The Bree problem with the primary load cycling out-of-phase with the secondary load

https://doi.org/10.1016/j.ijpvp.2017.06.004Get rights and content

Highlights

  • The Bree problem is analysed with both primary and secondary loads cycling out of phase.

  • Ratchet and shakedown boundaries are derived analytically.

  • Algebraic expressions are derived for the ratchet strains and cyclic plastic strains.

  • Ratcheting occurs more readily when the loads cycle out-of-phase than if they cycle exactly in-phase.

Abstract

Ratchet and shakedown boundaries are derived analytically for the Bree problem with both primary and secondary loads cycling, but out of phase. Ratchet strains and cyclic plastic strains are derived in terms of the X,Y position on the ratchet diagram. The ratchet boundary lies between the Bree ratchet boundary and that for both loads cycling in-phase (or in exact anti-phase), but generally lies closer to the former. This provides a warning for ratcheting analyses with two loads cycling: employing the simplifying assumption of cycling exactly in-phase (or exactly in anti-phase) may be substantially non-conservative.

Introduction

In this, the 50th anniversary year of Bree's classic analysis of ratcheting [1], it is particularly appropriate to return to the algebraic method of solving ratcheting problems. Analytical solutions of ratcheting problems are relatively rare, being confined to cases of sufficient geometrical and loading simplicity. In real engineering applications, numerical methods, such as the linear matching method [2], [3], are far more versatile and are now becoming mature techniques. Nevertheless, complete analytical solutions remain of considerable value, partly because they provide validation cases for numerical methods and partly because, in the few cases where they are tractable, they provide greater physical insight into the problem. This is exemplified by the Bree analysis, Refs. [1], [4], which is still the chief landmark in this field even after half a century. A further advantage of a complete analytical solution is that all load magnitudes are addressed, identifying the behaviour at all points on the X,Y ratchet-shakedown diagram.

The Bree problem addresses uniaxial ratcheting under a constant primary membrane load plus a secondary wall-bending load which cycles between zero and some maximum value. Whilst the application that Bree had in mind in his original analysis was fast reactor fuel clad, and hence a cylindrical geometry, he analysed the problem as if for uniaxial stressing. Consequently the problem may be considered as relating to a beam of rectangular section.

A modified Bree problem consists of considering the primary membrane load to also cycle. If the primary and secondary loads cycle precisely in-phase (Fig. 1a), or if they cycle precisely in anti-phase (Fig. 1b), the complete solution to the resulting ratcheting and shakedown problem has been published, [5], [6], [7]. The in-phase and anti-phase cases have an identical ratchet boundary, but this lies at substantially greater loads than for the original Bree problem, i.e., in-phase or anti-phase cycling of the primary load is a significantly more benign loading as regards ratcheting than if the primary load were held constant.

However, whilst realistic plant operation does frequently involve both the primary and secondary loads cycling, the occurrence of precisely in-phase, or precisely anti-phase, cycling is improbable. This raises the question as to where the ratchet boundary lies when the primary and secondary loads cycle with some intermediate phase. There are published analyses in which the primary load cycles as well as the secondary load, with various phase relations, for example [8], [9], [10]. However, none have presented the complete solution for the out-of-phase case addressed here.

Section snippets

Definition of the problem addressed

The problem may be considered as relating to a beam of rectangular section subject to uniaxial stressing: a primary membrane load and a secondary bending load. Both loads vary between zero and some maximum, their temporal variation being 'square waves' with the same period but not in phase, as depicted by Fig. 2. Note that because a square-wave profile is assumed, each load is either acting fully or not acting at all. This simplifying assumption makes the problem analytically tractable.

The

Formulation of the problem and solution method

All dimensions are normalised by the section thickness, t. (This may differ from the normalisation convention used elsewhere, e.g. [7], normalised using t/2). All stresses are normalised by the yield stress, σy, and all strains by the yield strain, εy=σy/E. The applied (elastic) primary membrane stress, after normalisation, is X (hence X=1 is plastic collapse). Similarly the peak (elastic) thermal bending stress after normalisation is Y. The through-thickness dimensionless coordinate x, is zero

Ratchet solutions

Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 10a are schematic diagrams of possible stress distributions. Each Figure shows the stress distributions for the first five quarter-cycles. The stress distribution for the fifth quarter cycle differs from that of the first quarter cycle despite corresponding to the same loading. Thereafter the distributions repeat, i.e., the sixth cycle has the same stress distribution as the second cycle, the seventh cycle has the same stress

Distributions of Fig. 11

For Fig. 11 to be valid it is required that a<0.5 which, using the parameter solutions of Table 9, implies Y>1X. This is merely the condition that yielding does occur when both loads act, and establishes that the region in question lies above the 'elastic line', UE in Fig. 15.

For Fig. 11 to be valid it is also required that σ8<1 which implies that Y<4(1X), i.e., that the region in question lies below line BE on Fig. 15.

Finally, for Fig. 11 to be valid it is also required that σ9>1 which

Ratchet-shakedown diagram

The resulting ratchet-shakedown diagram is shown, with all its sub-regions, in Fig. 15. Coordinates of salient points on Fig. 15 are given in Table 11a. Table 11b gives the algebraic equations of the various boundary curves on Fig. 15, except in two cases, curves GAB and JAD, which are given numerically in Table 11c. Fig. 16 compares the ratcheting and shakedown regions with the equivalents for the original Bree loading [1], [4], and also with the in-phase or anti-phase case, [5], [6], [7].

Ratchet strains

The ratchet strain (normalised by the yield strain εy=σy/E) is given by εratchet=2Y(af) for all ratchet regions except R1f for which the ratchet strain is εratchet=4aY. The parameters a and f may be found using Table 1, Table 2, Table 3, Table 4, Table 5, Table 6, Table 7, Table 8 for each of the eight ratchet regions separately. The results are plotted against Y for a range of X values in Fig. 17.

Cyclic plastic strains

In the cyclic plasticity regions, P1 and P2, the cyclic plastic strain ranges are given by εL=2Y(0.5+g) on the left-hand surface, and εR=2Y(0.5c) on the right-hand surface. Note that these are ranges, not amplitudes. (Recall that the two surfaces differ because the right-hand side is the positive side of the secondary bending load). The parameters c and g have been found numerically by applying the rules of §3.

It may sometimes be overlooked that a structure undergoing ratcheting will also be

Conclusion

The complete analytic solution for the positive out-of-phase cycling variant of the Bree problem has been derived. Fig. 15 shows the resulting ratcheting, shakedown and cyclic plasticity regions on the X,Y diagram. Fig. 16 compares the ratcheting and shakedown regions obtained for positive out-of-phase cycling with the original Bree loading [1], [4], and also with the in-phase or anti-phase case, [5], [6], [7].

In the region where design code limits on primary loading are likely to be respected,

References (10)

There are more references available in the full text version of this article.

Cited by (0)

View full text