Alwyn S. Tooth

A parametric study of metal-to-metal full face taper-hub flanges

The paper presents the results of a parametric study, using finite element analysis, of the behaviour of full face metal-to-metal taper-hub flanges. The important stress values in the flange have been obtained for a range of flange thickness, taper-hub thickness and length, when the shell/flange component is subject to internal pressure. The influence of the pre-stress in the bolts is examined. The results obtained have been compared with the predictions from the appropriate sections of the ASME, BS and the new European Unfired Pressure Vessel Standard (Draft BS: prEN 13445).

A study of the buckling behaviour of horizontal saddle supported vessels

Previous experimental work, by one of the authors, examined the behaviour of end supported cylindrical vessels loaded centrally. It was found that the vessels failed by buckling when the radius to thickness ratio (R/t) was greater than 150. These results provided the motivation for examining the buckling behaviour of such vessels when they are supported, in a more conventional way, by using two saddles. In the cases examined it was noted that the stresses that cause buckling behaviour are the longitudinal and circumferential membrane stresses. These occur at four vessel locations, i.e. the zenith and nadir (top and bottom) of both the vessel mid-span and the saddle centre profiles. Known buckling formulae based on simple loading patterns, such as an axially loaded cylinder and a cylinder under pure bending, will be utilized in determining the allowable buckling stress. Present British Code rules and European recommendations will also be discussed. The allowable buckling load will be compared with theoretical stresses obtained from a small displacement linear elastic analysis, using a Fourier series method. From these results a design method will be presented.

The Support of Horizontal Vessels Containing High-Temperature Fluids—A Design Study

This paper investigates the behavior of horizontal cylindrical vessels, subjected to thermal loading by high-temperature fluid, where the saddles are fixed to the supporting structure. In order to determine an optimum saddle design, three widely used saddle configurations, with differing saddle heights and top saddle plate extensions, are explored. Thereafter, one of the saddle designs is selected to illustrate a decoupling procedure, for the radial and axial expansions, whereby design charts are obtained to derive the maximum stress values for a range of vessel geometries. The finite element approach, using linear elastic, small displacement analysis, is used throughout.

Radial loading of a cylindrical vessel through a rectangular rigid attachment

Abstract This paper examines the behaviour of a cylindrical vessel which is radially loaded through a rigid attachment of rectangular plan form. The resulting radial and tangential interface forces between the vessel and the attachment are derived assuming that the attachment is fixed to the vessel at all points over the mating surface and subjected to a radial displacement. From these interfacial forces the stresses in an illustrative vessel are found to be some 37% higher than those which are obtained when a uniformly distributed radial interface pressure is assumed.

A study of the stability of horizontal cylindrical vessels filled with varying amounts of liquid— Part I: Experimental investigations

Abstract The instability behaviour of thin-walled horizontal cylindrical vessels supported by means of ring girders, or from rigid end plates, is examined experimentally during liquid filling. Selected results from a horizontal test on a large stainless steel horizontal storage vessel are presented to illustrate the way in which thin-walled vessels, of large radius/thickness ratios (300–650), behave during fill. Such results have provided the motivation for some 26 polyester film model cylinder tests. The models cover three different radius/thickness ratios (300, 400 and 500) and a range of length/radius ratios. They were supported at their ends and progressively filled with liquid. In each case the onset of instability was noted. In Part II 1 of this paper a theoretical approach is presented and compared with the experimental values in an effort to assess the validity of design methods currently in use.

The radial loading of cylindrical vessels—influence of attachment rigidity

Abstract The present version of the Pressure Vessel Standard, BS 5500 1 restricts the design of the local load type situations to certain vessel and attachment geometrics. This paper identifies the main underlying reasons for these restrictions; which are a limitation of the deflection and rotations by the use of small displacement analysis, and a neglect of the rigidity of the attachment. This paper addresses the influence of the attachment rigidity by considering a number of attachment thicknesses. The companion paper 2 considers the large displacement phenomenon and brings the two effects together in a design approach for these components. The method proposed provides a series of stress and displacement correction factors whereby the existing British Standard design method can be modified to cover the entire range of geometries up to ( C φ R ) = 0·25 for all values of ( R T ) up to 250.

The Support of Horizontal Cylindrical GRP Vessels—Saddles or Longitudinal Beams?

Horizontal cylindrical GRP vessels are normally supported on either twin saddles or longitudinal beams which extend the full length of the vessel. To examine the way in which these supports influence the strain distribution in the vessel, results are presented from a series of experimental investigations. These have been carried out on a 1 m diameter, 4 m long, CSM reinforced polyester vessel subject to hydraulic loading. In the first series of tests the vessel was supported on twin saddles of various geometries and vessel/saddle interface conditions. In a later series, longitudinal beams, extending the length of the vessel were used to provide support.

A study of the stability of horizontal cylindrical vessels filled with varying amounts of liquid—Part II: A theoretical approach and design considerations

Abstract A simplified theoretical approach to the problem of the stability of horizontal cylindrical vessels filled with liquid is proposed and is compared with the experimental results presented in Part I of this paper and with those of other investigators. In view of these comparisons certain design considerations are proposed for liquid storage vessels.

An alternative way to support horizontal pressure vessels subject to thermal loading

When storing liquids at high temperature in horizontal vessels, the current design methods aim to minimise the thermal stresses by introducing a sliding surface at the base of one of the twin saddle supports. However, regular site maintenance is required to ensure that adequate sliding is achieved. This may be difficult and costly to carry out. The aim of the present work, therefore, is to dispense with the sliding support and to provide saddle designs which, although fixed to the platform or foundation, do not result in the storage/pressure vessel being overstressed when thermal loading occurs. This paper provides general recommendations for the most appropriate saddle geometries, and details the way in which design-by-analysis and fatigue-life- assessments may be carried out using the stresses that arise from these designs.

Parametric equations for maximum stresses in cylindrical vessels subjected to thermal expansion loading

Abstract This paper derives a set of parametric equations for finding the maximum stresses developed in a cylindrical vessel which is supported by two saddles firmly secured to the foundation and subjected to thermal expansion loading. Three types of stresses are considered: maximum stress intensity, maximum circumferential stress and maximum axial stress. The maximum stresses in the vessel are found to be governed by the height and width of the saddle, the spacing between the two supports and the relative structural rigidity between the support and the vessel. This paper is an extension of the work reported by Tooth et al. (1996) [1] . Using a least square curve fitting procedure, parametric equations for the maximum stresses developed in the vessel have been established. Raw data used for the curve fitting were obtained from a comprehensive finite element study which covers a wide range of key dimensions. A total of 900 finite element runs have been performed. The derived parametric equations are subsequently validated against the raw data and their error bounds are established. In all cases the maximum errors are found to be within 20%. The established parametric equations can be used directly in design calculations. The curve fitting procedure outlined in this paper has wide application for any set of generated or measured data.