Anoverview ofthestructural designof minehaulage roads · Anoverview ofthestructural designof...

An overview of the structural design ofmine haulage roadsby R.J. Thompson * and A.T. Vissert

Synopsis

The structural design of a pavement relates to the ability of theroad to carry the imposed loads without the need for excessivemaintenance or rehabilitation. Current techniques for the structuraldesign of haulage roads are empirical and based primarily on theprevious experience of personnel assigned to pavement design,both in terms of the strength of the structure and the quality of thewearing course material. This has the potential for unwarrantedexpenditure on too thick a structure or, conversely, prematuredeformation leading to the need for excessive expenditure onmaintenance. Furthermore, the effect of larger haulage vehicles onthe structural performance of a pavement cannot be predicted.Thus, there is a need to develop a portable and practical method ofstructural design. This paper presents an overview of structural-design methods for mine haulage roads and, through an analysisand quantification of the structural performance of existingpavements, recommends the mechanistic design procedure,together with a revised structural design and associated limitingdesign criteria. It is concluded that the application of themechanistic method to a design project can reduce the costs ofhaulage roads construction and accrue additional benefit in termsof reduced operating and maintenance costs.

Introduction

Large-scale surface mining in South Africa is arelatively recent innovation. At the turn of thecentury, open-pit mining, almost withoutexception, was short-lived and quicklyreplaced by underground mining. Only in theearly 1960s did extensive open-pit miningbegin, with mines such as Palabora andSishen. A decadelater, the crisis in the oilindustry provided the impetus for the large-scale development of South Africa's coalreserves. production rose from 66 Mt of run-of-mine coal in 1974 to 260 Mt in 1994. Of thistotal amount, some 40 per cent, or 106 Mt, wasproduced by open-cast methods, which require,inter alia, the transportation of raw coal fromthe pit to the loading or transfer point.

In addition to the mechanical-engineeringchallenges involved in the design and construc-tion of large dump trucks, specific mining-engineering problems arise, primarily as aresult of the gross vehiclemass (GVM) and the

The Journal of The South African Institute of Mining and Metallurgy

resultant loading applied to the pavement.Taking the Caterpillar 789 as a typical exampleof a 170t capacity rear dump truck, the GVMof317.4 t results in a maximum dual-wheel axleload of 2086 kN. as illustrated in Figure 1. Incomparison, public road authorities permit alegal maximum dual-wheel axle loading of80 kN, which is similar in magnitude to thatassociated with a 25 t truck with tandem rearaxles.

Large mine haulage trucks impose axleloads ranging from 110 to 170 t, which areapplied to haulage roads that have been, atbest, designed empirically on the premise of'satisfactory' or 'failed'. This design methodserved its purpose in an era when off-highwaytrucks were lighter and less financial outlaywas required in terms of the initial costs ofpavement construction, the ongoing costs ofroad maintenance, and the costs of vehiclemaintenance. Currently, the costs of truckhaulage can account for between 20 and 40per cent of the total costs incurred by a stripmine and, as the trend in increasing truck sizecontinues, the current pavement systems willprove inadequate. Not only will the maintenancecosts of the existing roads of inadequate thick-ness increase, but vehicle operating and main-tenance costs will also increase prohibitively.

The operating performance of a pavementor road structure system can be subdividedinto the following three distinct designcategories.

~ Structural design. This concerns theability of a haulage road to carry theimposed loads without excessive defor-mation of the pavement and, consequent-ly, the need for excessive maintenance.

~ Functional design. This refers to theability of a haulage road to perform its

--- ----

*Department qf Mining Engineering, University qf

Pretoria, PretoJia, 0002.

t Department fll Civil Engineenng, University qf

Pretona, PreIOJia, 0002.@ The South Aji1'can Institute qf Mining and

Metallurgy, 1996. SA ISSN 0038-223X/3.00 +

0.00. Paper received Feb. 1995; revised paper

received Aug. 1995.

JANUARY IFEBRUWY 1991 29 ~

Structural design of mine haulage roads

Gross vehide mass 317t

~'--------,,--------

-

-, ---

--.-------.-----------

33'1(, 67'1(, 2086ltN

AXLE LOAD (FULL)LOAD DISTRIBUTION (FULL)

52I.51tN

WHEEL LOAD

Figure 1-Magnitude of pavement loading associated with a largehaulage truck

function, Le to provide an economic, safe, and vehicle-friendly ride. The selection of wearing-course materialsprimarily controls the functional performance.

~ Maintenance design. The maintenance aspect ofhaulage-road design cannot be considered separatelyfrom the structural and functional design aspects sincethe two are mutually inclusive. An optimal functionaldesign will include a certain amount and frequency ofmaintenance (watering, grading, etc.), and thus main-tenance can be planned and scheduled within the limitsof the expected road performance. The major problemencountered in the analysis of maintenance require-ments for haulage roads is the subjective nature of theproblem; the levels of acceptability or serviceability forthe road are user -specific.

When applied to the design of mine haulage roads, eachof these components is addressed singularly, and thencombined to produce specific and individual haulage-roaddesigns based on a highly portable, comprehensive, generaldesign strategy applicable to most surface-mining operations.The aim of this paper is to present an overview of structural-design methods for mine haulage roads and, through ananalysis and quantification of the structural performance ofexisting pavements, to recommend a mechanistic designprocedure in conjunction with specific design criteria. Througha comparison of the mechanistic design with existing structuraldesigns, it is shown how the mechanistic design can reducethe cost of haulage-road construction and accrue additionalbenefits in terms of reduced operating and maintenance costs.

Structural design of pavements

The terminology applied in the description of the variousstructural layers of a pavement is illustrated in Figure 2. Thestructural design of a pavement concerns the ability of theroad to carry the imposed loads without the need for excessivemaintenance or rehabilitation.

Pavements deteriorate with time owing to the interactiveeffects of the traffic load and specific sub-grade materialstrengths and structural thicknesses. Vertical compressivestrains induced in a pavement by heavy wheel loads decreasewith increasing depth, as illustrated in Figure 3. This permits

~ 30 JANUARY IFEBRUARY 1996

Road drain

Sub-base layer

Not to scale

Number of pavement layers may vary accordingto specific design adopted

Figure 2-Terminology used to describe the structure of haulage roads

Strong pavement

Load

\7

Weak pavement

Load

\7

Wearing course

~

Verticalstrain ~

~ ~0

Sub-grade

Figure 3-Load-distribution characteristics of a strong versus a weakpavement

the use of a gradation of materials and preparation techniques,stronger materials being used in the upper regions of thepavement. The pavement as a whole must limit the strains inthe sub-grade to an acceptable level, and the upper layersmust, in a similar manner, protect the layers below. Based onthis premise, the road structure should theoretically provideadequate service over its design life.

In an attempt to obtain satisfactory service over the designlife of a road, pavement design models can be used to predictperformance over a wide range of traffic loads and roadstructural designs. Pavement structural design is the processof developing the most economical combination of pavementlayers (in relation to both the thickness and the type ofmaterials available) that is commensurate with the in situmaterial and traffic to be carried over the design life.

Current state of structural design

The structural design of public (mainly paved) roads has beenthe focus of considerable attention from as early as the 1920s.The structural design of unpaved roads in the public domainhas received less attention, design guidelines having beendeveloped only recently!. Similarly, structural-designtechniques for haulage roads that promote safer, moreefficient hauling have received little attention until recently.

Early haulage road design techniques consisted of placingseveral layers of granular material over the in situ materialand, as excessive deterioration occurred, more layers wereadded to provide adequate strength. In 1977, a researchstudy carried out on behalf of the United States Bureau ofMines2 (USBM)considered the structural-design aspects of

The Joumal of The South African Institute of Mining and Metallurgy


mine haulage roads. It was recognized that a stable road baseis a fundamental consideration in road design. The placementof a road surface over any material that cannot adequatelysupport the weight of traffic using it will hamper vehicularmobility and controllability. The findings of the USBMreportwere that the structural-design techniques for haulage roadswere entirely empirical, mine operators often electing toforego the use of sub-base materials and accepting infringe-ments on mobility in the interest of economics. The USBMwork was one of the first annotated applications of theCalifornia bearing ratio (CBR)cover-thickness designtechnique for mine haulage roads.

CBR structural-design technique

The load-bearing capacity of a soil is directly related to itsshear strength, given by the Mohr-Coulomb equation. Tyreloadings for large haulage trucks generally exceed the bearingcapacity of most soils, and thus anything less consolidatedthan soft rock will not provide a stable base for a haulageroad, and other materials will need to be placed over the sub-grade to protect it and adequately support the road structureand traffic.

A survey conducted in 1928-1929 by the California RoadsDepartment to determine the extent and cause of pavementfailures concluded that failure was due to inadequate compac-tion of materials forming the road layers and insufficientcover over weak in situ material. These conclusions indicatedthe importance of material-compaction and shear -strengthconsiderations in road building, both in terms of a suitabledesign procedure and an associated method of materialstesting. The notion of the CBRvalue for a specific materialwas thus developed from a laboratory penetration test of asoaked sample of pavement material from which its shearstrength could be inferred.

The first indications of cover requirements over in situmaterials of specific CBR (percentage) values were reportedby Porter3, who enumerated the concept of cover thickness.Later modifications included a consideration of trafficvolumes, single wheel loads, and increased wheel loadsbased on an estimated maximum allowable shear stress forspecific materials. The problem of dual-wheel assemblies wasaddressed by Boyd and Foster4 through consideration of theequivalent single-wheel load (ESWL), by which a load iscalculated that generates the same tyre contact area andmaximum deflection as would a group of wheels. The conceptof equivalent deflection was used to equate an equivalentsingle wheel to the multiple-wheel group.

Despite the empirical origins of the technique, Turnbulland Ahlvin5 derived a mathematical approach to the calcula-tion of cover requirements using the CBRmethod. Thisapproach is adopted for the calculation of cover requirementsover in situ material as predicted by the CBRdesign methodfor ultra-heavy axles. For the complete mathematical treatise,readers are referred to Turnbull and Ahlvin5 and Ahlvinet al.6. A typical CBRcover thickness curve is illustrated inFigure 4 for the case of a Haulpak 630E fully laden dual-wheel rear axle (53 t per wheel). In addition, the actual CBRvalues for the in situ pavement layers are also shown. TheCBRcover thickness curve represents the minimum coverthickness required to eliminate over-stressing (and conse-quential deformation) of the pavement. The line representingthe CBRvalues for the in situ pavement layer should then lie


Pavement depth (mm)0

-1500

-1000

-1500

-2000

-2500

-3000

-31500

-40001 10 100 1000

CBR

Figure 4-Typical CBR cover curve design for a Haulpak 603E truck(fully laden dual-wheel rear axle) in comparison with in situ pavementCBR derived from analysis by use of the dynamic cone penetrometer(DCP)

to the right of the predictedcovercurve if the pavement isadequatelydesigned;to the left if the pavement is under-designed. In this case, layers two and three exhibit strengthsbelow the CBR-predicted minimum cover requirements andthe possibility of over-stressing and deformation exists.

Although the CBRapproach is an elementary and straight-forward empirical structural design technique, based on andimproved by considerable design experience, there arenumerous disadvantages in the application of the method tothe structural design of haulage roads.

~ The method has its base in Boussinesq's semi-infinitesingle-layer theory, which assumes a constant elasticmodulus for the material. Mine haulage road structureconsists of numerous layers of differing materials, eachwith its own specific elastic and other properties.

~ More specifically, the CBRmethod is based on empiricalresults from public roads subjected to a maximum axleload of 80 kN. Mine haulage roads are subjected toaxle loads that are 25 times as high. Simple extrapola-tion of these empirical design criteria to accommodatehigher axle loads can lead to serious errors of under-or over-design.

The method is thus recommended exclusively to haulageroad structural design cases that incorporate single layersonly. However, the method, when applied judiciously, can beused to determine safe (total) cover over in situ materials,although the extent of over- or under-design associated withthe method remains to be quantified. where cemented orstabilized layers are included in the design, or where theoptimal structural design is sought, use should be made ofother design techniques that can account for the differentpavement layer material properties and more accuratelypredict their performance under the action of ultra-heavy axleloads.

One of the tenets of the CBRcover-thickness designtechnique is the determination of the bearing capacity of thein situ pavement layers. This can be estimated by use of thedynamic cone penetrometer (OCP).

DCPpavement design and analysis

The extensive use of the DCP in the determination of the insitu shear resistance of a material has enabled predictive

JANUARY/FEBRUi' 31 ....1996


models of pavement performance to be developed for thin-surfaced, unbound-gravel, flexible road pavements, togetherwith the correlation of the DCPresults with the CBR7,8,aswell as with the 7-day soaked unconfined compressivestrength (DCS) of lightly cemented materials(DCS< 3000 kPa) , as discussed by Kleyn9 and De BeerlO,

Although the DCPinstrument is ideally suited to theevaluation of existing pavements, the original research7 wasused to establish a design method for new pavements. Thisdesign approach is, in principle, similar to the CBRapproachin that over-stressing of the lower layers is prevented througha balanced increase in layer strength. The DCPinstrumentcan be used in the investigation of haulage-road structuraldesign for the analysis of

~ the location of various pavement interfaces~ the CBRvalues of these various layers~ the overall balance of the structural design.

A first attempt at the recognition of pavement layers canbe made from a DCP-generated layer strength diagram. Thisrelates the depth of each layer (vertical axis) to the CBRonthe horizontal axis. The CBRvalues are derived from the DCPpenetration, as described by Kleyn and Van Heerdenll. Thetypical in situ pavement layer CBRproftle shown in Figure 4was derived accordingly.

Although the DCPdata afford an insight into actual roadstructure as opposed to the design structure and the strengthof each layer actually achieved in the field, the extent to whicheach type of design fulftls the structural performance require-ments can be determined only from an analysis of the responseof each layer to the applied loads. When the deficienciesinherent in the CBRcover-curve design technique and theDCPcharacterization are considered from the point of view ofthe structural design of mine haulage roads, it would appearthat a more formal technique is required in which pavement-layer material is assessed more rigorously in both single- andmulti-layered structures.

Mechanistic structural design

The mechanistic approach to pavement engineering involvesthe calculation of stresses, strains, and deflections in thepavement based on the extension of Boussinesq's single-layer theory to multiple layers. The approach adopted for amechanistic analysis and design embraces the prediction ofmaterial behaviour prior to construction through appropriatelaboratory and field tests. The response of a pavement to anapplied load is simulated, and the stresses and strainsgenerated in each pavement layer are compared with limitingdesign criteria. This contrasts with the CBRmethod, in whichempirical designs are based on data obtained from actualpavement behaviour in response to an applied load.

Empirical design criteria are to some extent a requirementof all structural-design techniques. While the CBR-basedapproach is entirely empirical and therefore subject tolimitations imposed by the characteristics of the data, themechanistic approach, although including an empiricalcomponent, relies on mechanistically derived data to whichempirical procedures are applied, therefore extending theapplicability of the technique. Typical benefits of themechanistic-empirical approach are as follows.

» It can accommodate changing loads and their impact onthe structural performance of a pavement.

~ 32 JANUARY IFEBRUARY 1 996

~ It can utilize available construction materials in a moreefficient manner.

~ It can accommodate new materials.~ The performance predictions are more reliable.~ The material properties used in the design process are

more closely related to the field performance of thestructure.

~ The definition of existing pavement-layer properties isimproved.

A simple and convenient method to assess the structuralproperties of pavements involves the application of a loadand measurement of the resulting depth-deflection profile.This is most readily achieved through the use of a multi-depth deflectometer (MDD), which is described by De Beeret al.12 and Basson et alp. In essence, an array of six linearvoltage-differential transducers is used to determine indivi-dual pavement layer deflections resulting from an appliedload. These deflections are compared with those generatediterativelyfrom the ELSYM5A14 multi-layeranalysis programunder various assumed values for the effective elasticmodulus of pavement layers until the observed and calculateddeflections agree to within pre-determined ranges. Once themoduli of the individual layers have been resolved, thestresses and strains can be determined and compared withestablished design criteria to verify whether critical stressesor strains have been exceeded.

As little published data exists concerning establisheddesign criteria for haulage roads, the most tractable approachis to identify the damage parameters applied in the design ofroads subject to standard axle loadings and, by categorizationof the structural performance of haulage-road test sections,deduce acceptable limiting design criteria for the structuraldesign of haulage roads. The development of empiricallimiting fatigue or distress criteria applicable to the structuraldesign of mine haulage roads can then be used to evaluatethe structural performance of a pavement and, if necessary,the efficacy of corrective measures or new structural designs.

Since much of the structural deterioration of a pavementis attributable to the stresses or strains developed in individualpavement layers, these parameters are important in a consid-eration of limiting design criteria. Vertical strains in the topof the sub-base and sub-grade layers are associated withrutting and deformation, while horizontal strains in the upper,stabilized layers are associated with cracking.

According to this strategy, two design criteria can beidentified from conventional road design.

Vertical (compressive) strain

Limitations are usually placed on the permissible compressivevertical strains at the top of sub-grade layers to prevent ruttingand subsequent deformation of the road. Three pavementcharacteristics influence the magnitude of sub-grade strainsunder the application of a constant wheel load:

~ resilient modulus of the wearing-course material~ thickness of the wearing-course layer~ resilient modulus of the sub-grade.

The design criterion for the layers below the wearing courseis that of horizontal tensile strain or vertical compressivestrain, depending on whether the layers have been stabilizedor not. These relate to the failure criteria of fatigue cracking



in the stabilized layers and rut initiation in the sub-graderespectively. There are no published data relating the limitingvalues for strains in a haulage road associated with adequatestructural performance. Recourse must be made to categoriza-tion of the performance from which limiting values of verticalstrain can be deduced, taking cognizance of the characteristicslisted.

Factor of safety (FOS)

Granular materials exhibit distress through cumulativepermanent deformation or inadequate stability. Both forms ofdistress are related to the available shear strength of thematerial and, to prohibit shear failure or excessive gradualshear deformation in the layer, the shear stresses generatedmust be limited.

While the Fas at any point in a layer can be deducedfrom consideration of the applied load and shear -strengthcharacteristics of the material 15, in the structural design ofhaulage roads, extrapolation of established Fas values isunreliable owing to the uncertainty surrounding the damageattributable to ultra-heavy axle loads. Thus, in the deductionof limiting safety factors, recourse must be made to thecategorization of actual haulage-road performance.

Sensitivity of the materials to stress

In addition to the characterization ofthe pavement-layerresponse in terms of limiting design criteria, it is importantthat the stress-sensitivity to of the material comprising theselayers should be assessed. Departures from the expectedbehaviour (especially the predicted stress-hardening ingranular materials)16 may lead to under-design. The stressdependency of unbound gravel layers comprising the structureof haulage roads needs to be assessed at the load levelsencountered with large haulage trucks so that the structuralintegrity of a particular road or design can be predicted.

The vertical compressive strain, FaS, and stress sensitivityof the materials comprising each pavement layer at a test sitewere determinedfrom the results of the ELSYMSAanalysisfor the various loads applied. The effective modulus ofelasticity (Eeff) and Poisson's ratio (J.l)define the materialproperties required for the computation of the deviator andsum of the principal stresses and strains (vertical Evandhorizontal Eh) in a pavement structure. In addition to thematerial properties, the layer thickness is specified, in thiscase with reference to the DCP-derived data for the re-definedstructural thickness of the layers. Figure 5 summarizes thelayered elastic model. the data requirements, and the designcriteria adopted in the analysis.

Comparison of various design methods

In a determination of the suitability of the previouslymentioned techniques of structural design as they apply tomine haulage roads, existing mine pavements were assessedin terms of a number of independent variables (factors) thatwere recognized as predominantly controlling the structuralperformance of a haulage road. These factors are appliedload, strength of the sub-grade, and structural thickness andstrength of the layers.

The Journal 01 The South African Institute 01 Mining and Metallurgy

. MDD unitinstallation

WHEELLOAD

Y'

i j Load radius r

Contact pressure p

FOS1 = f (0, shear strength) ....

IVertical strain Ev2

FOS2 = f (0. shear strenght) t...

!

Vertical strain Ev3 tT

!

Vertical strain Ev4 t

h 1 Eel'1 ~1 Gravel wearing course

h2Eeff2~? Base

h:JEefl:J~3 Sub-base

h4Eefl,,~4 Sub-grade

In situ infinite In depth

Figure 5-Model of a layered elastic pavement for mechanisticstructural design and analysis

The structural performance of existing haulage-roaddesigns were fully characterized through the analysis of eachcombination of factors at various levels by use of a designedfactorial experiment. Nine test sites were identified at threemines, which enabled 27 factorcombinationsto be assessed.For each of the independent variables, actual field valueswere recorded, together with the resilient deformations ineach of the test-site pavement layers and sub-grade. By thisapproach, the utility of each design method was assessedover a range of factor levels, enabling a comparative analysisto be completed for all combinations of haulage-truck loads,construction materials, structural thicknesses, and layerstrengths.

As a precursor to the analysis of the structural performanceof haulage roads and the derivation of limiting design criteria,a classification was drawn up of the structural performanceof haulage roads to indicate in broad terms the adequacyofthe various designs encountered at each mine test site, Thiswas achieved by the assignment of an index to each siterepresenting poor to excellent structural performance, togetherwith a short summary of the structural defects observed orreported by mine personnel. In addition, the maximumdeflectionrecordedby the l'v\OD in the structure was depictedfor each site as an aid to classification,

DCP analysis

Nine mine haulage-road test sites were evaluated in terms ofthe layer profile, layer CBR,and balance of the design. Arange of pavement designs were encountered. In general, thetest sites in which most of the pavement strength lies in theupper layers were found to be more sensitive to increasedwheel loads and consequential f~lilureof the upper layers. Incontrast, a test site where the supporting layers were strongwas less sensitive to any increase in wheel loads, butexcessive deformation in the weaker upper layers could occur.However, the extent to which these effects occur in haulageroads and the degree to which each type of design fulfils thestructural performance requirements could be confirmed onlyby correlation with the results of the mechanistic analysis.

JANUARY IFEBRU/\I 33 <till1')96


CBR analysis

The CBRmethod has been widely applied to the design ofsurface haulage roads in which untreated materials are used.In essence, the method relates the cover thickness required tothe in situ bearing capacity, thereby eliminating over-stressing and consequent deformation of the sub-grade dueto axle loading.

The results relating the required cover thickness and CBRvalue for a particular vehicle and the actual (as constructed)cover were compiled for each test site, using the particularvehicle for which the road was constructed and the largestvehicle currently operated. The results ranged from potentialunder -design of the road (inadequate cover) to probableover-design (excessive cover). In the case of the indicatedunder-design, there was a correlation between poor structuralperformance as determined from the performance classifi-cation and inadequate CBRvalues for the pavement layers.Recourse to mechanistic analysis is necessary to quantify theextent of any over- or under-design recognized from the CBRapproach, since each layer cannot be assessed independentlyin terms of its load-bearing capability by use of the lattertechnique. The technique can, however, be used to generatereliable estimates of the total cover over the in situ material,subject to the limiting assumption of a single layer. Wheremulti-layer structures are involved and the optimal design issought, the mechanistic approach is more appropriate.

Mechanistic analysis

Two design criteria were proposed for the assessment of thestructural performance of mine haulage roads: the verticalcompressive strain in each layer below the top layer, and theFOS for the two uppermost layers. Figure 6 relates thevertical strains in each layer for three typical test sites; theother sites exhibited similar responses. A comparison of

these data with the performance classification showed thatthe vertical strain criterion correlates well with the structuralperformance of the road; those mine sites exhibiting poorperformance and an associated excessive deformation ormaximum deflection were associated with large values ofvertical compressive strain in one or more layers. From ananalysis of all the data, it was found that the vertical strainvalues should not exceed 2000 microstrains. Higher strainvalues were found to be associated with unacceptablestructural performance.

with regard to the FOS for the upper layers, it was foundthat this criterion is not applicable to the design of haulageroads since the FOS is a function of the failure load carriedby each layer, which is not normally attained. In the absenceof any definitive criterion, a 200 mm layer of compacted (95to 98 per cent Mod. AASHTO) good-quality gravel isrecommended for the wearing course. This recommendationis derived from the observation of wearing-course layers atmine sites that exhibited adequate structural performance.

Regarding the propensity of the various pavement layersto stress-harden or -soften, some localized evidence ofstress-hardening and -softening was seen. However, this ismore a function of the specific construction material used ateach site rather than a universal phenomenon.

If the CBRcover-curve design (Figure 4) is compared withthe vertical-strain results for the same site (site 1 in Figure 6),it can be seen that the outcome of the CBRtechnique isunder -design of the pavement at site 1, specifically in termsof the excessive vertical strains in layers 3 and 4. This isevidenced by the approach of the top of layers 3 and 4 to lessthan the minimum cover requirements. In the case of site 2,under -design was also apparent. This is due in part to thevery different CBRprofile from that of the mechanisticallygenerated proflle of effective elastic moduli. The pavementdesign at this site included a layer of dump rock (layer 3) at

Vertical strain (microstrains)9000

8000

"""""""""""""""""""""--'-""""""""-""""""'"

-"""""""""""""'-""""-""""""""""""""""'--""""""""""

..............-........................

.............-.........-..............

Site 3: Layer 2 stabilizedSite 2: Layer 3 dump-rock layer too deep

in pavement structure

Maximum recommended vertical strain

= 2000 microstrains

7000

8000""""""""""""""""""""""""""""""""""""""""""""""""""..

5000

4000

""""""""""""""""" "'-"""""""""""""""""""""""""""""""

"""""""""""""'"

.................................

3000

2000 ...............................

""""""""""""""""""""""""""""""'"

. ."""""""'"

1000

0

... ."""""""""

-100081:2 81:3 81:4 82:2 82:3 82:4 83:2

Mine site: Layer number83:4

.........-............................

"-"""""""""""""""""""""""""""""""""""""""""""""""""""""""""".......................................................

""""""""""""""""""""""""""""""""""""""""""""""""""""'"....................................................................

. ..............................................................................................................................................................................

""""""""""""""""""""""""""""""" ..."""""""""""""""""""""""""""""""""""""""""""""""""""'"

"""""""""""""""""""""""""""""

83:3

~ 34 JANUARY/FEBRUARY 1996

Figure 6-Vertical strain recordings for various test sites and layers under a range of wheel loads



depth and, although vertical strains are much reduced, thedeep rock layer does not appear to improve the performanceof the road, as indicated by the CBR-predicted over-design.At site 3, it was seen that, despite the fact that most of thepavement strength lies in the upper layers, the road performswell and is not susceptible to the effects of high axle loads inthe upper layers, primarily because of the load-carryingcapacity of the stabilized layer (layer 2). This result hasimportant implications in terms of the optimal structuraldesign of a road.

The proposed structural design for pavements is based onthe findings from the mechanistic analysis of existing haulageroads. Of particular importance in this respect is a road thatincorporates a stabilized layer immediately below the wearingcourse. Although stabilization techniques are expensive andthe layers themselves are subject to cracking if not adequatelydesigned, the most tractable option is the use of mine spoilrock in place of a stabilized layer. This provides increasedresilience to the applied loads without the need for excessivestructural thickness.

Summary of comparative analyses

From the foregoing, it is evident that both the DCPand theCBRtechniques of structural design have only limited applic-ability in the design of mine haulage roads. The DCPtechniquecan be employed in the determination of the location ofvarious pavement layers and their associated CBRvalues. Ingeneral terms, the pavement strength profiles to be avoidedare those incorporating strong or stabilized layers deep in thestructure, which are associated with excessive vertical strainsin the pavement. Although the DCPdata afford an insightinto the actual road structure as opposed to the designstructure and the strength of each layer actually achieved inthe field, the extent to which each type of design fulfils thestructural performance requirements can be determined onlyfrom an analysis of the response of each layer to the appliedloads.

As regards the CBRstructural-design technique, under-design can be associated with poor structural performance,but the extent of any potential under- or over-design cannotbe quantified. When the DCPre-defined layer strengths wereanalysed in conjunction with the CBRcover curves generated,it was found that the DCP method, when applied judiciously,can be used in the determination of safe (total) cover over insitu materials. The method is thus exclusively recommendedfor single layer design cases. Where cemented or stabilizedlayers are included in a design, or where the optimal structuraldesign is sought, owing to the very different properties of thelayer from those of normal road-building gravels, use shouldbe made of design techniques that can account for the differentmaterial properties and more accurately predict theirperformance.

Based on the experimental work completed, themechanistic method appears to provide the most tractableand reliable approach to the structural design of haulageroads in which the limiting vertical strain in each layer isused as the design criterion. In addition, as shown by theanalysis of the various structural designs encountered, theinclusion of a rock layer immediately below the wearingcourse confers additional benefits in terms of reducedstructural thickness for similar load-bearing capacities. The


mechanistic-design procedure, together with the designcriteria and optimum design thus established, was applied toa typicalstructural design case study, and the results ofwhich are discussed in the next section.

Case study of mechanistic design

The optimal mechanistic structural design of a haulage roadfor a surface mine embodies the selection of target effectiveelastic modulus values for the construction materials and theplacement of those materials so as to optimize their perform-ance both as individual layers and over the entire structure.The performance was analysed in terms of the minimumthickness and compaction of the wearing course and thelimiting design criterion of vertical strain in the base, sub-base, and sub-grade layers. In addition, of the various designoptions analysed at each mine test site, the inclusionof arock layer immediately below the wearing course afforded thestructure increased resilience to the applied loads without theneed for excessive structural thickness.

For comparative purposes, two design options wereconsidered: a conventional design based on the CBRcover-curve design methodology, and a mechanistically designedoptimal equivalent as discussed previously. It was assumedthat the in situ and road construction material propertiesremained the same irrespective of the design techniqueadopted. For both options, a wearing course with a minimumthickness of 200 mm compacted to a density of 98 per centMod. AASHTO was adopted. A Euclid R170 truck was usedto assess the response of the structure to applied loadsgenerated by a fully laden rear dual-wheel axle, and theassumption was made that no load-induced deflections existbelow 3000 mm. The various design options are summarizedin Figure 7.

CBR cover-curve design

The cover-curve and layer-strength diagram was developedas discussed previously, based on a CBRvalue of 17 per centminimum for the compacted in situ material. Structural-design data are given in Table I. To determine the likelystructural performance of the road in the light of the criticaldesign factors previously identified, a mechanistic analysiswas conducted by use of the ELSYMSAprogram and theassignment of target values of effective elastic modulus foreach layer. These values were determined from a classifi-cation of the materials comprising the pavement layers interms of grading, Atterberg limits, CBR,swelling, and fieldcompaction17.

CBR COVER-CURVEDESIGN TECHNIQUE

0200 mm350 mm500 mm

800 mm

OPTIMAL DESIGN DERIVEDFROM MECHANISTIC ANALYSIS

200 mm ferric rete 0E ~ 150 MPa 200 mm

300 mm ferrlcrete2 layers of 150 mill 500 rnlll selected rock

E = 100 MPa E ~ 3000 MPa

300 mm selected rockE ~ 3000 MPa

700 mm

In silu compactedE=85MPa -

Base 3000 mm

Figure 7-Details of the structural-design options

JANUARY/FEBRU/.I 35 ....1996


Table I

CBR structural-design data

Materialdescription

Selectedferricrete

Selectedferricrete

Selectedferricrete

Selectedsandstoneex-pit<300 mmblock size

In situcompacted

Figure 8 presents the results of the mechanistic analysisof the CBR-derived cover-curve design. It is evident thatexcessive vertical compressive strains were generated in thetops of layers 2 and 3. The strains in excess of 2000 micro-strains were associated with an unacceptable amount ofrutting and pavement deformation. The surface deflectionsgenerated by the applied load of 3,65 mm did not appear tobe excessive but, when accompanied by severe load-inducedstrains, will eventually initiate structural failure. Thecomments made regarding the inapplicability and under-design apparent when the CBRdesign technique is used areborne out by these results, specifically the large verticalstrains developed in the pavement as the design strengths ofthe layers approached the cover-curve line.

OPTIMALMECHANISTIC

DESIGN

429 kN

J

Optimal haulage-road design

The design adopted is depicted in Figure 7 and described inTable H. The design was analysed mechanistically todetermine the likely structural performance of the road in thelight of the critical design factors previously identified.

Figure 8 also shows the results of the mechanisticanalysis of the optimal design. It is evident that no excessivevertical compressive strains are generated in the structure,primarily owing to the support generated by the shallowdump-rock layer. The maximum values of vertical strain donot exceed the limiting design value of 2000 microstrains,and the maximum surface deflections are approximately2 mm. The proposed optimal design thus provides a betterstructural response to the applied loads than does the thickerCBR-based design and, in addition, does not contravene anyof the proposed design criteria.

Table"

Optimal structural-design data

Materialdescription

Selectedferricrete

Selectedsandstoneex-pit<300 mmblock size

In situcompacted

CBRDESIGN

TECHNIQUE

429 kN

0

Vertical strain (microstrains)

3000I

4000I

E=150MPa E=150MPaII

200 i """""""""""""""'" .Optimaldesign I

350 . 1 """""""""""'-""""""""""""""""""""""""""'" .........:

500 I

::I

~ """""""""""""""""""""""""""""""""I~........................................................................................................II

!I1..........................................................................................................

'If

E = 3000 MPa

1200""""""""""""Depth (mm)

E = 85 MPaCritical limitingdesign criteria

2000 microstrains

E = 100 MPa

E = 100 MPa

""""""""CBR design """'""......

E = 3000 MPa

E = 85 MPa

~ 36 JANUARY/FEBRUARY 1996

Figure 8-Results in terms of design criteria for vertical strain as determined for the CBR and mechanistic optimal designs



Cost implications of optimal design

A cost comparison was compiled, based on contractor tenderunit costs, for the construction of a road designed accordingto the CBRcover-curve technique and one constructedaccording to the mechanistically derived optimal design.Included in the comparison were contractor's preliminary andgeneral costs, which were assumed to remain constantirrespective of the structural design adopted. It was alsoassumed that rock and borrow-pit materials were within thefree-haulage distance of the construction site.

The variable costs taken into account were those of thevolume and area of materials required, and the associatedcosts of placement and compaction. The equivalent cost ofthe optimal mechanistic design was calculated from the actualconstruction costs of the conventional design. It was foundthat there was a 28.5 per cent saving in variable costs realised,by virtue of the reduced material-volumetric and -compactionrequirements associated with the mechanistic optimal design.In terms of the total construction cost (including thepreliminary and general costs), there is a total cost saving of17,4 per cent. In addition, further benefits should accrue interms of the reduced operating and maintenance costs arisingfrom the superior structural performance of the road asevidenced from the foregoing analysis.

Summary

Various structural-design methods for pavements wereassessed in terms of their applicability to the design of haulageroads. Although the OCPdata afford an insight into the actualroad structure as opposed to the design structure and thestrength of each layer actually achieved in the field, theextent to which each type of design fulfils the structuralperformance requirements can be determined only from ananalysis of the response of each layer to the applied loads. Asa precursor to the analysis, the CBRdesign technique wasinvestigated in which data generated from the OCP investi-gation were compared with the cover requirements predictedfrom the CBR-design method.

The deficiencies inherent in the development of the CBR-design method, together with the potential for under-designassociated with multi-layer structures, limit the utility of thetechnique when applied to the structural design of minehaulage roads. The method can, when applied judiciously, beused in the determination of a safe (total) cover over in situmaterials, although the extent of over- or under-designassociated with the method cannot be quantified. The methodis thus exclusively recommended for the design of single-layer pavements. Where cemented or stabilized layers are tobe used, or where the optimal structural design is sought, themechanistic design technique should be employed.

The derivation of the design criteria for the mechanisticdesign of surface-mine haulage roads is based on categoriza-tion of the structural performance of those roads and, as such,reflects a range of structural designs and performance. of thetwo design criteria proposed, that of vertical strain correlateswell with the structural performance of a road, and an upperlimit of 2000 microstrains is proposed for the layers of apavement. The FOS criterion is not applicable to haulage-roaddesign since the FOS is a function of the failure load carriedby each layer, which is normally not attained. In the absenceof any definitive criterion, a 200 mm layer of compacted (95to 98 per cent Mod. AASHTO) good-quality gravel isrecommended for the wearing course.

The Journal of The South African Institute of Mining and Metal/urgy

The recommendations regarding the structural design ofsurface-mine haulage roads are centred on the inclusion of alayer of dump rock within the structure. The optimal locationof this layer is immediately below the wearing-course layer.Using this approach, a reduced structural thickness is realizedwithout the attendant deformation and reduction in structuralperformance that would otherwise be evident without a rocklayer. An analysis of the proposed optimal design in a casestudy comparing it with the conventional CBRdesign indicatedthat the mechanistically derived optimal design provides animproved structural response to the applied loads and, inaddition, does not exceed any of the proposed design criteria.As a result of the reduced structural thickness, a saving inconstruction costs of 17,4 per cent over those for the conven-tional design is realized, together with further benefits as aresult of the reduced operating and maintenance costs arisingfrom the superior structural performance of the road.

Acknowledgements

The permission of AMCOALto publish this paper, togetherwith their financial support for the research, is gratefullyacknowledged, as is the assistance and co-operation receivedfrom mine personnel.

References

1. COMMITfEEOf STATEROADAUTlll'I(JTIES.The structural design. constructionand maintenance of unpaved roads. Drqf{ TRH20. Pretoria. 1990.

2. KAUfMAN,W.W., and AULT, ).c. The design of surface mine haul roadsmanual. USDOIIrlfonnation CIICIIlar 8758. United States Bureau ofMines, 1977. pp. 2-7, 19-30.

3. PORTER,0.). Development of the original method for highway design. Proe.ASCE, vol. 75.1949.

4. BoYD, W.K., and FOSTER,CR. Design curves for very heavy multiple wheelassemblies. Ibid.

5. TURNBULL,W.)., and AHLVlN,R.G. Mathematical expression of the (BRrelationships. Proceedings 4th II/Icmatiol/al Corlference on Soil Mechanics

and Foundation Engineering. london, 1957 vol. 2, p. 178.

6. AHLVlN,R.G., ULERY,H.H., Hl'HlIIW1~, RI., and RIcE, ).L. Multiple wheelheavy gear load pavement tests. Basic report, USA Waterways

Experimental Station Report. AFWI-TR-70.tI3.1971.vol.1.

7. KLEYN,E.G. The use of the dynamic cone penetrometer (DCP). RepoltL2/74. Pretoria, Transvaal Provincial Administration Roads Branch, 1975.

8. KLEYN,E.G., MAREE,).H., and S,I\.\l'l', P.F. Application of a portablepavement dynamic cone penl'lrometer to determine the in situ bearingproperties of road pavement layers and suhgrades in South Africa. ESOPT

11. Amsterdam, 1982.

9. KLEYN,E.G. Aspects of pavement evaluation and design as determinedwith aid of a dynamic cone pelletrometer. University of Pretoria, M.Eng.thesis, 1984.

10. DE BEER, M. Permanent deformation of pavements with lightly cementedlayers. Drqft Research Report. DVPT -35. Pretoria. DRTT, 1988.

11. KLEYN,E.G., and VAN HEERDE~,M.I.I. Analysis and classification of DCPsurvey data. Proceedings 1983 AlIl/ual7i'al/sportation Conventioll.

Johannesburg, 1983. vol. 29.

12. DE BEER, M., HORAK,E., and VISSLH.A.T. The multi-depth deflecrometer(MDD) system for determining rhe effective clastic moduli of pavement

layers. ASTM SIP 1026. 1989.

13. BASSON,).E.B., WIJNBERGER,0.1.. ami SKUum, ). The multi-depth deflec-

tometer: a multi stage sensor for the measurement of resilient detlcctionand permanent deformation at various depths in road pavements. NITRRTechnical Report RP/3/81. Prctoria, CSIR, 1981.

14. ELSYM5A. Interactive versioll 511 users guide. FHW A, 1985.

15. FREEME,C.R. Evaluation ofpal'l'ment behaviour for maior rehabilitation ofroads. NITRR Technical Rep(1!r Rf'/19/83. Pretoria, CSIR, 1983.

16. MAREE,).H., VAN ZYL,N.).W., ami FIZEEME,CR. The effective moduli andthe stress dependence of pavcmenl materials as measured in some HVStests. Session H qfTransportLlrioll Irifrasrrlf,tUre, Annual TransporrationConvention (ATC). Pretoria, t 982.

17. THoMPsoN, R.)., and VISSER,A.T. The structural design of surface mine haul

roads. Haul roads research pn,/('u, Stage I contract reportjor AJlCOAL.University of Pretoria, Oct. 1994. .

JANUARY/FEBRU/I 37 ....1990

Anoverview ofthestructural designof minehaulage roads · Anoverview ofthestructural designof...

Documents

Transcript of Anoverview ofthestructural designof minehaulage roads · Anoverview ofthestructural designof...