Studies on designing of submarine optical fiber cable

Subthreshold flaws and their failure prediction in long-distance optical fiber cables

“…This approach cannot specifically insure against hostile environments, particularly hot or acidic aqueous contact which may promote premature "pop-in" of cracks. Design philosophies which can minimize this threat are to: 1) aim for [20][21][22][23][24][25][26][27][28][29][30] year targets in all applications, 2) use only the conservative n = 19 for subthreshold decay for the whole period, 3) simulate where possible the worst-case environment in conducting prooftests, extending the duration of the test where possible to promote crack formation.…”

“…It is shown in Appendix A that the probability of a failure F, in a length L , having survived a proof stress U,,, is F, = I -exp (Lu,"(uj; -(s/t,,)''l'")) (4) where S = E ( u " t ) for all applied stresses U for time t experienced by the length L, including the prooftest, and the fatigue constant n is derived empirically. This result is equivalent to that of [21], but is here derived with minimal dependence on fatigue theory, making use instead of empirical data and the equivalence of the fatigue processes in the field and in the prooftest. It is worth noting similarities between this approach and that employed for many other components, particularly optoelectronic ones, used in long haul systems [23].…”

“…Broadly, the methods may be classified into two categories: a) those which interpret the Griffith flaw treatment as being theoretically accurate, and attempt to predict conditions for survival with absolute confidence, based on numerical values for the constants A , Y, etc. (e.g., [19]), and b j those based additionally on empirical failure data and reliability theory, which are less sensitive to the numerical accuracy of the initial assumptions, e.g., [20], [21]. It should be stated at the outset that neither method is now consistent with the physical behavior described above, but in the absence of a fundamental theory of subthreshold flaw propagation and subsequent pop-in, it will be shown that method b) at least offers a plausible model for assessing fiber reliability in cables, and for the moment at least is the best one available.…”

“…The scale and slope parameters uo and m can be estimated from manufacturing yields at at least two proof stresses [21]. These values are needed to make the prediction.…”

“…The application of (4) to such cables has been reported by Miyajima et ul. [21], 1241, and we propose to add here some illustrative figures and a possible alternative design approach.…”