Stuctural design of pipe linings 1998 pot

1 STRUCTURAL DESIGN OF PIPE LININGS 1998 - REVIEW OF PRINCIPLES, PRACTICE AND CURRENT DEVELOPMENTS WORLDWIDE J E Gumbel International Technical Director, Insituform Technologies SYNOPSIS Current design methodologies for both non-pressure and pressure linings apply simple concepts introduced 15 or more years ago when renovation techniques using plastics pipes and components were still in their infancy. While these concepts have served a useful purpose in providing a widely accepted basis for growth of the industry hitherto, they are increasingly recognised as too conservative or too imprecise for the full range of applications of today’s more highly engineered pipe lining products. This paper reviews the principles underlying current methods of lining design, including those practiced in the Australasian region, and introduces new concepts of structural classification and analysis which are the subject of much discussion in the US and Europe and proving influential in shaping the next generation of design standards at both national and international level. The scope of current research activity in various countries is also briefly described. INTRODUCTION - HISTORICAL DEVELOPMENT The origins of current design concepts for pipe linings can be traced back to the earliest developments of the various classes of rehabilitation techniques using plastics pipes and components, as illustrated in Fig 1. The technique of sliplining using continuous or discrete loose-fitting pipes (Fig 1a) dates back to the 1960’s. Such pipes are clearly independent of their host and are designed accordingly. Their internal pressure rating is based on long-term burst pressures derived from tests on pipes without any external radial support. Resistance of the liner pipe to external pressure is provided by the ring bending stiffness S f defined as: S f = EI/D 3 (1) where E = short-term flexural modulus I = second moment of area (= t 3 /12 for solid-wall pipe of thickness t) D = median diameter of liner pipe (= outside diameter minus thickness) 2 Where the host pipe is not structurally sound or may degrade over time, it is usual to fill the annular space around sliplined pipe with hydraulic grout. This fixes the line and level of the liner pipe (important for gravity sewer applications), and also imposes a temporary external pressure during grouting which often governs the choice of ring stiffness. The associated design limit state is unrestrained ring buckling for which the critical pressure (P cr ) 0 is given by the classical formula: (P cr ) 0 = 24 S f = 2 E (t/D) 3 (2) The invention of the Insituform® (cured-in-place) lining process in the early 1970’s introduced a new class of flexible, close-fitting liner pipe (Fig 1b) which has important structural as well as hydraulic advantages, especially for gravity sewer applications. Apart from maximising the free bore of the renovated system, such linings are generally sufficient to stabilise the existing pipe-soil structure by arresting infiltration-erosion or internal corrosion which are the primary causes of most sewer degradation. With the existing structure thus stabilised, the only structural demand on the lining itself is to resist external groundwater pressure. Here the closeness of fit of the lining carries the further advantage that radial restraint from the host pipe enhances its hydrostatic buckling resistance. Based on experiments at Coventry University during the early 1980’s [1], the first edition of the UK Water Research Centre’s pioneering “Sewerage Rehabilitation Manual” [2] incorporated an enhancement factor K = 7 applied to the unrestrained buckling formula of Eq(2) in design charts for wall thickness of close-fitting, circular, cured-in-place liner pipe subject to sustained external water head. Thus the concept of beneficial interaction between a flexible liner pipe and its host was effectively born. The same principle has been exploited in the subsequent development of close-fit thermoplastic lining techniques, which apply a variety of diameter reduction or folding techniques to temporarily reduce the liner pipe cross section for insertion [3,4]. For grouted sliplined pipe which does not achieve the same degree of radial restraint the WRc manual applied a factor K = 4. This UK approach to the design of flexible sewer linings (classified by WRc as “Type II”) left open the question of what would happen if the existing structure continued to deteriorate after lining. Without giving much thought to the mechanisms available for the transfer of soil loads to the liner pipe, engineers in a number of countries, led initially by the USA [5], Scandinavia [6] and France [7], adopted what seemed a prudently conservative assumption of treating the liner pipe as though it had been conventionally buried. But whilst this approach was reasonable for the quite distinct class of rehabilitation illustrated in Fig 1c, in which lengths or segments of plastics liner pipe form a rigid composite with a layer of structural grout (WRc “Type I”, applicable in man-entry sizes only), it was and remains irrational for flexible linings which by their very nature cannot restore the stiffness and strength of a clay, concrete or iron pipe, but work instead by enabling the surrounding pipe-soil structure to maintain internal equilibrium. The shortcomings of “equivalent buried pipe” design methodology have been highlighted by the author at previous No-Dig conferences [4, 8], and in the present paper this theme is developed further by showing how a rational model of soil load transfer to a liner pipe requires a more detailed examination not only of the existing condition of the pipe to be renovated but also of its potential for further deterioration after lining. Finally, for pressure applications only, a further class of lining is the coated fabric hose (Fig 1d), originally introduced in Japan in the early 1980’s as an earthquake-resistant technique for renovating gas distribution mains. In all but the smallest sizes such hoses lack both ring stiffness and fully independent internal pressure capability, but provide effective long-term leak-sealing by bridging holes and gaps in otherwise structurally sound host pipe. The same interactive principle for pressure lining design is also increasingly applied to non-bonded cured-in-place and close-fit thermoplastic pipes which have ring stiffness and hence some capability to resist net external as well as internal pressure [9]. (It should be noted here that although hoseliners are usually backed with resin adhesive to resist collapse 3 when empty and/or as a barrier to tracking of gas along the annulus, the adhesive bond does not contribute significant external pressure resistance because its peel strength has to be limited to survive cracking or joint movement in the host [10,11]). Against this historical backdrop, the structural action of flexible liner pipe in response to both internal and external pressure will now be described afresh from first principles, with detailed reference where appropriate to both current design practice in various countries and proposals for more rational methods of calculation. PRINCIPLES AND PRACTICE (1): PRESSURE LINING DESIGN The concepts of independent and interactive pressure linings, as already adopted for the preparation of draft European product standards [12,13], are further illustrated in Fig 2. The independent action of a loose-fitting liner (Fig 2a) is clear, but a close-fitting plastics liner which is capable of resisting the internal pressure on its own (Fig 2b) will in practice transfer some or all of its hoop tensile hoop stress over time to a host pipe made of very much stiffer material: cast iron, for example, is around 500 times more rigid than polyethylene (PE). An independent close-fitting liner will thus in general only be fully stressed at localised points where the host pipe has already failed, and should in principle be capable of surviving the possibly dynamic event of host pipe failure at any time during its design life. Interactive liners (Fig 2c) offer a cost effective solution with minimal reduction of free bore for leak- sealing and protection of pipelines subject to corrosion or other attack predominantly from the inside. Such liners require flexibility of both their material and installation technique to ensure that they come into contact with the host pipe wall throughout their length, including, where applicable, inside bends or elbows. Apart from this aspect, the choice of wall thickness, and in the case of resin-fibre composite liners the design of the wall structure, are governed by considerations of short and long-term material failure rather than stiffness. Because the mechanisms of failure of the various interactive liner materials are very different, it is not possible in principle to derive common design equations for all products. For example, the ultimate limit state for long-term hole spanning is predominantly plastic shear failure for PE liners [14], tensile rupture for woven fabric hoseliners [11], and flexural failure in the case of thicker resin-fibre composite pipes [5]. Although the possibility of developing common functional tests for all interactive pressure liners is currently under discussion within European standardisation groups, this has yet to be realised, and for the time being the design of individual pressure lining products is generally based on a combination of stress analysis with conventional laboratory tests for material characterisation. In the structural analysis of pressure linings and the associated jointing systems or end seals it is essential to take account of longitudinal as well as sectional forces. These generally include some residual stresses from the installation process, the effects of temperature cycling, and the Poisson’s ratio component of internal pressurisation stresses needed to close any small initial annular gap. Once all the above factors have been taken into account, it is convenient to express the design performance of interactive linings in terms of the maximum hole and gap sizes that can safely be spanned under a given pressure. Since potential corrosion hole sizes are generally a function of the material and thickness but not the diameter of the host pipe, for situations in which joint gap spans are less critical than holes it turns out that the pressure rating of interactive liner pipe is a simple function of its wall thickness. This contrasts with independent pipe for which pressure rating depends primarily on the ratio of diameter to wall thickness, D/t (also referred to as Standard Dimension Ratio or SDR). Thus the relative benefits of interactive pressure linings in terms of both preserved capacity and cost saving tend to increase with increasing pipe size. 4 Some pipelines which normally operate under internal pressure may also be subject to net external hydrostatic pressure in the short or medium term, for example if exposed to vacuum surges or buried below the water table and periodically de-pressurised or emptied. Design for this load case requires checking buckling stability in the same way as for non-pressure applications, and for interactive linings of diameter greater than around 500mm will often govern the selection of wall thickness. PRINCIPLES AND PRACTICE (2): NON-PRESSURE LINING DESIGN The design of flexible liner pipe to resist external loading differs fundamentally from design to resist internal pressure in the following three important respects: a) The critical limit state for circular or near-circular linings is almost invariably buckling rather than material failure; this is turn allows a unified treatment of all materials based on a simple model of material stiffness, including, as is essential for plastics, a visco-elastic component which may manifest as creep (time-dependent strain under constant stress) or relaxation of stresses at constant strain. b) Changes in geometry along the axis of the pipeline (e.g. stepped joints, bends or other localised shape irregularities) tend to enhance the stability of a close-fitting lining, so that use of a two-dimensional structural model produces a safe design. c) In addition to uniform (radial) forces and deformations, the lining cross section may be subject to distortional loading and ovality; failure to appreciate that the balance of structural action and response in these two modes is not the same for linings as for conventionally buried pipes is perhaps the single greatest source of confusion in the design of renovation systems in current practice. There are further two distinct load cases to consider: 1. External hydrostatic loading due to groundwater pressure plus, where applicable, any internal vacuum; this loading action is purely uniform. 2. Soil and traffic loading transferred through the existing pipe to the lining; this may in principle include both uniform and distortional components (see Fig 3). The conditions for development of these load cases and the associated lining responses will now be considered in turn. External Hydrostatic Loading In any leaking pipe buried below the groundwater table it must be assumed that a non-pressure lining will be subject to sustained external pressure acting in the annulus. Fig 4 shows the two most common models of restrained hydrostatic buckling of circular linings applied in current design practice. Fig 4a depicts the concept of an enhancement factor K [2,5] applied to unrestrained ring buckling pressure (P cr ) 0 , and also illustrates the linear, elliptical buckling mode implicitly assumed. The term linear here refers to the assumption that pre-buckling deformations are small, and do not involve any circumferential strains. It is clear that the assumed buckling mode is not in fact consistent with the boundary conditions, which if strictly enforced would give an infinite K-value. 5 Fig 4b depicts the alternative “free arc” concept [6,7] developed originally to model the condition of an imperfectly grouted sliplined (or spirally-wound [15]) pipe assumed to be fixed around part of its circumference and effectively unrestrained around an arc of angle 2θ. Again this model is not fully consistent with the boundary conditions: it turns out that the associated buckling formula refers to an asymmetric deformation mode, as illustrated, and is once again linear, implying an infinite buckling pressure as the half-angle θ tends to zero. These theoretical shortcomings of current restrained buckling formulae do not entirely undermine their practical value, but they do mean that design values of K can only be determined empirically from actual buckling tests, and that θ, which can be expressed as a simple function of K, does not in fact correspond to any physically measurable angle. Extensive external pressure testing in the USA [16] has confirmed that the design value of K = 7 originally derived in the UK [1] is conservative for most cured-in-place products, but the results also highlight influences on buckling pressure which are not explained by the simple models of Fig 4. A more accurate depiction of the restrained ring buckling phenomenon is given in Fig 5. The only possible initial deformation of a close-fitting, rigidly encased liner pipe is uniform ring compression (Stage 1). The subsequent development of buckling is therefore by definition non-linear. As an initial annular gap develops, the inward radial deformation becomes concentrated in either one or two lobes (Stage 2), which with increasing pressure become flatter and narrower. Eventually “snap-through” instability of one or other lobe (due to changing geometry and ring compressive stress) occurs at critical pressure P cr (Stage 3). Deformation generally becomes visible only after buckling (Stage 4) when it may appear as a continuous (elastic) inward fold or as a cusped or “boat-hull” shape if due to the severe and rapid local straining immediately after the onset of instability the flexural breaking strength of the liner pipe has been exceeded at mid-lobe. These phenomena have been observed and recorded in numerous experiments [1,16,17,18]. An analytical solution for single lobe buckling of this type, given perfect geometry (i.e. rigid circular encasement of initially tight-fitting ring) has been available since 1977 [19] and reduces to the simple form: (P cr ) 1 = E (t/D) 2.2 (3) The results of short-term tests conducted with near perfect geometry have been found to conform closely to this formula, both qualitatively (reflecting the power law in t/D) and quantitatively, and recent theoretical studies [20,21,22] have focussed on assessing the impact on restrained hydrostatic buckling of three types of imperfections: initial annular gap, local geometric imperfections of the liner wall, and ovality of the host pipe. Because the non-linear buckling mechanism of a close-fit circular lining is initiated by opening of an annular gap, buckling pressure is quite sensitive to initial lack-of-fit or any gradual shortening of the liner pipe perimeter due to compressive creep strain. Likewise there is sensitivity to local imperfection and ovality when each is considered separately. Some authors [20,21] have developed theoretical reduction factors in respect of each type of imperfection which they apply in turn to the formula of Eq(3), but more recent analysis [22] of combined imperfection effects has shown that the associated reduction factors are not in fact multiplicative. Where significant ovality (2 - 5% or greater) is present it turns out that this is the dominant imperfection, which for example forces buckling into the more favourable two-lobe mode (Fig 5); the additional impact of gap is then relatively minor. Use of the traditional buckling formula (Fig 4a) with its associated ovality factor [5] has in any event been demonstrated to be safe for typical site and installation conditions. The next generation, fully consistent, design equation for hydrostatic buckling is likely to take the form: 6 P cr = C i E (t/D) m (4) in which the index m as well as the coefficient C i varies continuously with degree of imperfection, from a value m = 2.2 for perfect encasement, to m = 3.0 for a completely unrestrained lining. The most significant benefit of an improved theoretical model for restrained hydrostatic buckling is that it enables a rational treatment of the important effects of material creep on long-term stability, which is usually the governing design condition for selection of liner pipe wall thickness. Hitherto, based on the simple buckling concepts of Fig 4, the practice worldwide has been to assign a reduced long-term value E L to flexural modulus E in the linear buckling equation. Like the enhancement factor K, however, the creep factor F L = E L /E is an empirical quantity which bears no clear relationship to any independently measurable material characteristic. A proper understanding of the non-linear buckling mechanism shown in Fig 5 suggests two distinct influences of creep deformation: a) For near perfect, close-fit circular systems, the effects of compressive creep of the liner pipe ring will dominate, and may conveniently be represented as a time-dependent component of annular gap incorporated in Eq(4) through the factor C i . b) For systems with significant gap or ovality imperfection, any influence of time- dependent flexural strain as the index m approaches 3 may be incorporated by way of a further reduction factor which is a function of both m and a relevant measurable index of creep factor. Recent research at the University of Bradford [23] is pointing the way to a simplified presentation of Eq(4) in the form of a design chart for any given lining product which relates buckling pressure at a specified design life to dimension ratio D/t and host pipe ovality, with the effects of characteristic annular gap of the system and measured axial/flexural creep behaviour of the lining material already incorporated. In the meantime full-scale long-term external pressure testing on individual renovation products [16] provides the only sure way to validate values of overall hydrostatic buckling creep factor F L used in current design formulae. Conditions For Transfer Of Soil Loads To Liner Pipe One of the common misconceptions in sewer lining design is that longitudinal cracking and associated vertical deformation of an existing clay or concrete pipe, a form of damage frequently observed by CCTV inspection, can of itself lead to full transfer of soil and traffic loads to a flexible liner pipe used for renovation. This idea is fundamentally wrong for two principal reasons. First, it neglects the important rôle of the soil as a component of the structure as well as a potential loading medium. When a rigid pipe cracks and deforms it remains stable despite total loss of its flexural ring stiffness by developing passive lateral support from the soil surround in the same way as a flexible pipe (Fig 6a). The only structural action required from cracked pipe to maintain stability is to continue resisting ring compression due to the uniform component of earth load (see Fig 2); the distortional component will be fully carried by redistribution of stress within the soil [24]. Since traditional clay and concrete sewer materials, whether in the form of prefabricated pipe or in-situ masonry, are particularly well adapted to resist compression, it is only in exceptional circumstances, such as where part of the pipe wall is missing or severely corroded (Fig 6b), or in the case of brick construction where the mortar is partially washed out or severely weakened, that any earth load at all can be transferred to the lining. 7 The second common error of reasoning concerns the timing of damage or potential damage to the host pipe which could impact on lining performance. Typically the “equivalent buried pipe” approach is only invoked if the sewer to be renovated appears on inspection to be in poor condition - or in the language of current US practice [5] to be “fully deteriorated”; yet if the host pipe has already lost its flexural stiffness (Fig 6a) or even its compression stiffness (Fig 6b) prior to lining, the fact that it is still standing proves that a new equilibrium has been reached in the surrounding soil, so that the potential for future soil load transfer to the liner pipe is in fact less than if the sewer were still intact or “partially deteriorated”! The most important causes of sewer deterioration in practice are those associated with leakage (erosion of the surrounding soil by groundwater infiltration or stormwater exfiltration/backwash), and internal corrosion (especially by sulphate reducing bacteria in cementitious pipes). Since both these decay mechanisms are effectively arrested by lining with close-fitting plastics pipe, the only scenarios in which the subsequent behaviour may even approach that of a conventionally buried pipe are where substantial additional loads are applied at the ground surface or the soil adjacent to or above the pipe is excavated and then replaced after renovation. Otherwise, as already argued at length elsewhere [8] the behaviour of flexible liner pipe is more comparable with that of a tunnel lining. Response To Transferred Soil Loads Notwithstanding the arguments just presented for discounting the possibility of transferred soil load in most practical situations, it is helpful and in keeping with prudently conservative engineering practice to consider how a plastics liner pipe would in fact respond in a theoretical “worst case” scenario where the host pipe is fully intact at the time of renovation, and afterwards disintegrates completely so as to become indistinguishable from the soil (Fig 7a). The first point to note is that the maximum vertical soil pressure P v reaching the liner pipe-soil structure (see Fig 2) will be appreciably less than the full height of overburden which would act on a conventional buried pipe. Since the soil pressure previously acting on the host pipe would be transferred by unloading at the pipe-soil interface (as occurs in tunneling), due to elastic arching effects alone both uniform and distortional components of interface pressure would be reduced, typically by around 30%, compared with effect of adding load at a rising ground surface (as occurs during burial) [25]. This is even before taking account of frictional arching action, which typically limits the maximum soil effective stress acting on bored tunnel linings to less than the equivalent of one or two diameters of cover. The distortional component of earth pressure P y reaching the liner pipe will be further reduced by the natural consolidation of the soil, aided by groundwater movements and the passage of surface traffic, which will have occurred during the service life of the original sewer pipe if, as assumed, it has remained intact. Such vertical consolidation usually also restores the lateral earth pressure P h , even around a buried rigid pipe, to close to its “at rest” value prior to the original trench excavation. This load redistribution, combined with the enhanced stiffness of the naturally consolidated soil and the absence of the directly distorting effects of soil compaction, reduces the maximum potential out-of- round deflection of the liner pipe after insertion to such a small fraction of that predicted by the usual formulae for conventionally buried flexible pipe as to be effectively negligible. (For reasons already discussed, if the host pipe loses its flexural stiffness before lining, subsequent deflection of the liner pipe will be limited to the effect of reconsolidation of any adjacent softened soil, which again will be an order of magnitude less than that associated with direct burial). This leaves just response to the uniform earth pressure component P z to be considered, for which the relevant limit state is multiwave buckling as illustrated in Fig 7b. The associated design formula currently used in almost all countries for buried flexible pipe design has the form: 8 (P z ) cr = (32 K b E' . S f ) 1/2 (5) where E' is an index of soil modulus derived from buried pipe deflection data, and K b is an empirical coefficient derived from buried flexible pipe buckling experiments which in the most widely used formula published by the American Water Works Asssociation [26] also includes an unverified theoretical water buoyancy factor. In applying this formula to lining design, as for example in ASTM F1216-93, further modifications introduced on an intuitive basis only include the use of long-term modulus E L in the computation of S f , and application of the same ovality factor as used in the hydrostatic buckling formula of Fig 4a. The theoretical inconsistencies in this approach are too numerous to explain in detail here; indeed there is almost nothing right with it! For a start, the square-root power law in Eq(5) arises from a one- dimensional model of soil support which neglects its most important characteristic of shear stiffness; the parameter E' turns out to be as much a function of the assumed load distribution as of the true soil stiffness; and the empirical coefficient K b is based largely on observations of a different mode of buckling applicable only to relatively shallow buried pipe under vacuum pressure which has no relevance to deep-buried sewers [25,27]. Added to this, it has been demonstrated experimentally [27] that ovality does not affect multiwave buckling pressure in the relevant mode range; also the use of long-term ring stiffness in the formula is questionable because it negates the beneficial effect in this case of compressive creep of the liner pipe which leads to relaxation of the applied soil pressure. To cap it all, it is usual to apply a larger global factor of safety to this design case than to the hydrostatic buckling case, ranging from 2.5 in the USA and France [5,7], to as high as 6.4 in some parts of Japan! The combined effect of these theoretical inconsistencies, the overestimate of maximum theoretical soil loading and the accumulation of layer upon layer of implicit as well as explicit safety factors is to make earth pressure buckling invariably appear to be the critical design limit state. This in turn encourages risk-averse engineers to check response to transferred soil load as a rule rather than by exception, resulting in routinely over-conservative designs. If instead a balanced and consistent analysis is made of the “worst case” scenario of Fig 7a, it is frequently found that the hydrostatic buckling limit state is in fact more critical, especially if 2-3m or more groundwater head is present. The simplest fully consistent theoretical analysis of the combined loading and supporting action of the soil surrounding a renovated pipeline is provided by a two-dimensional elastic continuum model. The resulting expression for earth pressure buckling has the form: (P z ) cr = C n (E s ) 2/3 . (S f ) 1/3 (6) where E s is a true soil modulus, and C n is a coefficient with value close to 1. This formula has been shown [27,28] to give far more accurate predictions of a wide range of published buried pipe buckling data than Eq(5), which it should logically replace in due course. Australia is to be commended for being the first - and so far the only - country to have incorporated Eq(6) in its national design standard for buried flexible pipe [29], which serves also as a source reference for the design of sewer linings. Yet while this is an important step towards the overall consistency of approach advocated by this paper, there still remain several aspects of renovation design practice in the Asia-Pacific region as a whole, for example the treatment of lining ring deflection and associated bending strains, which require correction. To conclude this section, the importance of the soil as a key element of the structure of non-pressure pipelines is re-emphasised. One of the least appreciated benefits of trenchless renovation of sewers is that non-disturbance of the soil surround preserves a natural state of consolidation and favourable 9 distribution of stresses in the ground which cannot be replaced in the short-term even by the most careful compaction of fill around a new trench-laid pipe. The enhancement of soil stiffness by natural consolidation also tends to be discounted, sometimes to the extent of using lower values of E' or E s than for conventionally buried pipes, which is quite illogical. This practice simply adds another hidden factor of safety to the already over-conservative design treatment of the transferred soil load case. SUMMARY OF KEY STRUCTURAL CHARACTERISTICS OF LINER PIPE In the light of the design principles set out above, the following notes briefly review some further important aspects of the short and long-term properties of plastics pipes used for renovation. Ring stiffness and flexibility Stiffness is the key characteristic of circular liner pipe for resistance to external pressure. For the most important design case of buckling under sustained groundwater pressure (and also where applicable under transferred soil pressure), appropriate account needs to be taken of the effects of liner material creep in both flexural and compressive modes. There is however no particular advantage in providing ring stiffness greatly in excess of that required to resist buckling with a projected safety factor of more than, say 1.5-2.0 after 50 years. This is because there is inevitably a trade-off between stiffness and flexibility, and it is by virtue of its flexibility that a close-fitting liner can absorb without significant stress any small deformations due to readjustment of soil-pipe equilibrium after renovation. Some design specifications nevertheless enforce a minimum ring stiffness (S f ) requirement for liner pipe, ranging from 1 or 2 kPa in some European countries to reportedly as high as 6 kPa in Japan. This requirement originates from handling difficulties and resulting damage experienced in the 1970’s in the laying and compaction of soil around very flexible (S f < 1 kPa) GRP pipe, but these considerations simply do not apply to renovation techniques, for example cured-in-place pipe, which allow placement of thin-walled linings moulded to the shape of the existing pipe wall without any initial stress. Generally it is counterproductive to add thickness to a circular liner pipe much in excess of that required for stability, since any ring deflection after lining will still be controlled primarily by the surrounding soil, and the extra material will simply increase liner flexural stresses without any benefit. Flexural stresses are in practice only a relevant design limit state in the design of non-circular linings, such as boxes and egg-shapes, which have a relatively flat wall over part of their perimeter [2]. Short versus long-term stress and strain limits Where material failure is the critical limit state, it is in principle necessary to consider long-term strength characteristics of a plastics liner pipe, but it is important as in the case of stiffness to distinguish uniform and distortional deformation modes and the effects of geometric restraint. The phenomenon of tensile creep rupture is well documented for all types of independent plastics pipes for which a safe pressure rating is invariably based on a relationship between applied pressure and time-to-failure determined by testing. Similar design factors on short-term strength also need to be applied to flexural stresses where bending of the liner pipe wall is unrestrained, such as on the flat side of a non-circular conduit or where an interactive pressure lining bridges holes and gaps. In these cases however time-dependent stress redistribution across the thickness of the liner wall in bending means that the long-term design factors applicable to the stress resultants of moment and shear are less than those for pure tensile creep rupture. Where deformation of the liner wall is restrained, for example over the majority of the surface of an interactive pressure liner, or where bending or shear is imposed by soil movement, creep deformation 10 has the effect of relaxing the imposed stresses with time. For such cases the appropriate design check is against the short-term strain limit of the material. RESEARCH NEEDS AND CURRENT ACTIVITY WORLDWIDE In conclusion, a move towards more principled design procedures as outlined in this paper, and the associated removal of irrational layers of conservatism especially in the analysis of soil load transfer to sewer linings, should encourage even more widespread use of trenchless pipeline renovation by increasing customer confidence and understanding of its technical benefits, and helping to remove artificial obstacles to the economic yet safe use of the full range of field proven techniques using well characterised and tested lining materials now available. Progress towards agreeing the basis for a common rational design methodology for circular sewer linings has recently been made at international level [30], and key researchers and customers in Europe and North America have identified a number of items requiring further research including: o Extension of current UK and US studies of restrained hydrostatic buckling to enable relevant characterisation of a range of different lining techniques/materials o Design of non-circular (especially egg-shaped) sewer linings and clarification of the related boundary between buckling and material failure as critical limit states o Effects on system behaviour of seasonally varying water table compared with the usual assumption of a constant external head acting throughout the design life o Replacement of arbitrary “global” safety factors with a proper limit state approach applying relevant partial safety factors to loads and material properties The discussion in this paper of the relevant modes of interaction between lining, pipe and soil has further highlighted the need for: o An improved system of damage classification for existing sewer pipe which enables assessment of the current and potential future loss its compressive as well as flexural ring stiffness. o Laboratory and/or field studies to directly measure the extent for potential further deterioration under different conditions of various pipe materials after lining, to determine the related applicability of interactive lining design in the case of pressure pipes, and maximum possible soil load transfer in the case of non-pressure pipes. Even before the results of such research become available, the scope for beneficial adjustments to current renovation design practices are, in the author’s view, considerable, and the pooling of international experience should provide added confidence for individual countries or specifying bodies to take the necessary steps. REFERENCES [1] Aggarwal, S.C. & Cooper, M.J., “External Pressure Testing of ‘Insituform’ Linings” Internal Report, Coventry (Lanchester) Polytechnic, April 1984 [...]... I.D “Parametric Study for Buckling of Liners: Effect of Liner Geometry and Imperfections” Trenchless Pipeline Projects: Practical Applications, ed L.Osborn, 416-213, ASCE, June 1997 [21] Falter, B., “Structural analysis of linings for sewer renovation”, 5th Int Conf on Pipeline Construction, Hamburg, October 1997 [22] Boot, J.C., “Elastic buckling of cylindrical pipe linings with small imperfections... Technical Report 11295, "Techniques for Rehabilitation of Pipeline Systems by the use of Plastics Pipes and Fittings", August 1992 [4] Gumbel, J.E.,"Rehabilitation of Sewers by Structural Lining", Asian Trenchless Tech '93, Dubai, April 1993 [5] ASTM F1216-93 “Standard Practice for Rehabilitation of Existing Pipelines and Conduits by the Inversion and Curing of a Resin-Impregnated Tube”, American Society for... imperfections subject to external pressure” Trenchless Technology Research 1998 (in the press) [23] Boot, J.C and Javadi, A.A., “The structural behaviour of cured-in-place pipe Paper to be presented at the Plastics Pipes X Conference, Göteborg, Sweden, 14-17 September 1998 [24] Gumbel, J.E and Wilson, J., “Interactive Design of Buried Flexible Pipes - a Fresh Approach from Basic Principles”, Ground Engineering,... Discussion of “Elastic solutions for a deep circular tunnel” by M.J.Pender, Géotechnique, Vol.31, No.3, 434-435, 1981 [26] AWWA C950-88 “Standard for Fiberglass Pressure Pipe , American Water Works Association, Denver CO, June 1988 [27] Gumbel, J.E., “Analysis and design of buried flexible pipes” Ph.D Thesis, University of Surrey, 1983 [28] Moore, I.D “Elastic Buckling Of Buried Flexible Tubes - A Review Of. .. of sewer pipelines - Chapter 5: Renovation design and installation - General principles] Danish Technology Institute (also Swedish VAV P66), September 1989 [7] AGHTM (France) “Recommandations pour la réhabilitation des réseaux díassainissement Methode de calcul chemisage et tubage” [Recommendations for rehabilitation of sewer pipe networks Structural design method for cured-in-place and sliplined pipes],... 2nd Ed 1998 [8] Schrock, B.J & Gumbel, J “Pipeline Renewal 1997”, North American No-Dig '97, Seattle, April 1997 [9] Gumbel, J.E, Heavens, J.W., & Boot, J.C Design and selection criteria for plastics lining systems for the renovation of pressure pipelines” 1st Int Conf on Coatings and Linings for the Water and Wastewater Industry, Prague, June 1995 [10] Smith E.P & Gumbel, J.E., “Renovation of Low... “Experimental determination of buckling capacities of cured-inplace pipe linings , Paper to be presented at the 11th International Conference on Experimental Mechanics, Oxford U.K., 24-28 August 1998 [19] Glock, D., “Überkritisches Verhalten eines starr ummantelten Kreisrohres bei Wasserdruck von Aussen und Temperaturerhöhung” [Post-critical Behaviour of a Rigidly Encased Circular Pipe Subject to External... Flexible Pipelines - Part 1: Structural Design Australian/New Zealand Standard, 1997 [30] Trenchless Technology Research Colloquium, Rehabilitation Group Communiqué “Structural performance of close-fit non-bonded flexible linings for nominally circular gravity systems: agreed basis for a rational design methodology”, published January 1998 on http://www.latech.edu/~guice/TTC/forum.htm 12 FLEXIBLE LINER PIPE. .. 1997 [14] Boot, J.C., Guan, Z.W., & Toropova, I., “The structural performance of thin-walled polyethylene pipe linings for the renovation of water mains” Trenchless Technology Research Vol.11, No.1, 37-51, 1996 [15] ASTM F1741-96 “Standard Practice for Machine Spiral Wound Poly(Vinyl Chloride) (PVC) Liner Pipe for Rehabilitation of Existing Sewers and Conduits”, American Society for Testing and Materials,... and textile hoses for renovation of gas mains], Gas-Erdgas, Vol.136, No.1, 38-46, 1995 (describes basis of draft German standard DIN 30658-1, April 1996) [12] Elzink, W and Gumbel J., "The Development of International Standards for Pipeline Renovation", International No-Dig '94, Copenhagen, June 1994 [13] CEN/TC155/WG17 “Guidance on classification and design aspects of plastics piping systems used for . importance of the soil as a key element of the structure of non-pressure pipelines is re-emphasised. One of the least appreciated benefits of trenchless renovation of sewers is that non-disturbance of. hidden factor of safety to the already over-conservative design treatment of the transferred soil load case. SUMMARY OF KEY STRUCTURAL CHARACTERISTICS OF LINER PIPE In the light of the design principles. how a rational model of soil load transfer to a liner pipe requires a more detailed examination not only of the existing condition of the pipe to be renovated but also of its potential for further