International Journal of Offshore and Polar Engineering Vol. 2, No.

2, June 1992 (lSSN 1053-5381) Copyright by The International Society of Offshore and Polar Engineers

Egil Skomedal Veritec, H~vik, Norway

Pipeline free spans are often calculated as a continuous beam, without the effective axial force taken into account and, particularly, without the sliding of the pipeline towards the span taken into account. Correspondingly, oscillations are usually calculated as for a straight beam with zero axial force. Such calculations are not correct, except for "medium" spans if the effective axial force happens to be zero. Shorter spans are usually subjected to a compressive effective axial force; longer spans may be subjected to tension and may have considerable sag. The tension and the sag imply that the span acts partly as a suspended cable, so that the static capacity is large compared with a beam of the same dimensions. The tension and the sag may also affect the dynamic behaviour of the span. One reason a pipeline span does not behave as a beam is the pipeline's tendency to elongate when set in operation. A pipeline that is not restrained will elongate. If it is restrained, a compressive effective axial force will act in it instead. A free span is a special case of reduced restraint, where the pipeline takes the opportunity to elongate towards the span from both sides in order to get rid of the compressive effective axial force in it. For longer spans even tension, combined with sag, may be necessary for the vertical equilibrium of the span.

if the pipeline was laid empty and with (residual) horizontal pull H.

v Poisson's ratio a coefficient of thermal expansion L1T = temperature increase relative to the installation temperature E modulus of elasticity
Note that Pe disappears from Eq. 2 if the pipeline is laid empty (airfilled), which usually is the case. So is an important quantity even for pipelines in which So can not exist (due to insufficient restraint). If only a force S can exist, an elongation per unit length equal to

= __

S-S

EA
will necessarily take place. As certain points on a pipeline are fixed, the elongation must be taken up by deflection. One must assume that this deflection usually takes place in the horizontal plane. A span is a special case where the deflection takes place in the vertical plane, and under the (additional) action of the submerged weight of the pipeline. What happens in a span is otherwise much the same as what may happen at a location with less than average lateral restraint. The effective axial force (Ss) that can (or must) exist in a span is usually less negative (or more positive) than S some distance away from the span.

The notion effective axial force is used above to make clear that it is not the force acting in the pipe wall. The effective axial force is an important quantity in the calculation of pipeline behaviour, particularly the behaviour of free spans. The relation between the effective axial force (S) and the pipe wall force ("true" axial force, N) is:

Pi = internal pressure Pe = external pressure Ai = internal cross-sectional area (= Ae Ae = external cross-sectional area A = cross-sectional (steel) area

A span of length L will usually cause deflection of a somewhat larger length (L1) and elongation of a considerably larger length (L2) (which may be several kilometres for a larger pipeline).
-

A)

Received February II, 1991: revised manuscript received by the editors March 2, 1992. The original version (prior to the final revised manuscript) was presented at The First International Offshore and Polar Engineering Conference (lSOPE-91), Edinburgh, United Kingdom, August 11-16, 1991. KEY WORDS: Pipelines, free spans, effective axial force, true axial force.

Fig. 1 Idealized situation if I SJ ~ not buckle.

Sth

l , i.e.,

if the pipeline does

arc length within the deflected part of the pipeline, and the change in arc length must be equal to the elongation due to the change in axial force caused by the span.
~ 'P

---

"i

--

------

E~-;;;'~iint Mea:- 'lrue"aNa

ELongation proportional to maded area_

Referring to Fig. 1 the elongation taking place within L2 shall be taken up by deflection within LI. The practical calculation can be carried out in different ways. Iterations are in any case necessary due to the nonlinear relation between the sag and the effective axial force in the span. Methods using "exact" shape of the deflected portion (LI), including use of nonlinear finite element programs, seem to be somewhat troublesome. However, considering all uncertainties involved in real cases, methods based on approximate deflected shape may be sufficiently accurate. Such a method is described by Skomedal (1991), who made the following two simplifications: The shape of the second order bending moment diagram (that from the effective axial force) is described approximately by two factors, one defining the area compared with that of a known rectangle, and one defining the position of its centre of gravity. The relation between the elongation and the sag is simplified. Considering that the maximum curvature is at the shoulders, one will realize that little of the elongation will be taken up within the lifted length (II) on each side of the span. The elongation in relation to the sag is taken a little larger than that corresponding to a parabola within the span (L).

Fig. 2 "Missing" compressive force causing elongation "pushed" into the span. I SJ >I StJ Obviously, L2 depends on the difference between S8 and the force (S) immediately outside L2, and the longitudinal friction. See Figs. 1 and 2. (S above represents So or Sth') The problem is easiest if the pipeline outside L2 is sufficiently restrained to resist So (so that S = So), This could be the case for a buried pipeline (with an exposed portion over a depression). Otherwise S may be estimated on the basis of lateral friction, general bottom conditions and "snakedness" of the pipeline. Fig. I illustrates a case where the pipeline can resist So except in the span, and Fig. 2 illustrates two cases where this is not the case. The maximum effective axial force that can be resisted is Sth' and some places it is even smaller, depending on the actual lateral restraint, which naturally is not constant over the length. In the case shown in Fig. 2a the depression existed before the pipeline was laid. In the case shown in Fig. 2b the depression occurred later due to erosion. The elongation taken up by sagging in the span is different in the two cases. This is illustrated by the shaded areas, which are proportional to the elongation taken up by the respective span. The essential point in the calculation of a span is that the effective axial force is connected with the deflection (sag) in such a way that equilibrium requirements and geometrical compatibility requirements are satisfied at the same time: The combination of effective axial force in the span and the deflection of the span must satisfy certain equilibrium requirements. A change in deflection must be consistent with the change in

The main conclusion one can draw from calculations carried out along the above lines is that a pipeline free span may behave considerably differently from a continuous beam. The reason for this is the effective axial force, both that actually acting in the span and some distance away from it, and that which would have acted if the pipeline were completely restrained. An important consequence is the great static strength of a long span compared with that of a beam of the same dimensions.

Skomedal, E (Aug 1991). "Static Calculation of Pipeline Free Spans - A Simple and Quick Method," Proc J st Int Offshore and Polar Eng Con/, ISOPE, Edinburgh, Vol 2, pp 281-289. Skomedal, E (Sept 1991). "The Concept of Effective Forces and Its Application in Stability Analyses," Proc 4th Int Colloquium on Structural Stability, Mediterranean Session, Istanbul, pp 351-358.