles. These data can be stored in microprocessors and made available for analyses. The conversion of the force and displacement records to damage by fatigue crack growth is the challenge of this work. Appeal is made to a dual scale micro/macro fatigue crack growth model that has the capability to delineate micro- and macro-cracking. The model makes use of three parameters ¼^*, d^* and σ^*. They account for the interaction between the micro/macro effects in terms of the, respective, relative shear modulus ¼_micro¼_macro, the micro-tip characteristic length d/d_o and the crack surface tightness ratio σ_o/σ_∞ that controls the opening of the fatigue crack. These parameters are assumed to vary in a random fashion as the crack size increase with the repetition of loading. Random normal distribution of statistical values of μ^*, d^* and σ^* is investigated for different mean <μ> and standard deviation <σ>. The amount of dispersion around the mean is examined for the crack length a and crack growth rate da/dN in addition to the compliance C and its rate dC/dN. It is demonstrated that randomness of the material microstructure has an effect on the local compliance which will increase with crack growth in a stable fashion and then rises more quickly with the number of cycles. The characteristic is reminiscent of the compliance to crack length relation used for determining the energy release rate in fracture mechanics.

The ability to account for micro/macro damage transition is considered to be fundamental for assessing the meaning of the compliance data. Randomness of the material microstructure parameters provides the means for studying the dispersion of the results in terms of fatigue crack growth. The preliminary findings are encouraging and suggest several new directions of research for health monitoring.