Pavement engineering in developing countries by C. 1. Ellis 1979 I Digest SR 537 PAVEMENT ENGINEERING IN DEVELOPING COUNTRIES Author C I Eflis PAVEMENT ENGINEEMNG IN DEVELOPING COUNTRIES by C I Eflis Since most developing countries lie in the tropics or sub-tropics the differences between pavement engineering in temperate industriafised countries and developing countries are often thought of almost exclusively in terms of climatic differences. Wst these differences are substantial, a considerable body of knowledge exists that enables experienced designers to incorporate reasonably satisfactory provisions within their designs for all but the most extreme effects of tropical climates. ~udy important differences between pavement engineering in developing and industrialised countries are the greater variabfity of construction materials, quality of construction, and the larger fluctuations in the volume and weight of road traffic that are typica~y encountered in developing countries. . Because of the large variability of these factors the design of road pavements in developing countries must either include a higher ‘factor of safety’ than is usual in industrialised countries, or a higher risk of failure must be accepted. The latter course is the one most commonly adopted. It is an advantage in that it minimises the demands made by road-building on scarce capital resources and the disadvantages of partial or premature pavement failure are much reduced by the short ‘design tife’ generally adopted and the relative ease with which repairs can be made on the typically uncontested roads. In the few fortunate oil-rich developing countries the pavement engineering situation has, however, radically changed in recent years. Low risks of faflure and long design lives are now often demanded in these countries; designers respond by adopting generous safety factors to compensate for the large uncertainties that remain in the prediction of traffic and the variability of materials and quality of construction. An important aspect of pavement engineering in developing countries that has no parallel in most industrialised countries is the extent to which unsurfaced roads contribute to national road networks. Unsurfaced roads of all types play a vital role in the economic and social fife of many of these countries, and such roads, carrying several hundred vehicles per day, are not uncommon. In this respect consideration is given to the traffic bearing ability of earth roads related to the important factors of route selection and drainage. The design of gravel roads is considered in terms of structural thickness and the selection of suitable gravels. The wide range of pavement design methods for bitumen surfaced roads are described and discussed by reference to the pavement thicknesses recommended by such methods. Such comparisons are expressed in terms of ‘Tropical Structural Number’ (TS~ which is a modified form of the AASHTO Structural Number. The work described in this Digest forms part of the programme carried out for the Overseas Development Administration 1 but any tiews expressed are not necessarily those of the Administration. If this information is insufficient for your needs a copy of the full report, SR 537, may be obtained on wn.tten request to the Technical Information and Library Sewices, Transport and Road Research Laborato~, Old Wokingham Road, Qowthorne, Berkshire. Crown Copyright. Extracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged. 260 240 220 200 G 0_ ~ 180 5 % k ; 160 : 140 120 100 80 k...~..~ Fluid A Markey Test o--+ Fluid B Markev Test ●—0 Fluid A In car tests on track o 1.0 2.0 3.0 4.0 5.0 Percentage of water Digest Fig. 1 VAPOUR LOCK TEMPERATURES OF BRAKE FLUID AT DIFFERENT WATER CONTENTS 32 r National sample– Total 1000 cars -- (Rear wheel cylinder samples) 28 - 24 - .:.:.:.:.:.:.:.:.:.:. .:.:.:.:.:.:.:.:.:.:. ..................... ..................... ..................... ..................... 20 - ..................... ..................... ..................... ..................... ..................... ..................... ..................... ..................... ..................... ..................... ..................... ..................... 16’ - ..................... ..................... ..................... . . . . . . . . . . . ..................... ..................... ..................... ..................... 12 - Over Not 7% known 8 - ..................... ..................... ..................... ..................... 4 - ..................... ..................... n 1 2 3 4 5 6 7 Percentage of water in brake fluid Digest Fig. 2 WATER CONTENT IN THE BRAKE FLUID OF A NATIONAL SAMPLE OF CARS The work described in this Digest was carried out in the Vehicle Safety Division of the Safety Department of TRRL, If this information is insufficient for your needs a copy of the fill report, LR 903, may be obtained on written request to the Technical Information and Libra~ Semites, Transport and Road Research Laborato~, Old hbkingham Road, Oo wthome, Berkshire. Crown Copyright. hy views expressed in this Digest are not necessarily those of the Department of the Environment or of the Department of Transport. Extracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged. TRANSPORT and ROAD RESEARCH LABORATORY Department of the Environment Department of Transport SUPPLEMENTARY REPORT 537 PAVEMENT ENGINEERING IN DEVELOPING COUNTRIES by C I EUS The work described in this Report forms part of the prograrnme carried out for the Overseas Development Administration but any views expressed are not necessarily those of the Administration Overseas Unit Transport and Road Research Laboratory Crowthorne, Berkshire 1979 ISSN0305–1315 CONTENTS Abstract 1. 2. 3. 4. 5. 6. htroduction Highway pavement engineering 2.1 Cost models Earth and gravel roads 3.1 &rth roads 3.2 Gravel roads 3.3 Recent developments Bitumen surfaced roads 4.1 ~rrent design methods 4.1.1 4.1.2 4.1.3 4.1.4 4.1.5 TRRL Road Note 31 CEBTP pavement design manual for tropical countries AASHTO interim guide Shefl modified design charts Other methods 4.2 Comparison of pavement design methods 4.3 Recent developments 4.3.1. 4.3.2 4.3.3 4.3.4 4.3.5 tie loads Heavy vehicle sirnuktor The design of bituminous surfacings for high service temperatures Overlay design Impact (square) ro~ers Concrete roads Assessing risk 6.1 Decision mtig 6.2 hproved forms of contract Page 1 1 2 3 3 4 4 5 6 6 7 7 8 9 9 9 11 11 13 14 15 15 16 16 16 17 7. Future needs 8. Achowledgements 9. References Page 17 18 18 (C) CROWN COPYRIGHT 1979 Extracts from the text may be reproduced, except for commercial purposes, provided the source is actiowledged PAVEMENT ENGINEEMNG IN DEVE~PING COUNTRIES ABSTRACT This Report discusses the reasons for the differences between pavement engineering in temperate climates and in developing countries with tropical or sub-tropical cfimates. The importance of earth and gravel roads in developtig countries is emphasised, and commody used methods of pavement design for bitumen surfaced roads are described and compared. Recent developments in techniques and equipment for improving construction standards and for assessing road performance are described. The use of appropriate forms of contract is briefly discussed and the need to ‘ improve knowledge of the factors which would enable total transport costs to be minimised is emphasised. 1. INTRODUCTION Since most developtig countries lie in the tropics or sub-tropics the differences between pavement engineering in temperate industriatised countries and developing countries are often thought of almost exclusively in terms of cfimatic differences. Wtist these differences are substantial, a considerable body of knowledge exists that enables experienced designers to incorporate reasonably satisfactory provisions within their designs for ti but the most extreme effects of tropical climates. ~u~y important differences between pavement engineering in developing countries and industrirdised countries are the greater variability of construction materials, quahty of construction, and the krger fluctuations in the volume and weight of road traffic that are typica~y encountered in developing countries. Because of the large variability of these factors the design of road pavements in developing countries must either include a higher ‘factor of safety’ than is usual in industrialised countries, or a higher risk of faflure must be accepted. The latter course is the one most comody adopted. It is an advantage in that it minimises the demands made by road-buflding on scarce capital resources, and the disadvantages of partial or premature pavement failure are much reduced by the short ‘design life’ genera~y adopted and the relative ease with which repairs can be made on the typicdy uncontested roads. In the few fortunate ofl-rich developing countries the pavement engineering situation has, however, radic~y changed in recent years. hw risks of faihrre and long design tives are now often demanded in these countries; designers respond by adopting generous factors of safety to compensate for the large uncertainties tht remain in the prediction of traffic and the variability of rnaterirds and quahty of construction. An important aspect of pavement engineering in developing countries that has no para~el in most industrirdised countries is the extent to which unsurfaced roads contribute to national road networks. In developing countries unsurfaced roads carrying several hundred vehicles per day are not uncommon, and unsurfaced roads of afl types play a vital role in the economic and social hfe of many of these countries. The techniques of constructing and maintaining unsurfaced roads are thus an important part of pavement engineering in developing countries. 1 The selected statistics shown in Table 1 illustrate the differences between roads and road transport in developing countries and in industritised countries. TABLE 1 Selected statistics of roads and road transport in developing and industridised countries Country United States France United Kngdom Brazil Sierra Leone India Ethiopia * Trucks and buses. Gross national product per capita us $ (ref. 1) 6670 5 Ho 3590 920 190 140 100 bngth of bitumenised or concrete road km x 103 (ref. 2) 2925 707 331 77 1 409 3 Length of earth or gravel road kmx 103 (ref. 2) 3215 87 12 1235 6 486 6 Density of road networks km per krn2 (ref. 3) 0.66 1.4 1.49 0.15 0.10 0.27 0.007 Numbers of commercial vehicles* per km of road 4.29 2.78 5.36 0.76 0.64 0.46 1.19 2. HIGNAY PAVEMENT ENGINEERING Highway pavement engineering may be defined as the process of designing, constructing, and maintaining highway pavements in order to provide a desired ‘level of service’ for traffic. In the pavement design part of this process, designers make assumptions about the methods of construction that WN be employed and the level of maintenance that the pavement wfll receive throughout its design fife. Clearly both the initial standard of construction and the level of maintenance affect the level of service provided by the pavement, these two factors being very closely interrelated. In the USA Hudson4 has drawn attention to the interdependence of these factors and the fact that the design and construction of a road pavement is not a sin~e-phase process, but is merely the beginning of a continuing process in which maintenance, traffic, chmate and the required level of service are.~ contributory factors. In Britain the costs of constructing and maintatiing road pavements over a 50 year design life have been andysed5. h the case of low-cost roads, typical of most developing countries, the influence of maintenance on the level of service provided by the pavement wi~ generally be much stronger than is the case with heavier duty pavements. For unsurfaced roads the level of maintenance is at least as important as the initial construction standard in determining the level of service provided. Idedy at the design stage, the designer, in consultation with the authority that wifl maintain the road, should select the combination of initial construction standard and subsequent maintenance that wfll provide the level of service required over the design life of the road in the most cost +ffective way. In practice, in developing countries the designer rarely has sufficient information to make a quantified assessment of the appropriate balance between the initial construction standard and the level of maintenmce, and because of uncertainty about the provision of maintenance when it is needed, he will often tend to enhance the standard of initial construction so as to minimise 2 future maintenance requirements. Some advances are being made however in providing designers with quantitative information on the ‘trade~fP between the standards of construction and maintenance of roads typical of those found in developing countries. In recent years models have been produced that Mow designers not ody to quantify the interaction between construction and maintenance standards, but also to calculate their combined effect on vehicle operating costs, and hence to produce a design and specify a maintenance strategy that @ minimise the sum of construction, maintenance, and vehicle operating costs over the design hfe of the pavement. The sum of these three costs has been cafled the ‘total transportation cost’, and w~st these costs do not by any means represent dl the costs involved in road transport, they are by far the most significant costs on non-urban roads in developing countries. 2.1 Cost models The initiative for developing cost models for roads in developing countries was taken in the early 1970s by the World Bti. The World Bank recognised the need to improve knowledge of the interaction between road construction costs, road maintenance costs, and vehicle operating costs, in order to improve the qudty of investment decisions in the roads sector in developing countries. As a first step a computer model was buflt on the basis of efisthg knowledge 6 . Subsequently a major study was undertaken in East Africa by the Overseas Unit of TRRL7’8 h co~aboration with the World Bank to improve knowledge of the effects of road conditions on vehicle operating costs and the rates of deterioration of road pavements. As a result of these field studies empirical relationships were derived that have been incorporated into a computer model (RTIM)9. This model calculates the sum of road construction costs, road maintenance costs, and vehicle operating costs over the ‘design fife’ of the road for a nonurban road project @ a developtig country. Further improvements to these empirical relationships wi~ be forthcoming from World Bank sponsored studies which are in progress in Brazd.10 ~d India and from TRRL studies in the Caribbean. In addition to these field 9 studies, model development is being undertaken by the Massachusetts hstitute of Technology with the objective of producing a model capable of evaluating ‘total transport costs’ for road networks rather than just for sin#e road hks. 3. EARTH AND GRAVEL ROADS Not surpristi@y the engineering of earth and gravel roads has received much less attention from highway engineers and research workers thm the engineering of surfaced roads. Most of the technical literature on unsurfaced roads is concerned with the techniques and the organisation of maintenance, and the selection and specification of naturrd materials for constructing gravel roads. Several methods have been pubtished for designing the thickness of unsurfaced roads on a structural design basis, and three of these methods are described below. However, the relevance of the concept of structural design to unsurfaced roads is not widely accepted. In recent years there has been a growing awareness of the need to improve knowledge of the relationships between the construction and maintenance standards, the climate, the traffic, and vehicle operating costs on unsurfaced roads. The need arises because of the desire of transport planners to improve the quality of investment decisions in the rural road sector in developing countries. A considerable amount of unrecorded local knowledge about the interaction bet ween many of these factors efists in developing countries, but quantified data is very scarce. 3 3.1 Earth roads The distinction between earth and gravel roads is rarely well-defined, but a commonly accepted distinction is tit earth roads are constructed ody of the natural materials that are encountered on the road line or immediately adjacent to it. Gravel roads on the other hand may incorporate imported material, normally a selected natural gravefly material, but sometimes processed gravels maybe used. The best service wfll be obtained from earth roads if they are located on the better drained parts of the terrain, and on the more granular sods if any choice of soil type is available. In most situations skifl in locating earth roads in the landscape is thus fundamental to their subsequent performance. Often the location of earth roads is undertaken on the basis of a brief reconnaissance of the ground by an engineer, but in complex terrain much better results wdl be obtained by using terrain evaluation techniques 11,12,13 assisted by aerial photography and other remotely sensed ground data. Clearly the traffic-bearing abflity of earth roads depends heavfly on the type of sod forming the running surface, and on the prevatig moisture conditions. The abflity of earth roads to carry traffic can be substantially enhanced if they are ‘engineered’ by such measures as raising the formation in low-lying areas, clearing trees and shrubs we~ back from the road so that the sun and wind can more readily dry out the road surface when it is wet, cambering the surface, and cutting suitable side-drains. The principles of good construction and maintenance practice for earth roads are illustrated in Figure 1. In favorable circumstances earth roads can carry substantial volumes of traffic, as is evidenced for example by the behaviour of ‘sabkha’ roads 14 in certain arid areas, and the roads built on the ‘red-coffee’ sods in Kenya. A comprehensive guide to the construction and maintenance of earth roads has been written by MeUierl 5. More recently some interesting research on the trafficabitity of soils has been undertaken at Vicksburgl 6, the results of which are illustrated by the nomograph shown in Figure 2. This nomograph indicates the number of repetitions of wheel loads of different magnitudes that can be carried by soils of different strengths. The definition of the terminal condition in this work was when the rut-depth in the soil exceeded 75 mm. If the strength of an earth road is known (in terms of its California Bearing Ratio (CBR) value), the nomograph permits predictions to be made of the load carrying abflity of the road. 3.2 Gravel roads There are two basic attitudes to the design of the thickness of gravel roads. One of these is the quantitative 15 and Ha~ttl 7. This structural design approach, examples of which are the methods suggested by Mellier approach presents the designer with the difficult problem of deciding what moisture content should be assumed for the subgrade soil and the gravel layer, since clearly in seasonal climates there wfll be a large change in moisture content and hence the strength of these materials between the wet and dry seasons. The other approach to the design of gravel roads is that described by O’Reifly and ~ardl 8 in which the assumption is made that thickness design of gravel roads is unnecessary, and the important issue is the selection of gravel material with specified grading and plasticity characteristics which is placed in a layer of standard thickness (150 to 200 mm). This approach is the one usually adopted in developing countries in the tropics in which 4 subgrades under adequately maintained gravel roads are commody strong enough for the greater part of the year to support the imposed wheel loads with no more than 150 mm of gravel cover. Table 2 gives values for the plasticity characteristics of gravel surfacing materials typically specified in different climatic zones. In either case provision has to be made for the replacement of gravel that is lost over a period of time due to the action of traffic and weather. TMLE 2 Plasticity characteristics preferred for gravel surfacings 8 Ctimate Liquid limit not Plasticity index Linear to exceed (%) range (%) shrinkage range (%) Moist temperate and wet tropical 35 4–9 2.5–5 SeasonMy wet tropical 45 6–20 4–lo Arid or semi-arid 55 15–30 8–15 h practice the choice of gravel material is limited by what is avtiable, and economy in design depends on the sW1 and resources that can be deployed to locate the very best materials that exist within an economic haul distance from the road. The terrain evaluation approachl 1‘12>13 can be of great assistance in the location of gravel materials wittin the landscape. In areas with stron~y seasonal chmates, it may not be economically feasible to bufld gravel roads on plastic subgrades that are capable of sustaining traffic throughout the wet season without excessive deformation. In these circumstances the level of service provided can usudy be maintained at an acceptable level by increasing the frequency of grading, or in some areas where this is not possible, roads are closed for short periods after heavy and prolonged rain to prevent traffic causing excessive deformation when the subgrade is saturated. Maintenance has a very strong influence on the performance of gravel roads. At traffic flows of more than 30 or so vehicles per day most gravel roads require grading at regular intervals of time to remove corrugations, to restore the transverse profile, and to bring back into the centre of the road gravel that has been thrown to the sides by the action of trafficl 9. In addition, the clearing of side ditches and the restoration of material removed by erosion are required at regular intervals. ~ these operations are necessary if the level of service provided by a gravel road is to be maintained, but perhaps the crucial operation is the maintenance of an adequate camber. Camber must be sufficient to prevent rainwater being retained on the surface of the road. If it is not, the road wti deteriorate very rapidly in wet. weather. 3.3 Recent developments The tectiques of designing and constructtig earth and gravel roads have not cbnged significantly for many years. There are, however, two recent developments that are of some interest. One of these is the growing interest among highway planners and engineers in quantifying, for a particular cfimatic zone, the relationships between traffic volume, construction standards, rate of deterioration, and maintenance. fiowledge of these relationships is 5 required if cost-effect decisions are to be made about the construction of unsurfaced roads, or economic maintenance strategies are to be devised. In a number of developing countries empirical relationships have been devised in the past that fink the frequency of grading necessary to maintain an acceptable level of service to the volume of traffic, and also in some cases to the type of gravel surfacing. Only recently, however, have quantified relationships of this kind been based on systematic measurements of road surface conditions, or included quantification of the effect of surface condition on vehicle operating costs. For example, in the study in East Africa 78‘ relationships tig the fo~owing variables were established for several types of gravel road: (a) the volume of traffic; @) the longitudinal roughness of the road surface; (c) the depth of ruts formed by the traffic; (d) the amount of gravel ‘lost’ from the road surface due to the action of traffic and erosion; (e) the longitudinal gradient of the road; (~ the rainfa~. Further studies are in pro~ess that WWimprove these relationships and will extend them to other materials and cfimatic zones. A second recent development in the field of earth and gravel road construction is the widespread renewed interest in employing labour-intensive methods of construction in developing countries. In several Asian countries hbour-titensive methods have traditiondy been employed on a large scale for citi construction projects. In many other developing count ries, however, the construction of roads by labour-intensive methods has not hitherto been undertaken except on a sma~-scale and piecemeal basis. Recent studies sponsored by the World Bank*” and the htemationrd hbour Organisation 21 have shown that labour-intensive methods can in general produce the same qudty of construction as equipment-intensive methods, and that they are economica~y justifiable in countries where the shadow ddy wage* for labour is less than about US $1.50 to $2.00. This is the situation in more than forty countries, in several of which national prograrnrnes for labour-intensive road construction have recently been started or are planned. 4. BITUMEN SURFACED ROADS 4.1 &went desi~ methods Very few of the various methods of pavement design that are in general use throughout the world have been devised specific~y for the design of pavements in tropical developing countries. Two exceptions are TRRL Road Note 3122, and the French CEBTP design manual for tropical countries 23 . Other popular methods of pavement design, such as the AASHTO method 24 and its derivatives, though derived empiricdy in industriahsed countries with temperate climates, are nevertheless often used for the design of pavements in the tropics. * In many developing countries the wages paid often do not reflect the true value of labour to the economy. It is this latter which is relevant when comparing labour and capital intensive methods and it is termed the ‘shadow wage’. For further amplification see Ministry of Overseas Development – Guide to the Economic Appraisal of Projects in Developing Countries. hndon, 1972 (H M Stationery Office). 6 Comments on the appropriateness of these methods for designing pavements in tropical developing countries are made below. 4.1.1 TRRL Road Note 31: This pavement design guide gives recommendations for the design of bitumen-surfaced roads carrying up to 2.5 mi~on equivalent standard axles per lane in tropical and sub-tropical countries. The guide offers the designer the choice of two simple standard pavements, and provides for the thickness of the sub-base to be varied to suit the strength of the subgrade. Particular attention is given in the guide to two aspects of pavement design that are of special importance in most developing countries: (1) detafled consideration is given to the influence of tropical ctimates on moisture conditions in road subgrades; (2) attention is drawn to the advantages of adopting a stage construction approach to road budding in situations where traffic growth rates are high or longterm predictions are uncertain. The guide expresses the strength of the subgrade sofl in terms of its CBR measured at a moisture content equal to the wettest moisture condition likely to occur in the subgrade’ after the road has been constructed. Traffic loading is expressed in terms of equivalent standard axles, on the same basis as that used in the MSHTO method. The expression used by the Overseas Unit, TRRL, for defining the equivalence factor of any atie-load traversing typical roads in developing countries is given by: k) 4.55 ~= L — s ie one application of the afle-load L causes the same amount of structural damage as N applications of the standard tie-load Ls (8160 kg). The guide, which has been derived empiricdy, suffers the disadvantages that afl such methods share, namely that it cannot be applied with confidence to design situations that lie outside the range of conditions within which it has been derived. Nevertheless it does offer the designer a wide choice of pavement material options and is suitable for the design of medium and hghtly trafficked roads in any tropical or sub-tropical environment. 4.1.2 CEBTP pavement design manual for tropical countries: Ttis manual is widely used in French-speaking developing countries in the tropics. Traffic is categorised into four classes and a catalogue of four basic pavements is recommended to match these. Subgrade strength is expressed in terms of CBR, and the thickness of the base and sub-base of each class of pavement can be varied within timits to take account of differences in subgrade strength. The categorisation of the traffic loading can be made in two ways: (1) on the basis of the average volume of traffic per day (all vehicles) assuming a design life of 15 years and 30 per cent ‘heavy’ veticles in the traffic stream; 7 (2) on the basis of the cumulative number of ‘heavy’ vehicles passing over the pavement during its design life. A heavy vehicle is defined as one that has a total weight of more than 3 tonnes. The guide also suggests that if more than 10 per cent of axle loads are greater than 13 tonnes extra pavement thickness may be required. The manual, which is very simple to use, is essentially a modification of the original CBR method of design. It does not attempt to differentiate between axle loads of different magnitudes, and hence it cannot take account of the big differences in damaging power that can exist between similar volumes of traffic in different countries, or even on different roads within the one country. 4.1.3 AASHTO interim guide: This guide, which is based on empirical relationships derived from the AASHTO Road Test25 , was produced primarily for the design of pavements in the USA. Nevertheless it is sometimes used, not always very appropriately, for the design of pavements in developing count ties in the tropics. h the guide the predicted distribution of the axle loads that WN t raverse the pavement throughout its design hfe is expressed in terms of the equivalent number of 8.2 tonne ‘standard’ afles. This is achieved by the application of ‘equivalence’ factors (see reference 56). These factors relate the damage done to the pavement by an afle load of any magnitude to the damage done by a standard 8.2 tonne axle load. The factors follow approximately a ‘fourth power law’; that is, if an a~e load is doubled, the damage it wi~ inflict on the pavement wfll be sixteen times greater. This concept of atie load equivalence factors is now widely accepted, ahd is incorporated into several other methods of pavement design. However, there is a great deal of uncertainty about the magnitude of the power in the ‘power law’, and the extent to which it is influenced by overall pavement strength and the types of materials used in the pavement. The AASHTO guide defines subgrade strength in terms of a ‘soil support value’, ranging from one for very weak subgrades, to ten for subgrades as strong as crushed rock base material. Pavement thickness is expressed in terms of a Structural Number (SN), ranging from 1.0 to 6.0, which is an index of the strength of pavement required. This is then modified by the application of a Regional Factor which adjusts the design to suit local climatic conditions. Finally, layer thicknesses are determined, the contribution of the layers of the selected pavement materials to the pavement strength is estimated by applying coefficients (sometimes called ‘layer equivalences’) that have been assigned to the commoner materials on the basis of the results of the AASHTO test and subsequent experience. The problem facing a designer applying the AASHTO guide to the design of a pavement in a typical developing country situation is the difficulty in estimating the appropriate Regional Factor to use, and in determining the appropriate coefficients to apply to the pavement materials available. The same problem arises whenever an empirically derived design method is apphed in conditions outside the range of experience on which it has been based. The range of Regional Factors employed in the USA is large, and is rather arbitrarily applied. In dry frostfree regions the Regional Factor used can be as low as 0.5, whilst in regions where subgrades freeze in the winter but thaw out in the spring the Regional Factor used may be as high as 4.0 or 5.0. This range represents an increase 8 in thickness of 50 per cent for pavements with a SN of 2.5 or less, typical of most surfaced roads in developing countries. h South Africa, where a modified form of the AASHTO design method is employed and the climate ranges from moist sub-tropical to arid, Regional Factors in the range 0.3 to 0.75 are recornmended26. h Mexico a Regional Factor of 0.2 is used in semi-arid situations. As regards the appropriate coefficients to apply to different pavement materials, a considerable amount of information exists about the commoner pavement materials used in the USA, such as asptitic concrete surfacing materials and crushed rock and gravel base materials. In developing countries in the tropics the indigenous pavement materials avaflable will often not have any close parallel in North America and designers using the AASHTO method will have to rely on judgement in assigning coefficients to the materials selected. 4.1.4 Shefl modified design charts: The shell design charts were first pubhshed in 1963 27, but have been modified subsequently to take accollnt of the effects of high road temperatures on the moduh of bituminous materials28. They are therefore now better suited to the design of road pavements in developing countries in the tropics. The charts were derived from consideration of the basic properties of bituminous mixes as detertied by mechanical tests in the laboratory, and the Mowable strains in the subgrade and asphalt surfacing layer. The analytical approach to pavement design, which is the basis of both the original Shefl design charts and the more recent more comprehensive method 28, is fully described in reference 56. The original charts are held to have been we~-supported by the evidence of the performance of roads, but the recently modified charts have yet to be similarly vatidated. In using the modified She~ charts it is necessary to estimate the weighted mean annurd air temperature (MAAT) of the location in question. This can be obtained from mean montMy air temperatures by use of a given weighting curve. Simple meteorological information of this kind can usua~y be obtained in most countries, or can be estimated with adequate accuracy. 4.1.5 Other methods: Various other methods of pavement design are used from time to time in developing countries. The most significant of these are RRL Road Note 29 29, the Asphalt Institute method30, and the Qnadian Good Roads Association method 31 . None of these methods is well suited to the design of roads carrying less than about 500 vehicles a day, which form the majority of the surfaced roads in most developing countries. Moreover, wtist these methods of pavement design can be used to design heavily-trafficked roads in developing countries, they do not have the facility for adjusting designs to take account of different cfimatic environments. 4.2 Compan.son of pavement desi~ methods Vafid comparisons between the different methods of pavement design described above are difficult to make. If common factors of safety and criteria for pavement faflure were shared by au the methods, a comparison on the basis of the recommended thicknesses of simdar materials would be a measure of the relative ‘efficiency’ or design economy of the different methods. 9 In fact, the faflure criteria and factors of safety inherent in the various methods are all different, and in several methods they are not even clearly defined. Nevertheless some measure of the relative efficiency of the different design methods can be gained by comparing the pavement thicknesses recommended by the various methods for simflar levels of traffic loading and subgrade strength. k order to make these comparisons it is helpful to express pavement thicknesses in terms of a ‘structural number’, simflar in concept to the SN used in the AASHTO design method. In the AASHTO method SN is given by: SN = alD1 + a2D2 + a3D3 where al, a2, and a3 are layer. coefficients, and D1, D2, and D3 are layer thicknesses in inches. It should be recognised that assigning coefficients to particular materials can only be a very rough guide to the relative contribution different materials can make to the structural strength of a pavement. It is well known that the effective coefficient of a pavement layer is not only a function of the type of material but is also dependent on the thickness of the layer, the position of the layer in the pavement, the moduli and thicknesses of the other layers, the subgrade characteristics, and other factors. There are thus many reservations about the concept of layer coefficients, but nevertheless it wifl suffice for the purpose of making a rough comparison between current methods of pavement design. Because the service temperatures of bituminous surfacings in most developing countries in the tropics are much higher than those experienced in the AASHO Road Test, it is necessary to assume different layer coefficients for bituminous layers from those given in the AASHTO guide if even rough comparisons between different design methods are to be made in the context of tropical environments. The AASHTO guide suggests a coefficient of between 0.44 and 0.40 for asphdtic concrete surfacings, but in a typical developing country situation it is assumed for the purposes of this comparison that 0.20 would be more appropriate. This assumption is made on the basis of the temperature/stiffness relationship for bituminous mixtures implicit in the modified Shell design charts , 28 from which can be deduced the relations shown in Figure 3 between MAAT and layer coefficient. From Figure 3 it can be seen that in tropical chmates the effective layer coefficient for asphalt surfacings ranged from about 0.10 for CBR 20 subgrades in the hottest climates, to about 0.37 for CBR 2.5 subgrades in sub-tropicrd cfimates. Using values of 0.2 for al, and the same values for a2 and a3 as are given in the AASHTO guide (eg granular base a2 = 0.14, sub-base a3 = 0.11 ), the modified structural numbers imptied by the different design recommendations are compared in Figure 4. For convenience the modified structural number is plotted as ‘Tropical Structural Number’ (TS~, and Figure 4 shows how this is related to traffic loading on an average subgrade of CBR 8. h calculating the TSN values that are plotted in Figure 4 it was assumed that the pavements had the minimum thickness of bituminous surfacing permitted by each design method. This minimises the influence on TSN of the value adopted for the coefficient a ~, and is typical of developing country pavements which rarely have thick bituminous surfacings. It is also assumed in Figure 4 that the subgrade strengths are assessed in the same way for each of the design methods. In fact each method su~ests different ways of estimating the appropriate subgrade moisture content and 10 density at which to assess the subgrade strength. These variations increase even further the differences between designs produced by the different methods shown in Figure 4. the Simfiarly, each design method suggests rather different ways of expressing traffic loading. Most methods convert the predicted afle load distributions into an equivalent number of standard ties, but there are considerable differences in the factors used for this. Accepting these reservations about the basis of this comparison, it can nevertheless be concluded that there are very large differences between the designs produced by the various methods of pavement design, even when the same assumptions of subgrade strength and traffic loading are made. The TSN is in effect an index of pavement thickness, and it can be seen from Figure 4 that the more conservative design methods recommended pavements rdmost twice as thick as the least conservative. It can dso be seen that the methods of pavement design issued specifictiy for use in tropical developing countries produce the most economical designs. This partly reflects the different standards of serviceabfity and failure criteria that are inherent in the different methods, but it also reflects different modes of ftiure and the absence of the damaging effect of frost in the tropics. The comparisons made in Figure 4 demonstrate the degree of uncertainty that etists about the design of more heavily trafficked road pavements in the tropics, and points to the need for more research in this field. 4;3 Recent developments 4.3.1 tie loads: In afl countries increases in the volume of traffic and the size and weight of vehicles over the last two or three decades have necessitated the butiding of much stronger road pavements than those that sufficed earlier. In developing countries the rate of growth of traffic loading has been most dramatic. In part this is a reflection of the relatively sma~ number of commercial vehicles that were in use in Third World countries twenty years ago, but it also reflects the absence of effective control of the weights and afle loads of vehicles in many count ries. Indeed in some countries there is no restriction at au on ade loads. As a result of the high rate of growth of traffic loading in developing countries, it is not uncommon for roads to fti prematurely because the traffic loading in service is several times greater than was assumed in the design. In recent years a considerable amount of information has become available about the de loads on the roads in a number of developing countries 32’33’34. This has shown the strong influence that a relatively small number of vehicles with M@ atie loads can have on the rate of pavement damage of a typical main road in a developing country. For example, it is possible to envisage a situation in which the regular use of as few as one or two hundred vehicles with 20 tonne tie loads could severely damage a significant proportion of the paved road network of a developing country within a few months. This is because when overafl volumes of traffic are low and tie loads are modest, the introduction into the traffic stream of a small proportion of vehicles with 20 tonne tie loads can double or treble the damaging power of the traffic. It is thus very important to control the upper hrnit of tie loads in developing countries. This conclusion arises directly from the assumption that an afle load of 20 tonnes is rou@y 300 times as damaging as one of 5 tonnes. This is the ratio given by atie load equivalence factors derived from the AASHO Road Test that are used in most methods of pavement design in current use. However, there are considerable 11 doubts about the validity of the AASHO Road Test equivalence factors for use in developing countries. They may under-estimate or over-estimate the damaging effect of heavy tie loads. The main reasons for these doubts are: (1) (2) (3) (4) The maximum afle loads used in the AASHO Road Test were 13.6 tonnes for sin@e axles, and 21.8 tonnes for tandem axle sets. In developing countries, sin~e ade loads in excess of 20 tonnes are not uncommon, hence it is necessary to extrapolate well beyond the AASHO test results to derive equivalence factors for these loads. There are considerable difference: between the various interpretations that have been made of the AASHO Road Test results. ~lst hddle’s analysis35 is most widely used, the analysis made by Shook and Finn 36 is also used to a considerable extent and these two interpretations give quite different results for the damaging effect of very heavy axle loads. Frost was a critical factor in the deterioration of the AASHO Road Test pavements, but it is a ne~igible factor in most developing countries. In the interpretation of the results an attempt was made to separate the effects of frost on pavement deterioration from the traffic-induced effects, but this was not wholly successful. The Road Test pavements were constructed on a weak subgrade, whereas most pavements in developing countries are budt on strong subgrades. Clearly there is a need to reduce the uncertainty that exists about the damaging power of heavy tie loads in typical developing country situations. Eventually the analytical pavement design approach will be developed to the point where consideration of empiricdy derived tie load equivalence factors wifl become irrelevant. However, this goal is stifl some way off, and in the shorter term research such as that being undertaken in South Africa using heavy vehicle simulators37 offers the best chance of providingpractical answers to these and other pavement design problems. In recent years the world-wide growth in the movement of freight by road, and the production of a new generation of larger and more powerful commercial vehicles, has prompted governments in many countries to re-exarrdne the maximum tie loads permitted on their national road networks. In general, the larger the vehicle and the heavier the aAe load, the cheaper is the tonne -tiometre operating cost. In both developing countries 38 and industriahsed countries 39 attempts have been made to deduce the optimum axle load at which the combined cost of vehicle operation, road construction, and road maintenance is minimised. Figure 5 indicates the nature of the relationships between axle load and the tonne-tiometre costs of vehicle operation and road construction and maintenance. There are of course big differences between the ideal ‘optimum’ axle load, the legal ade load hrnit, and the range of afle loads that are actually applied to a road system. The influence of the legal tie load limit on the actual distribution of ade loads naturafly varies from country to country, depending on the degree of enforcement and other factors. Clearly this aspect must be carefufly considered when making an analysis of the merits of a particular legal axle load limit. Simflarly if an increase in legal afle load fimit is being considered careful thought must be given to the effects on existing bridge structures and roads; it may we~ be that the costs of strengthening structures and adding expensive overlays to existing roads outweigh the benefits that would accrue from reduced tonne-tiometre operating costs. 12 Nevertheless, studies in some countries, notably Austraha w , have concluded that raising afle load limits would bring substantial national economic benefits. It is probable that simflar conclusions could be reached in several developing countries, and that in the future legal axle load limits in the Third World WU tend to rise, but that in pardel there wi~ be stricter enforcement of these fimits. The Overseas Unit of TRRL has developed a portable weighbridge, and has gained experience with it in tieload surveys in many parts of the World. The results of this experience have been incorporated into TRRL Road Note 4041, which gives guidehes for carrying out axle-load surveys, using the weighbridge, on paved roads in developing countries. 4.3.2 Heavy vehicle simulator: An important development in the field of pavement engineering in recent years has been the buflding of four Heavy Vehicle Simulators (HVS) in South Africa 37 . Whilst the contribution these machines are Wely to make to the advance of pavement design is of world-wide interest, the fact that they will be used, initiafly at least, in the sub-tropical climate of South Africa, will be of special interest to those concerned with improving pavement design in developing countries in the tropics. Repeated loading tests of pavement materials at laboratory scale as described by Brown57 has an important part to play in the development of a dependable analytical method of pavement design. However, many research workers believe that the essential vahdation by fu~-scale experience of theoretical models based on laboratory experiments udn be greatly assisted by the use of intermediate forms of testing of the ‘road machine’ type. Indeed some research workers befieve that without this intermediate step very tittle further progress can be made towards a reliable mechanistic method of design. The great asset of ‘road machines’ is that they apply the load to test pavements through a rolling wheel, and hence they induce the same sort of changes in the directions and magnitudes of the stresses experienced by all elements of the pavement as are experienced in normal service. Several different types of road machine have been used for pavement design research in the past, but the HVS is the first fun-scale machine that can be moved easfiy from place to place and can be used to load in situ both ordinarily-constructed roads as wefl as specia~y-built fu~-scde experimental roads and pilot scale pavements. The HVS can apply a load of up to 100 kN through a dual or singe wheel assembly which is traversed over the pavement for a distance of 8m at a speed of up to 14 km/hour. The wheel assembly reciprocates in a heavy steel frame and can apply up to 1400 repetitions of load per hour to the pavement under test. It is thus possible to apply hdf a m~on repetitions of load to a pavement in 20 to 30 days. The prototype HVS has been in use for some time and some limited but useful results have been obtained. It has been used to verify a model that predicts the initiation of traffic-induced cracking in cement -treated roadbases. The HVS proved to be successful for this purpose and it enabled load equivalence factors for a limited number of examples of this type of pavement to be studied 42 . It was found that a ‘power law’ of 6.0 to 6.7 applies to such pavements, as opposed to the power of 4.0 to 4.5 that has been generally assumed hitherto. This result has considerable significance for pavement engineering in tropical developing countries where cement bound materials are more commonly used for roadbases than in most industriafised countries. 13 4.3.3 The design of bituminous surfacings for high service temperatures: The large increase in recent years of the length of road in tropical developing countries that requires a premixed bituminous surfacing has focused the attention of highway engineers on the need to improve the design of surfacings for tropical conditions. For many years bituminous surfacings of the asphaltic concrete type have been widely used in hot climates, and the mix design procedures for achieving high stability mixes are well-established. Experience has shown, however, that such mixes, whilst having a high resistance to deformation, are prone to cracking and are difficult to compact. It is now recognised that even for high temperature service conditions high stability is not a~-important, and that more attention should be given in designing mixes to the properties of flexibility, durability and workability y. The high stability of asphaltic concrete is mainly dependent on aggregate interlock and inter-particle friction which are enhanced by the careful proportioning of coarse and fme aggregates to produce a continuously graded mixture. h Britain, with its temperate climate, such mixes have never been popular and virtua~y all heavy-duty bituminous wearing courses are composed of a gap-graded type of mix known as hot rolled asphalt. In effect this is a stone-fdled sand asphalt in which the stiffness of the sand-fdler-binder mortar provides the overall stability of the mix. ~ilst such mixes are more prone to deformation than continuously graded mixes, they are more tolerant of variations in mix composition, more flexible, and easier to compact. At British temperatures gap-graded mixes have proved to be very satisfactory, and interest is growing in the use of such materials for road surfacings in the tropics43. It has been found that modified forms of hot-rolled asphalt can be produced with adequate stabihty to resist deformation at high temperatures without sacrificing the advantages of this type of mix 28’M. To enhance the stability of these mixes it is desirable to use a stiffer bitumen (60/70 pen, or 40/50 pen if avaflable) than that normally used for asphaltic concrete (80/100 pen). The problems of mix design for both gap-graded rolled asphalt and continuously-graded asphaltic concrete are fully discussed in reference 58. For less-heavdy trafficked roads in developing countries, which nevertheless require a premix bituminous surfacing, bitumen-macadams are being used successfully. Dense bitumen-macadam is virtua~y the same as asphaltic-concrete except that it has a rather higher voids content. Open-textured bitumen-macadams, however, have a much higher voids content and are permeable. For situations where impermeability is not required they provide a low-cost mix which is easy to make and lay, and which has a wide tolerance to variations in mix composition. They are thus well suited for overlaying lightly trafficked roads that still retain an impermeable surface but require a levelling or friction course. In many developing countries the most economic and appropriate surfacing for the majority of paved roads is a single or multiple surface dressing 45 . Unfortunately in many countries the difficulty in achieving adequate cent rol of the surface-dressing process means that the lives of surface dressings are much shorter than they could be. Great difficulty is experienced in maintaining binder distributors so that they can apply an even film of binder at the required rate of spread. Similarly it is very difficult to maintain adequate standards of aggregate size, shape, and cleanness. Some of the effects of these deficiencies can be minimised by such techniques as the pre-treatment of the chippings with a light coating of tar, creosote or diesel to suppress dust and to aid adhesion. Deficiencies in binder distribution are virtually impossible to counter, though they can be mitigated to some extent by the application of a light ‘fog-spray’ of binder and a sprinkling of sand or rock dust after the chippings have been spread on the main binder film. This technique has value whether or not the main binder film has been sprayed satisfactory. 14 Slurry seals46, used either alone or in conjunction with a surface dressing, provide very satisfactory surfacings for medium and lightly trafficked roads in many developing countries. 4.3.4 Overlay design: There is no difference in principle between the process of designing strengthening overlays for roads in developing countries and in industriafised countries 47 . There are, however, differences in the appropriate methods to use for assessing residual pavement strength and in the stiffnesses of the pavements being strengthened. In developing countries Benkelman beams are very appropriate for assessing the residud strength of existing roads. More sophist icat ed equipment such as the deflectograph can equafly we~ be used, but in many cases the cost of such equipment and the difficulty of servicing and operating it in remote areas, WMoutweigh the advantages it offers as compared with the use of deflection beams. It has been found that wefl-trained teams equipped with two deflection beams and a loaded truck can achieve rates of survey adequate for a typical road strengthening programme in a developing country, and survey techniques have been developed for this purpose48. The effect of temperature on the stiffness of bituminous materials needs to be carefufly considered when undertaking deflection surveys in the tropics. For example, in Britain it is recommended that deflection measurements are not made when road surface temperatures rise above 30°C, and all measurements of deflection are corrected to their equivalent value at a standard reference temperature of 20°C. In many developing countries road surface temperatures rarely, if ever, fa~ below this temperature in dayfight hours, thus it is necessary to establish by investigation, the typical local deflection/temperature relationships and to select the appropriate standard reference temperature. Tentative deflection criteria relating deflection to the subsequent traffic-carrying abfity of the pavement have been suggested for typical developing country pavements 21 . Further research is needed to extend and vatidate these criteria in different conditions. The thickness of overlay required to reduce pavement deflections to the required ‘design’ level is best established by local experience, but if this is lacking the curves derived from British experience49 can be used as a guide, suitably modified by judgement to take account of differences in materials and service conditionsso. 4.3.5 Impact (square) ro~ers: The idea of an impact roller was first considered in the 1930s and five-sided prototype machines were built in the early 1950s. Athough these machines were used successfu~y in the 1960s5 1 to compact material in thick layers, the machines were impractical because of the problems created by the widely fluctuating loads at the drawbar. The Council of Scientific and Industrial Research (CSIR) in South Africa52 have studied and solved these problems and have developed a four-sided ‘square’ rofler of 8000 kg or 10000 kg weight. Its compactive effort is derived from the energy of the mass fafling from the corners to the faces w~st being towed at about 12 km/hour. This compactive effort is m~y times greater than that obtained from the static mass of the ro~er. The ro~er can be raised free of the ground onto carrier wheels to pass over bridges or culverts which might be damaged by the impact blows. 15 The main advantage of impact rollers is their ability to compact non-cohesive materials in thick layers and at low moisture contents. A considerable amount of experience has now been gained in the compaction of uniform Kalahari sands with impact roflers. It has been shown that these rollers are very effective in compacting material below 0.5m and up to depths of 4m below the surface. The upper 0.5m may sometimes be loosened by the impact compaction process, and must subsequently be compacted by other machines. In very loose materials compaction may have to be carried out through a blanket layer of more cohesive material. In a series of tests52 to compare the effectiveness of impact and vibratory rollers on a uniform marine sand it was suggested that the depth influence for the impact roller was about 3m compared with 1.8m for the 4500 kg vibratory roHer and 2.2m for the 9000 kg vibratory roller. It has also been su~ested that a 15000 kg vibratory roller may give a performance comparable to a 10000 kg impact ro~er. The main disadvantage of impact compaction is the surface deformations produced by the roller and the loosening of the top layers in certain materials. For this reason this compaction technique is only suitable for earthworks and subgrade preparation and not for the compaction of pavement layers. 5. CONCRETE ROADS Concrete roads are uncommon in developing countries for a variety of reasons. For instance in many countries cement has been in short supply for many years and road construction projects are often long distances from the nearest cement factory. Hence cement, if it is obtainable, is likely to be relatively very expensive by the time it is transported to site. Also in developing countries the construction of roads in stages makes sound economic sense, and concrete construction is not well-suited to stage construction. Furthermore, the advantages of concret e pavements over flexible pavements are most marked on weak subgrades, whilst in general, subgrades in tropical developing countries are relatively strong. For these and other reasons very few concrete roads have been built in developing countries. On the other hand, the rapid rise in the price of bitumen in recent years may prompt a renewed interest in concrete road construction in those countries that have adequate supphes of cement. It is also possible that more concrete roads may be buflt if they prove to be well-suited to labour intensive methods of construction, as seems possible. 6. ASSESSING RISKS 6.1 Decision wking Formal decision analysis is not widely used in highway engineering, nor indeed in civil engineering as a whole. This may be because in most civd construction much emphasis is placed on design against faflure, rather than on design for the optimum use of available resources. This attitude is perfectly legitimate in cases where the failure of a structure would be catastrophic, such as the failure of a dam or a major bridge. There are other types of civil engineering construction, however, in which ‘failure’, if it occurs, is not catastrophic, and the costs of accepting different levels of risk of failure can be quantified and should be inherent in the design process. Highway pavements are in this category since pavement ‘failures’ rarely have disastrous economic or social consequences, and seldom result in a road becoming completely unusable by vehicles. Indeed even the definition of the ‘failure’ of a road 16 pavement is very arbitrary, and is usually expressed in terms of a certain degree of cracking or surface deformation. Formal decision analysis, which provides a means of balancing technical and financial risks, should therefore be part of the pavement design process. In the process it may be necessary to make subjective as well as quantitative assessments of uncertainties53, but this is much better than ignoring entirely the lack of precision that is inherent in pavement engineering, and it is tikely to result in a more efficient use of the available resources. 6.2 Improved forms of contract k most developing countries road construction is undertaken both by direct labour and by contract. The proportion of the total that is carried out by contract varies considerably from country to country, depending on the size and efficiency of the direct labour organisation and the magnitude of the national road construction and maintenance programme. The tradition forms of contract used in industrialised countries are often inappropriate for road construction” projects in developing countries, where the degree of uncertainty about many aspects of a project may be high. W@ risks mean high contract prices, and clearly it is in the chents’ interest to apply a system of contract that can operate efficiently and fairly in a situation of uncertainty without incurring excessive costs. Recent research54 that has been undertaken on appropriate forms of contract for use in conditions of uncertainty is particularly relevant to road construction in developing countries. This research indicates that an adaptation of the target price form of contract is likely to be most suitable. Target price contracts are essentially cost-plus contracts with specific incentives buflt into them to encourage afl parties to reduce costs. The form of contract tends to compel the client to make a detafled appraisal of the true objectives of a project and the risks likely to be encountered in its execution. Such an approach is likely to focus the client’s attention on policy issues guch as the appropriateness of specifications and the use of low-grade materials. 7. FUTURE NEEDS There is an urgent need to improve knowledge of the many factors that influence the cost of transport on earth and gravel roads in developing countries. Better knowledge is required of the interaction between vehicles and the roads on which they run, with a view to balancing both vehicle design and road design so as to minimise total transport costs. Improved quantification of the rates of deterioration of earth and gravel roads is necessary in order to assess maintenance needs in economic terms, and to plan cost-effective maintenance strategies. In quantifying the rat es of deterioration of unsurfaced roads it will be necessary to separate the effects of traffic and ctimate, and to define better those surface characteristics of unpaved roads that have the greatest influence on vehicle operating costs. The influence on total transport costs of the method of construction of roads, whether plant-intensive, intermediate, or labour-intensive, also merits investigation. h the field of paved roads the development of an improved method of pavement design appropriate to developing countries in the tropics is certainly needed. Initia~y the extension of the existing methods will no doubt be undertaken, assisted by laborato~-scale studies of pavement materials and full-scale pavement studies using accelerated loading devices such as the heavy vehicle simulator. In the longer term, however, a universal analytical method of pavement design is required, sufficiently well calibrated to enable it to accommodate the wide range of environmental conditions and materials that are encountered in developing countries . 55 17 The interaction between initial construction standards and subsequent maintenance merits further attention, as does the extent to which maintenance can counter pavement deterioration. There is dso a need to develop appropriate specifications for locafly-occurring road-building materials, many of which may not meet existing specifications, but are nevertheless quite capable of providing satisfactory service. h conclusion, it is necessary for pavement engineers in developing countries to think more in terms of the design, construction, and maintenance of road pavements as being parts of a system, all parts of wtich interact with each other and with the vehicles that use the roads. The function of the pavement engineer is to manage this system in such a way that the service the system supphes to the community over a period of time is as costeffective as possible. 8. ACKNOWLEDGEMENTS The preparation of this Report forms part of the research programme of the Overseas Unit (Head: Mr J N Bulman), of the Transport and Road Research Laboratory. 1. 2. 3. 4. 5. 6. 7. 8. 9. REFERENCES WORLD BANK. World Bank Atlas. Washington DC, 1976 world Bank). 11th Edition. INTERNATIONAL ROAD FEDERATION. Road and motor vehicle statistics for 1975. Geneva, 1976 (hternationd Road Federation). UNITED NATIONS. Statistical Year Book 1975. New York, 1976 (United Nations). 27th Edition. HUDSON, W R. State of the art in predicting pavement reliability from input variability. US Army En~.neer, Watemays Experiment Station Contract Report 5-75-7. Vicksburg, Mississippi, 1975 (US Army Engineer Waterways Experiment Station). ROAD RESEARCH LABORATORY. The cost of constructing and maintaining flexible and concrete pavements over 50 years. Ministry of Transport, RRL Report LR 256. Crowthorne, 1969 (Road Research hboratory). MOAVENZ~EH, F. Investment strate~es for developing areas: andyticd model for choices of strategies in highway transportation. Massachusetts Institute of Technolo~, Department of Civil Eng.neen.ng Research Report No 72-62. Cambridge, Massachusetts, 1972 (Massachusetts Institute of Technology). HODGES, J W, J ROLT and T E JONES. The Kenya road transport cost study: research on road deterioration. Department of the Environment, TRRL Report LR 673. Crowthorne, 1975 (Transport and Road Research bboratory). HIDE, H, S W ABAYNAYAKE, I SAYER and R J WYATT. The Kenya road transport cost study: research on vehicle operating costs. Department of the Environment, TRRL Report LR 672. Crowthorne, 1975 (Transport and Road Research Laboratory). 18 9. ROBINSON, R, H HIDE, J W HODGES, S W ABAYNAYAKA and J ROLT. A road transport investment model for developing countries. Department of the Environment, TRRL Report LR 674. Crowthorne, 1975 (Transport and Road Research Laboratory). 10. REPUBLIC FEDERATIVE DO BRASIL. Research on the inter-relationship between costs of highway construction, matitenance and utilization. Inception Report, Empress Bresileira de Planejamento de Transported (GEIPOT). Brasflia, 1976 (Repubhc Federative do Brasfl). 11. BEAVEN, P J and C J LAWRANCE. The application of terrain evaluation to road engineetig. Hoceedings of the Conference on Road Engineen”ng in Asia and Australasia, Kuala Lumpur, 11 –1 6 June, 1973, sponsored by the Institution of En@.neers, Malaysia and the Public Works Department, West Malaysia. Kuala Lumpur, 1973 (hstitution of Engineers), p7. 12. DOWLING, J W F and F H P WILLIAMS. The use of acrid photographs in material surveys and classification of land forms. (Paper No. 9) Conference on Civil Enp.neen.ng Problems Overseas. bndon, 1964 (kstitution of Citi Engineers) pp 209–36. 13. TRANSPORT AND ROAD RESEARCH LABORATORY. Terrain evaluation for highway engineering and transport planning. A technique with particular value for developing countries. Department of the Environment Department of Transport, TRRL Report SR 448. Crowthorne, 1978 (Transport and Road Resea~ch bboratory). 14. ELLIS, C I and R B C RUSSELL. The use of salt-laden sofls (Sabkha) for low-cost roads. Conference on Lo*cost Roads, Kuwait, 25–28 November, 1974, sponsored by Arab Engineers Federation and Kuwait Society of Engineers. Kuwait, 1974 Kuwait Society of Engineers). 15. MELLIER, G. La route en terre. Paris, 1968 (Editions EyroUes). 16. AHLVIN, R G and G M HAMMITT 11. had supporting capability of low volume roads. Boceedings of a Workshop on low-volume roads held June 16–19, 1975 at Boise, Idaho, USA, Transportation Research Board Special Report 160. Washington DC, 1975 (National Research Councfl, National Academy of Sciences), 17. HAMMITT II, G M. Thickness requirements for unsurfaced roads and airfields, bare base support. US Army Engineer Waterways Experiment Station, Project 3782-65, Technical Report S-70-5. Vicksburg, Mississippi, 1970 (US Army Engineer Waterways Experiment Station). 18. OREILLY, M P and R S MILLARD. Roadmaking materials and pavement design in tropical and sub-tropical count ries. Ministry of Transport, RRL Report LR 279. Crowthorne, 1969 (Road Research hboratory). 19. TANNER, J S. Corrugations on earth and gravel roads: Their formation, treatment and prevention. Rds and Rd Constr., 1958,36 (421), 4–10; (422), 32–5. Inds Trav. d’outre mer, 1958,6 (55), 328–33. (In French). 19 20. HARRAL, C G, I K SUD and B P COWS. Scope for the substitution of labour and equipment in civd construction. Transport decisions in an age of uncertainty. Proceedings of the third World Conference on Transport Research, Rotterdam, 26–28 April, 1977. The Hague, 1977 (Martinus Nijhof~. 21. ALLAL, M and G A EDMONDS. Manual on the planning of labour-intensive road construction. Geneva, 1977 (International bbour Office). 22. TRANSPORT AND ROAD RESEARCH LABORATORY. A guide to the structural design of bitumensurfaced roads in tropical and sub-tropical countries. Department of the Environment Department of Transport, Road Note 31. hndon, 1977 (H M Stationery Office). Third Edition. (Available in French from the Transport and Road Research bboratory). 23. CENTRE EXPERIMENTAL DE RECHERCHES ET D’ETUDES DU BATIMENT ET DES TWVAUX PUBLICS. Manuel de dimensionnement de chausse~s pour les pays tropicaukSecretariat d’Etat aux Affaires Etrang6res. Paris, 1972 (Centre Exp~rimental de Recherches et d’Etudes du B~timent et des Travaux Publics). 24. AME~CAN ASSOCIATION OF STATE HIGHWAY AND TRANSPORTATION OFFICIAN. AASHTO interim guide for design of pavement structures 1972. Washington DC, 1974 (American Association of State Highway and Transportation Officials). 25. ~GHWAY RESEARCH BOARD. The AASHO Road Test, Report 7. Summary Report. Highway Research Board Special Report No. 61 G. Washington DC, 1962 (National Research Council). 26. COUNCIL FOR SCIENTIFIC AND INDUSTWAL RESEARCH. Asphalt pavement design for national roads 1970. National Institute for Road Research Technical Recommendation for Highways TRH 4. Pretoria, 1971 (National Institute for Road Research). 27. SHELL INTERNATIONAL PETROLE~ COMPANY. Shefl 1963 design charts for flexible pavements. London, 1963 (Shefl International Petroleum Company). 28. CLAESSEN, A I M, J M EDWARDS, P SOMMER and P UGE. Asphalt paving design, the She~ method. Fourth International Conference on the Structural Design of Asphalt Pavements, University of Michigan, 22–26 August 1977. fioceedings. Ann Arbor, Michigan, 1977 (University of Michigan). 29. ROAD RESEARCH LABORATORY. A guide to the structural design of pavements for new roads. Department of the Environment, Road Note 29. hndon, 1970 (H M Stationery Office). Third Edition. 30. ASPHALT INSTITUTE. Thickness design – fufl depth asphalt pavement structures for highways and streets, Manual Series No. 1. CoUege Park, Maryland, 1970 (Asphalt Institute). Eighth Edition. 31. CANADIAN GOOD ROADS ASSOCIATION. A guide to the structural design of flexible and rigid pavements in Canada. Canadian Good Roads Association Pavement Design and Evaluation Committee. Ottawa, 1965 (Canadian Good Roads Association). 20 32. ELLIS, C I and F P POTOCKI. Me load distribution on loads overseas: Abu Dhabi and Qatar 1970–71. Department of the Environment, TRRL Report LR 572. Crowthorne, 1973 (Transport and Road Research bboratory). 33. JONES, T E and R ROBINSON. 1975 Turkey traffic survey. (Ankara-Istanbul Expressway): tie loading. Department of the Environment, TRRL ReportLR713. Crowthorne, 1976 (Transport and Road Research bboratory). 34. JONES, T E. Atie loads on paved roads in Kenya. Department of the Environment Department of Transport, TRRL Report LR 763. Crowthorne, 1977 (Transport and Road Research Laboratory). 35. LIDDLE, W J. Application of AASHO Road Test results to the design of flexible pavement structures. First International Conference on the Structural Design of Asphalt Pavements, Universiw of Michigan, 20–24 August 1962. Proceedings. Ann Arbor, Michigan, 1963 (University of Michigan), pp 42–5 1. 36. SHOOK, J F and F N FINN. Thickness design relationships for asphalt pavements. First International Conference on the Structural Design of Asphalt Pavements, University of Michigan, 20–24 August 1962. Proceedings. Ann Arbor, Michigan, 1963 (University of Michigan), pp 52–83. 37. RICHARDS, R G, W D O PATERSON and M D LOESCH. The NITRR heavy vehicle simulator: description and operational logistics. NITRR Technical Report RP/9/77. Pretoria, 1977 (National Institute of Transport and Road Research). 38. ETHIOPIAN ROAD AUTHORITY. The economics of selecting appropriate legal tie load limits. Second Pan African Conference on Highway Maintenance and Rehabilitation, Accra, Ghana, 22–29 November 1977, organised by the Economic Commission for Africa with the cooperation of the Governments of the United Kingdom, France and the Federal Republic of Germany. Addis Ababa, 1977 (United Nations Economic Commission for Africa). Copies also avadable from TRRL. (Avtiable in French). 39. BRINCK, C E. Optimalt axeltryck (Optimum axle loads). Statens Va~.nstitut Meddelande 92. Stockholm, 1966 (Statens Vaginstitut). 40. NATIONAL ASSOCIATION OF AUSTRALIAN STATE ROAD AUTHORITIES. A study of the economics of road vehicle limits – features, recommendations and implementation. Proceedings of the 8th Conference of the Australian Road Research Board, Perth, 1976. Vermont, 1976 (Australian Road Research Board). 41. TRANSPORT AND ROAD RESEARCH LABORATORY. A guide to the measurement of afle loads in developing countries using a portable weighbridge. Department of the Environment Department of Transport, Road Note 40. London, 1978 (H M Stationery Office). 42. PATERSON, W D O. An evaluation of the heavy vehicle simulator as a tool for measuring pavement behaviour. NITRR Technical Report RP/5/76. Pretoria, 1977 (National Institute for Transport and Road Research). 21 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 22 MARAIS, C P. Gap-graded asphalt surfacings: the South African Scene. Proceedings of Second Conference on Asphalt Pavements for Southern Africa, Durban, 29 July – 2 August 1974. Pretoria, 1974 (National Institute for Road Research). GROTH, P J. Overlay design in Natal. Proceedings of First Conference on Asphalt Pavements for Southern Africa, Durban, 1969. Pretoria, 1969 (National Institute for Road Research). NATIONAL ASSOCIATION OF AUSTRALIAN STATE ROAD AUTHORITIES. Principles and practice of bituminous surfacings, Vol. 1, Sprayed work. Sydney, 1975 (National Association of Austrahan State Road Authorities). NATIONAL INSTITUTE FOR ROAD RESEARCH. Technical recommendations for highways TRH 3, bituminous surface treatments for newly constructed rural roads. Pretoria, 1971 (National Institute for Road Research). ELLIS, C I and G LIAUTAUD. The design and construction of strengthening overlays for roads. Conference on Highway EnD.neen’ng in Afn.ca, Addis A baba, 1974, organised by the Economic Commission for Afn.ca with the cooperation of the Bn.tish and French Governments. Addis Ababa, 1974 (United Nations Economic Commission for Africa). SMITH, HR. A deflection survey technique for pavement evaluation in developing countries. Department of the Environment, TRRL Report LR 525. Crowthorne, 1973 (Transport and Road Research hboratory). LISTER, N W and C K KENNEDY. A system for the prediction of pavement life and design of pavement strengthetig. Fourth International Conference on the Structural Design of Asphalt Pavements, Universip of Michigan, 22–26 August 1977. Proceedtigs. Volume 1. Ann Arbor, Michigan, 1977 (University of Michigan), pp 629–48. BULMAN, J N and H R SMITH. Pavement performance and deflection studies on Malaysian roads. Department of the Environment Department of Transport, TRRL Report LR 795. Crowthorne, 1977 (Transport and Road Research bboratory). CLEGG, B and A R BERRANGE. The development and testing of an impact roller. Civ. Engr S. Afr., 13 (3). 1971, CLIFFORD, J M. Impact rolling and construction techniques. Proceedings of the 8th Conference of the Australian Road Research Board, Perth, 1976. Vermont, 1976 (Australian Road Research Board). ELLIS, C 1. Risk and pavement design decision in developing countries. Department of the Environment, TRRL Report LR 667. Crowthorne, 1975 (Transport and Road Research bboratory). PERRY, J G and P A THOMPSON. Target and cost-reimbursable construction contracts – a study of their use and implications. Construction Indust~ Research and Information Assoctition, Report No. 56. London, 1975 (Construction Industry Research and Information Association). 55. CRONEY, D and J N BU~AN. The influence of climatic factors on the structural design of fletible pavements. Third International Conference on the S~ctiral Design of Asphalt Pavements, London, 11–15 September 1972. Proceedings, Volume 1. Ann Arbor, Michigan, 1972 (University of Michigan), pp 67–71 . 56. PEATTIE, K R. Flefible pavement design. Developments in highway pavement engineering. tindon, 1978 (Apphed Science Publishers). 57. BROWN, S F. Material characteristics for analytical pavement design. Developments in highway pavement engineen.ng. hndon, 1978 (Applied Science Publishers). 58. BRIEN, D. Asphalt mh design. Developments in highway pavement en~.neering. hndon, 1978 (Apphed Science Publishers). 23 Water from ditch and surrounding country draining on to road Typical sunken road Right Earth bladed from side drains -- —-- - . ---- --- Properly maintained earth track Fig. 1 CORRECTLY AND INCORRECTLY MAINTAINED EARTH TRACK —— This scale should be redrawn I I Sngle or equivalent single wheel load in kips II Yx A A v I 1 I I I I I I /l A/z I t I I I r w r \r I 1 1 I 1 I I I I I 1 1 I 15 25 50 70 100 150 200 300 psi Tyre pressure 27 mm to the right before CBR the nomogram is used. 100 80 E i Cove rages 10000 8000 6000 4000 2000 I 1000 800 600 400 200 F iOO 80 60 40 E r 20 10 8 [ Fig. 2 RELATION BEWEEN LOAD, REPETITION, TYRE PRESSURE AND CBR ,L FOR UNSURFACED SOILS (after Ahlvin and Hammittl 6 ) 0.6[ 0.5( 0.4C o.3a I ‘- m 0.20 — These curves have been drawn using information derived from the modified Shell design curves. For each design situation al has been calculated as al= total pavement thickness of unbound layers ~ 0,4 total pavement thickness of asphalt layers . where 0.14 is the layer coefficient of unbound base material Developing countries —— —— I g I m< I a. -c 6 3 I I 1 Ill 1 0.10 0 5 10 “15--20.25.30 35 40 MAAT (Shell) ‘C Fig. 3 POSSIBLE VARIATIONS IN SURFACE LAYER COEFFICIENTS FOR DIFFERING CONDITIONS *.U 3.0 2,0 R = 1.0 PSI = 2.0 AASHTO (2” /: / / L= 0.5 PSI = 2.0 AASHTO (2” / /1 / / ‘/‘J.0 ESA/veh CEBTP / < S~ll (all base+ 2)t .A.’...y . . . . . ,,’ ad Note 29 <, hell (all base + SD)t / / ‘%::. ]lO,O ,,A,veh ~EBTp : / 0“, .. . . . . . 7 . . . . . . . . . . . / ~’/ :/ / ,.’ .“” jz~Shell (8’ base+ SD)t . ‘,/ : A J ~&;@:~ ;;;”volved I I 1 1 I 0.1 0.5 1.0 , 5.0 10.0 Traffic x 106 standard axles Fig. 4a COMPARISON OF DESIGN METHODS FOR CBR 8 SUBGRADE Notes ———— — —-— —- ------------- .. . . . . ..- . . . . .---------- R=l,O PSI=2.O AASHTO methti with R = 1.0 and PSI = 2.0 AASHTO (2’ AC) for a pavement with 2 inches (50mm) of asphaltic wncrete (al= 0.44) and using UTAH correlation between S and CBR R = 0.5 PSI = 2.0 AASHTO method as above but with regional AASHTO (2’ AC) factor R = 0.5 Shell (all base + SO) Shell modified design method assuming surface dressing on a crushed rock base (take al D1 = 0.05). All unbound materials are assumed to be of base quality Shell (8” base + SO) Shell method as above but assuming wrface dressing + 8 inches (200 mm) mushed rock base + subbase to make up total thickness of unbound material Shell (all base + 2“) Shell modified design method assuming 2 inches of dense asphalt on a crushed rwk base. All unbound materials are assumed to be of base quality. Road Note 29 Road Note 29. TRRL method for U.K. Asphalt Institute Asphalt Institute design method assuming quoted minimum thickness of surfacing asphaltic mncrete, 6 inches (150 mm) of crushd rock base if neces=ry and the balance of subbase material if required -.-. -.— .- —.. —.. - Road Note 31 Road Note 31. TRRL method for developing muntries ................. ................. 1.0 ESA/veh CEBTP CEBTP method for tropiwl developing muntries assuming that traffic is equivalent to 1.0 standard axle per wmmercial vehicle 10.0 ESAlveh CE8TP CEBTP method as above for traffic equivalent to 10 standard axles per mmmercial vehicle AC) AC) Fig. 4b COMPARISON OF DESIGN METHODS \ .\. TC (Total costs) \ . \. \. . . \ ●✼✎ . i I i I Y . I I I I CMC (Construction and maintenance costs) \., ● . > / ..~ - = ‘(Vehicle o~r~ting costs) ./- 0 (Optimum) Axle load limit Fig. 5 THE NATURE OF THE RELATION BEWEEN TONNE-KM COSTS AND AXLE LOAD (2293) Dd0536361 1,500 11/79 HpLtd So’ton G191s P~INTED.lN ENCLAN~ ABSTRACT PAVEMENT ENGINEERING IN DEVELOPING COUNTRIES: C ZEllis: Department of the Environment Department of Transport, TRRL Supplementary Report 537: Crowthorne, 1979 (Transport and Road Research bboratory). This Report discusses the reasons for the differences between pavement engineering in temperate chmates and in developing countries with tropical or sub-tropical climates. The importance of earth and gravel roads in developing countries is emphasised, and commonly used methods of pavement design for bitumen surfaced roads are described and compared. Recent developments in techniques and equipment for improving construction standards and for assessing road performance are described. The use of appropriate forms of contract is briefly discussed and the need to improve knowledge of the factors which would enable total transport costs to be minimised is emphasised. ISSN 0305–1315 ABSTRACT PAVEMENT ENGINEERING IN DEVELOPING COUNTRIES: C1 Ellis: Department of the Environment Department of Transport, TRRL Supplementary Report 537: Crowthorne, 1979 (Transport and Road Research bboratory). This Report discusses the reasons for the differences between pavement engineering in temperate climates and in developing countries with tropical or sub-tropical climates. The importance of earth and gravel roads in developing countries is emphasised, and commonly used methods of pavement design for bitumen surfaced roads are described and compared. Recent developments in techniques and equipment for improving construction standards and for assessing road performance are described. The use of appropriate forms of contract is briefly discussed and the need to improve knowledge of the factors which would enable total transport costs to be minimised is emphasised. ISSN 0305–1 315