Fatigue and Reliability Assessment Incorporating Computer Strain Gage Network Data

This paper details a procedure by which reliable engineering components may be designed and produced to withstand specified fatigue loading situations. The procedure is a modification of the current aircraft industry's damage tolerance approach as outlined, among others, by Goranson [1]. For the design and manufacture of a specific component, the procedure incorporates a knowledge or specification of: (1) a typical fatigue loading history; (2) the determination of the stress time history at a location or locations of most concern; (3) typical fatigue crack geometries; (4) the mechanical properties, including the fatigue crack growth and closure characteristics, of the material from which the component is or will be manufactured; and, (5)in stochastic process terms, a time or cycle dependent description of the inherent statistical scatter that always accompanies fatigue crack growth.

While the procedure is quite general, its applicability is illustrated by an application to the improved design of mountain bicycle frames and components. Specifically and as an equivalent to standard aircraft flight-load histories such as TWIST, Mini-TWIST or FALSTAFF, the projected use of the results generated by the specialized data acquisition system whose development is described in a companion paper [2], is illustrated by estimating the fatigue crack growth characteristics and reliability of a mountain bicycle crank-arm. The procedure utilizes the loading history generated by “smart” strain gages situated on a mountain bicycle while the bicycle and rider are traversing a demanding mountainous trail. The acquisition system samples and captures strain data at 1 kHz with 12-bit resolution; performs peak detection and averaging calculations;transfers via digital radio the data to a Windows-based PC station at the trailhead; and analyzes these data both to determine stress profiles and to develop typical fatigue loading time-histories.Using this information, crack growth rate estimates based upon these load spectra, crack geometries typical of those found in mountain bicycle crank-arms, the crack closure concept, and the da/dN versus ΔK_eff for A17075-T6 may be obtained. Coupling this information with a stochastic process interpretation of the scatter in these crack growth estimates leads to a meaningful description of component reliability.