One of the challenges of South African civil engineers is to develop techniques for infrastructure provision appropriate to local conditions and suitable for creating employment. Local technology, consisting of welded plastic cells known as Hyson-Cells, is identified as a potential solution. After tensioning the cells, coarse aggregate is placed and compacted, after which a sand-cement grout is vibrated into the voids in the coarse aggre- gate. Alternatively, ready-mixed concrete can be used. The design cata- logue for geocell pavements for low-volume roads is presented, and the suitability of the catalogue is demonstrated by evaluating streets that had been in service for more than 5 years. The catalogue of pavement struc- tures was developed from the results of a laboratory study and field trials. An evaluation of eight road projects that had been in service for more than 5 years supported the structural design but also showed that construction control was important to ensure good performance. An economic analysis demonstrated that it was competitive with other pavement types normally used on access streets. One of the challenges of South African civil engineers is to develop techniques for infrastructure provision appropriate to local condi- tions and suitable for creating employment. In many urbanizing com- munities, the potential for involvement of traditional contractors is rather limited. This limited potential exists because there is little space to maneuver large equipment, and the communities dislike long-term disruption of the street system while earthworks and layer- works are being constructed. The need is, therefore, to develop a method that will provide an all-weather road, preferably incorporat- ing the drainage, without the disruptive processes of traditional con- struction, and simultaneously to develop local expertise and provide long-term employment opportunities. A South Africanâdeveloped and patented welded plastic geocell, commercially known as Hyson-Cells, had been successfully used for canal linings, earth reinforcement, and small dam structures. The properties of this product held potential for fulfilling the above need of pavements for local access streets and result in a flexible portland cement concrete pavement. The cells form a square with a length of side of 150 mm, as shown in Figure 1. After tensioning the plastic, which acts as formwork, alternative methods of placing portland cement concrete can be used. One method is to provide premixed or ready-mix concrete. A more suitable labor-intensive method is to place the coarse aggregate in the cells and then apply compaction, and a cement grout is used to fill the voids between the coarse aggre- gate. The process is a cement-grouted variant of the well-known waterbound macadam. The plastic cells limit the amount of shrink- age of the blocks formed in situ. Single-sized coarse aggregate has a void content of 40 percent. By first placing the coarse aggregate only, the grout, which forms 40 percent of the volume of the slab, has to be mixed and handled. TRANSPORTATION RESEARCH RECORD 1652 121 The aim of this paper is to present the design catalogue for geocell pavements for low-volume roads and to demonstrate the suitability of the catalogue by evaluating streets that have been in service for more than 5 years. The catalogue was developed from laboratory studies carried out to gain a fuller understanding of the mechanism of the system and from the results of controlled field experiments. In the field experiments, the structural strength of combinations of pavement layers and geocell surfacings were evaluated. A brief overview of the experimental work and the manner in which the catalogue was derived is first described. The design proce- dure and catalogue are then presented. An analysis of the performance of a number of in-service roads and streets is then provided, and further innovations in the use of the system are discussed. Finally, conclusions are stated. EXPERIMENTAL INVESTIGATION Laboratory Investigation The aim of the laboratory investigation was to determine the proper- ties of the materials that would be used during the field trials and that would be appropriate for execution by semiskilled labor. The prop- erties of ready-mixed portland cement concrete are well known and need no further investigation. Mixes for labor-intensive construction would need to be proportioned volumetrically. This step may lead to potentially large variability of the quality of the product, but this result has to be accepted. Although admixtures to improve work- ability were considered, their dosage in the field was considered to be a problem, and they were not investigated. The flakiness index of the fraction passing the 53-mm sieve and retained on the 37-mm sieve was 42 percent, and that of the frac- tion passing the 37-mm sieve and retained on the 26.5-mm sieve was 37 percent. There are no specifications for the flakiness index of concrete aggregate, but flaky particles in concrete road slabs help increase the flexural strength (1), although this would not be a requirement with the relatively small blocks. On drilling cores from the laboratory slabs, it was found that the grout had not filled completely under the flaky stones; it was therefore considered that the flakiness index for this type of construction should be limited, provisionally to 30 percent. In South Africa, 150-mm concrete cubes are used to define the strength of concrete. Three cubes were made at each of two cement/ water ratios (2.4 and 2.7), using ordinary portland cement and port- land fly-ash cement. The coarse aggregate was first placed into the mold and compacted, and then a 1;1 by volume sand-cement grout was vibrated into the matrix. The 28-day compressive strength of all the mixes was not significantly different from 49 MPa. As there are no specifications for this type of mix, it was assumed that a com- pressive strength of 30 MPa would be sufficient for general-purpose Flexible Portland Cement Concrete Pavement for Low-Volume Roads ALEX T. VISSER AND SALLY HALL A. T. Visser, University of Pretoria, Pretoria, 0002, South Africa. S. Hall, Hyson-Cells, P.O. Box 319, Muldersdrift, 1747, South Africa. applications. As a result of the high laboratory strengths obtained, the field experiment also included 2;1 sand-cement sections. The laboratory study on the load transfer capability used a 1.2-m by 0.8-m slab on different thicknesses of rubber, representing soft subgrade conditions (2). Slabs with a nominal thickness of 150 mm, 100 mm, 75 mm, and 50 mm were tested. As expected, there was a general trend of decreasing stiffness with decreasing slab thickness. The thicker slabs had better load-distribution properties as the thin- ner blocks could rotate more freely. Initially, the effect of the slab on the reduction of deflection of the rubber was small. However, as the load increased, the effect of the slab became larger and the slab âlocked up.â Good mechanical interlock can be obtained with thor- ough compaction of the coarse aggregate as this action deforms the thin plastic membrane in the vertical plane. In the case of ready- mixed concrete, the vertical deformation of the cells is achieved by stretching the cells in both horizontal directions, and by pushing them down slightly. It was found that the three-dimensional inter- lock was highly effective, as the blocks had to be separated by means of a heavy hammer after the experiments were completed. The stiffness of the geocell concrete layer was relatively low com- pared to normal asphalt and portland cement concrete slabs, hence the designation of a flexible portland cement concrete pavement. Stiffnesses of less than 100 MPa were measured (2) in the laboratory study on the small slabs, but in more recent field experiments, which are as yet unpublished, values of between 1000 and 1500 MPa were recorded, which are similar in value to block paving. Field Investigation High Traffic Volumes The aim of the field experiments was to evaluate the performance of the geocell pavement under accelerated traffic. Access streets typi- cally carry up to about 200,000 equivalent-standard 80 kN axle loads during a design life of 20 years (3). Therefore, the site had to carry substantial volumes of heavy traffic. Such a site was available 122 TRANSPORTATION RESEARCH RECORD 1652 on the exit from Ready Mix Materialsâ Honeydew operation, where about 300 laden five-axle trucks left the site daily. The existing decomposed granite unpaved road had been well compacted. The surface was skimmed with a motor-grader to pro- vide an even working platform. Dynamic Cone Penetrometer (DCP) tests were done along the length of the test section. The wearing course was about 250 mm thick, and had an in situ California bear- ing ratio (CBR) of about 150. Along the eastern part of the section, the DCP could not penetrate beyond the wearing course. At the western end, the roadbed below the wearing course had a CBR of 25 to a depth of 600 mm. This result meant that the existing gravel road had a substantial pavement structure, and the main form of evaluation would be the wear and crushing resistance of the plastic cell slab, as the deflections would be low. The experimental design, therefore, consisted of three different mixes, as discussed below. The road width was 6 m, and the plastic cells were feathered into the side channel to provide a drainage channel. Three test strips were constructed with 150-mm-high cells, namely a 2;1 (by volume) sand-cement grout (Mix 1) and a 1;1 sand-cement mix (Mix 2). Commercially produced Portland Fly Ash Cement, PC25FA, from Blue Circle Cement conforming to SABS 1466, was used. Mix 3 was a 1;1 mix, but the cement consisted of a Blue Circle-patented Rock- fast PCA system specially formulated to set quickly and yield high early strengths. During construction, 150-mm cubes were made and cores drilled out of the plastic cell slab. The process of making the cubes was the same as that used during the laboratory study: namely, the stone was placed, and thereafter the field-mixed grout was vibrated into the voids. Table 1 gives the results of the cube and 100-mm-high cylinder strengths. The cylinders were capped with sulphur mortar in accor- dance with SABS Method 865. Note that batching of the materials was done volumetrically, and that water was added to provide a workable consistency judged by the manner in which the grout pen- etrated the stone. This could explain some of the variability. All the cores drilled showed that the grout had penetrated to the bottom of the layer and that the construction process was effective. Without making corrections for the fact that cylinders were used, there was considerable agreement between the compressive strengths of the laboratory-prepared and cured cubes, and the cylinders. There was adequate strength gain despite the sections receiving little curing dur- ing the dry, windy winter high-veld conditions. This probably reflects what would be encountered in practice. It is also interesting to note that the cube strength of the 1;1 field mix (30 MPa at 28 days) was substantially lower than that obtained during the laboratory investi- gation (49 MPa at 28 days). This result reflect on the method of field control and mixing. The road sections were opened to heavy truck traffic on Monday, May 25, 1992. At that time Mix 1 was 7 to 9 days old, Mix 2 was 6 days, and Mix 3 was less than 72 hours old. At opening, the com- pressive strength of Mix 1 was about 10 MPa, and Mixes 2 and 3 about 15 MPa. On the first day, about 200 laden five-axle trucks used the road; and during the first month, a total of about 4,000 of these trucks passed over the section. By the time the section was abandoned after 21 months (because of relocation of the roads), about 90,000 heavy trucks had used the road. This represents in excess of 400,000 equivalent standard axles, considerably more than would be expected on an access street during its design life; and the section continued to provide excellent service. On Mix 1, the weakest section, the grout wore away quickly to leave the upper stone surface exposed within days. In isolated FIGURE 1 Schematic view of the welded plastic cell after placing. patches, probably indicative of nonhomogeneous mixing, the wear extended to a few millimeters below the stone surface. After the ini- tial wear, no further deterioration of the surface was noticed. Mixes 2 and 3 showed slight polishing and smoothening of the surface, but no significant wear. The sections were wetted several times during the day as the adjacent gravel road was watered. Fine material, then attached to the tires of the trucks, was deposited on the concrete and acted as a grinding paste. Even under these abnormal conditions, a concrete compressive strength of 15 MPa at opening should be ade- quate for heavy traffic. Light vehicles should be able to use a road within 24 hours without fear of exceptional damage. At the south- east corner, the trucks took a short cut over the side drain and caused a deformation of about 10 mm. The cell slab accommodated this movement without any structural damage. The experience with these test sections corroborated the findings of an earlier trial, which involved a 200-mm-thick slab and a concrete compressive strength of 40 MPa at 28 days. This particular section carried about 125,000 laden trucks on the same exit road. Polishing of the surface was also prevalent, although a stronger mix had been used. On part of this trial, the underlying excavation was poorly com- pacted, and a permanent deformation of 100 mm took place under traffic. Despite this movement, no structural distress was evident. This lack of structural distress further demonstrates the flexible nature of the pavement, and indicates the potential for use on weaker subgrades or areas where soil movement could cause damage to other pavement types. Thin-Slab Field Experiment Many access streets carry only a few vehicles per day, and from the laboratory experiment a 50-mm-thick cell slab appeared suitable. A small-scale field experiment was constructed on the entrance to the Blue Circle Cement/Ready Mix Materials Joint Testing Laboratory at Honeydew to evaluate its suitability. The existing road of old con- crete cubes and fine sand was removed and replaced with locally available decomposed granite fill material. Compaction was vari- able and the in situ CBR with the DCP ranged from 35 at one end to 70 at the other end of the 15-m-long by 5-m-wide strip. Because of the slab thickness, 19-mm dolerite aggregate was used instead of the 53 mm used on the thicker sections. The slurry was Visser and Hall 123 again made with the Rockfast PCA System. The compressive strength of 100-mm cubes manufactured in the laboratory with the materials used on site was, after 28 days, between 17 and 23 MPa. The section was trafficked by a 1-ton pickup about 90 minutes after placing, and there were no signs of damage. Besides normal laboratory traffic a 5-ton water tanker made about 4 passes over the section every day over a 4-month period, and the section was per- forming well. These results show that a 50-mm cell slab is feasible for lightly trafficked access streets. PLASTIC CELL DESIGN GUIDE AND CATALOGUE The design guide (2) was compiled based on field observations of the test sections, response during the laboratory study, and the UTG2 (4), TRH4 (5) and the low-volume street guidelines document (3). Note that this design only applies to geocells made of thin-walled smooth plastic film. The design process is based on the widely accepted process used for pavement design in South Africa [TRH4, UTG3 (6)]. It will be discussed briefly, and the catalogue of thickness designs will be presented. Step 1: Type of Street The design catalogue is focused on low-volume streets. These could be access streets to houses, or minor arterials that carry limited bus traffic. The type of application must first be identified. The designs are not recommended for heavily trafficked arterial routes under normal circumstances. Step 2: Design Strategy Under this topic the structural design life of the facility has to be decided, and a suitable analysis period selected for carrying out the economic analysis. The economic analysis is important for compar- ing the plastic cell pavement with other types of pavement construc- tion. The plastic cell pavement will require little, if any, maintenance, TABLE 1 Comparison of Cube and Cylinder Strengths After Construction and will often be used in areas where a low-maintenance capability exists. Therefore, the recommended structural design period will be the same as the analysis periodânormally 20 years. Step 3: Traffic Loading Heavy vehicles are responsible for the primary traffic damage to the road. The range of axle loads has to be converted to the equivalent damage caused by a standard 80-kN axle load, E80 [TRH16, (7)]. This is done according to the so-called fourth power law. The damage factor multiplied by the number of axles of that load, and summed for all the axle loads per day, gives the total E80s per day. By consider- ing the growth in E80s and the design period, the design E80s can be calculated (7). In the calculations, passenger cars and minibuses may be neglected, as their influence on the road structure is minimal. Typically, access streets carry less than five heavy vehicles per day, particularly if the layout design discourages through traffic. In the analysis of the traffic, the amount of building construction traffic has to be considered. On low-volume streets, construction traffic could account for as much as five times the damage of normal traffic. In many cases, it may be cost-effective to delay pavement construction until the suburb is substantially developed. Step 4: Material Considerations The intent is to place the plastic cell slab directly on the in situ mate- rial, except for the more heavily trafficked routes where additional layers may be necessary. The 4-day soaked laboratory CBR of the in situ material is used as the design criterion. For design purposes, the 10-percentile CBR is used. That is to say, no more than 10 per- cent of the CBR values are lower than the design value. In dry regions, the CBR at optimum moisture content or in situ conditions may be used if conditions are expected to remain unsoaked. Engi- neering judgment and local experience should be used in determin- ing if relaxation is feasible. Considerable care should be taken in the soil survey to define areas with soaked CBR values of 3 or less, where special treatment would be necessary on advice of specialists. Problem materials, such as swelling clays, collapsing sands, or dispersive soils, would also require special attention before construction of the plastic cell slab. These remedial treatments are beyond the scope of this simplified guide. On the more heavily trafficked routes, additional supporting layers may be necessary. The material designations used in the catalogue are those used in TRH14 (8). A G5 material has a soaked laboratory CBR of at least 45, a CBR swell of at most 0.5 percent, and a plasticity index of not more than 10. It would be compacted to 95 percent Mod. AASHTO density. The stabilized C4 material, compacted at 100 per- cent Mod. AASHTO density, would have a 7-day unconfined com- pressive strength of between 0.75 MPa and 1.5 MPa. In the field, it would be compacted to 95 percent Mod. AASHTO density. These guidelines may be modified based on local knowledge and experience. Materials for the slab should fulfill typical requirements for clean- liness and the absence of organic materials applicable to concrete. The coarse aggregate does not have to be crushed, but should be essen- tially single-sized, as a graded aggregate could inhibit ready penetra- tion of the grout. For 50-mm and 75-mm slabs, the aggregate should pass a 26.5-mm sieve and be retained on a 19-mm sieve, whereas for the thicker slabs, minus 37-mm plus 26.5-mm stone should be used. A maximum flakiness index of 30 percent should be required. Vol- 124 TRANSPORTATION RESEARCH RECORD 1652 ume batching using approximate sand to cement ratios of 1;1 or 2;1 should be adequate. Care should be taken with the amount of water added, as a thick paste is required. The required cement to water ratio for the desired strength is selected, and the volume of sand to give the desired consistency is added. Well-rounded plaster or wind-blown beach sands will reduce the water demand and facilitate construction. Step 5: Environment The climate will largely determine the weathering of natural materi- als, the durability of weathered natural road-building materials, and also depending on drainage conditions, the stability of untreated mate- rials in the pavement. In this guide, the Weinert classification is used to distinguish climatic regions. The Weinert N-value is the ratio of the evaporation during the warmest month divided by the mean monthly rainfall. The wet region is the area where the Weinert N-value is less than 2. This encompasses the southern and eastern coastal regions in South Africa, as well as the escarpment areas [see map in TRH4 (5)]. The remainder of the country is considered as dry or moderate, termed dry for convenience sake. Note that the selection of a climatic region is not a surrogate for poor engineering practice, such as the lack of adequate drainage pro- vision. The plastic cell pavement can be used to mold side drains, or even drains in the center of the street. Special attention has to be given to drainage, particularly in steep terrain, to prevent the flooding of houses. Step 6: Proposed Catalogue of Structural Designs Selection of layer thickness is only one small element in the overall design process. This means that the proposed design catalogue in Table 2 and Figure 2 should not be used without considering the complete design process. The primary function of the plastic cell slab is to provide an all- weather surface, and to keep water out of the underlying layers. By controlling water ingress, the strength of the layers is utilized effec- tively. Since the cell slab consists of individual blocks, a small amount of shrinkage (10 to 20 microns) occurs at the plastic mem- brane. Consequently, there are no large joints where problems can occur. It is, thus, expected that only minor maintenance would be required. The plastic cell slab is also able to accommodate signifi- cant permanent deformation without structural damage, as was found during the field trials. The proposed design catalogue is shown in Table 2 and Figure 2. The in situ material would have to be scarified and compacted to a depth of 150 mm if any reshaping is necessary. Often, traffic com- paction would obviate the need for further compaction if the existing material is not disturbed. The material codes were discussed in Step 4. Step 7: Practical Considerations Side drains or drains down the center of the road can be constructed and formed with the plastic cells, as the geocells are flexible and will mirror the preshaped underlying surface. Construction is, therefore, an integral process, and, thus, obviates the danger of damage to partly constructed elements. The finish of the street is highly dependent on the evenness of the compacted coarse aggregate. It is strongly recommended that a straightedge be used to ensure evenness, as the grout is ineffective in taking out depressions, or to smooth over stones that are standing proud. The surface can be brushed to a rough texture for steep inclines. Optionally, the concrete can be enhanced architecturally by adding pigment or by imprinting. On steep slopes, it is recommended that 200-mm by 200-mm keys be cut into the underlying layer across the width of the street and cast Visser and Hall 125 integrally with the plastic cell slab at regular intervals. The purpose of these keys is to prevent slippage of the slabs and the potential for- mation of drainage paths under the slabs. The spacing of the keys would depend on local circumstances. The main benefit of the plastic cell pavement is that it is ideally suited to labor-enhanced construction by small entrepreneurs. It is expected that a team of 20 people could construct more than 300 m2 per day. Small hand implements and wheelbarrows are the only equipment needed. During construction, the grout should be placed from the highest point on the street working downhill. This proce- dure facilitates the movement of materials, and avoids construction problems when water, used to moisten the coarse aggregate, washes out the grout if it is dammed by the oncoming construction. This construction procedure would eliminate the same type of problem caused by rain. A plate vibrator was found to be more effective than a hand- driven vibratory roller for vibrating the grout. This may, however, depend on the particular circumstances of the site and the available equipment. Subsequent to these experiments, a mixing and pump- ing system was developed by Putzmeister (Pty) Ltd., which pro- motes improvements in quality control and which permits a consistency of grout such that vibration is no longer necessary for the grout to penetrate to the bottom of the layer. There are, thus, two options available for constructing plastic cell slabs. In rural areas where clean stone and fine sand are not available, on-site concrete mixes were developed, where the cement and sand and water are first mixed before the coarse aggregate is added. This method has shown higher strengths to water added to the dry mix. PERFORMANCE EVALUATION Since the catalogue was first developed, more than 560 000 m2 of roads and streets were constructed, and many of these facilities were older than 5 years. To validate the design catalogue, an exper- imental design was developed that took into account traffic vol- umes (very low traffic, low traffic, and high traffic) and the visual TABLE 2 Plastic Cell Block Pavement Codes for Use in Figure 2 FIGURE 2 Catalogue of structural designs. condition (good and poor). On each section, the following tests were executed: ⢠Visual inspection, ⢠Road roughness with the MERLIN apparatus (9), ⢠Dynamic Cone Penetrometer, and ⢠Cores for concrete compressive strength. Visual Inspection The visual inspections provided valuable guidance for the construc- tion of geocell roads, as no structural failures were noted. None of the cells had cracked through the block, although there was some evi- dence of spalling. Most of the problems were construction-related, and included scaling of the concrete surface as a result of placing a thin levelling layer over the already placed mix. On the first sections built with ready-mix concrete, step forming occurred between adjacent blocks on the fairly steep slope because there was insufficient inter- lock between the adjacent blocks. This problem was subsequently resolved by tensioning the plastic before the concrete was placed. Edge spalling occurred on sections where traffic moved over the edges in areas where such traffic movements were not supposed to take place. Such actions should be anticipated at the design stage, and pos- itive traffic-control measures should be instituted or alternatively proper edge constraints should be constructed. The observed deterioration supports the contention that the geo- cell layer is flexible and the loads are carried by interlock between the blocks. Unlike portland cement concrete slabs, which need to resist significant bending moments, the small blocks carry the loads by compressive resistance. This means that the geocell layer can be placed thinner than an equivalent slab. Road Roughness The MERLIN device is suitable to measure the roughness over short sections of road. Over those road sections where more than the min- imum of 200 readings were taken, the correlated PSI values ranged from 2.7 to 3.1, and the evenness was found to be mainly a function of the ability of the contractor. These roughness values are perfectly adequate for low-volume roads and streets. Bearing Capacity of Supporting Layers One of the benefits of the geocell pavement is that it can be constructed directly on the in situ material in the case of low-volume roads and streets. This facilitates construction and minimizes disruption in built- up areas. Rapid-setting cements allow traffic within 4 hours of finish- ing the area. At the Tutuka power station, the 150-mm-thick geocell layer was placed directly on a poor-quality swelling clay, as a replace- ment for a portland cement concrete slab which had disintegrated in the ash loading area. The in situ CBR varied between 3 and 9 to a depth of 500 mm below the slab. Although some deformation had taken place, there was no structural distress. On some of the sites, the in situ CBR of the layers below the slab was between 10 and 20 to a depth of 200 mm below the slab. The area was trafficked 24 hours a day with a front-end loader and 20-ton axle heavy-duty haulage trucks for 6 years. These results support the catalogue in that the geocell layer can be constructed directly on the in situ material for low-volume roads and streets. 126 TRANSPORTATION RESEARCH RECORD 1652 Concrete Quality Most of the sections evaluated had a geocell layer of less than 100 mm. Only at Tutuka power station, the 150-mm layer permitted the extrac- tion of cores suitable for testing. The strength of the 100-mm-diame- ter and 100-mm-high cylinders was more than 50 MPa. Over a small part of one of the road sections, the concrete was starting to disinte- grate. Cores taken adjacent to the distress showed that the sand-cement grout had only penetrated about 20 mm, resulting in disintegration. In another case the grout had not penetrated the full depth, but no adverse performance could be seen. The reason for the acceptable performance is that the coarse aggregate at the bottom of the layer is constrained and provides the necessary support because of stone interlock. HEAVY-DUTY APPLICATIONS AND FURTHER DEVELOPMENTS At a diamond mine in Botswana, a hardstand area was constructed with 150-mm-high cells directly on the compacted Kalahari sand. The DCP CBR of the supporting sand was about 20 to a depth of about 200 mm, and about 80 up to a depth of 700 mm. The sand has a low CBR when wet, and in the areas where cores were taken the in situ CBR was about 5. The hardstand area is used for refurbishing the trucks, and a mobile crane with outriggers is used to take the tipper off the dumptrucks. No wooden planks are used under the outrigger feet, and the slab is performing well after 4 years of service. The fin- ish on the slab had to be of a high quality as the diamond-bearing sands were retrieved by vacuum cleaning. Interestingly, an adjacent hardstand built with conventional concrete slabs deteriorated to such an extent over a 2-year period that it had to be replaced. Cores taken from the 10 000-m2 area averaged 38.5 MPa, whereas the design specification was 40 MPa. The concrete was made from 50-mm round river pebbles, and the grout was made from wind-blown sand in the ratio of 1;1 to cement. As a result of the success on the low-volume roads, ongoing research is moving into the heavy-duty industrial applications, such as container terminals and storage areas. In addition, a further devel- opment in the provision of urban infrastructure is the integral design of the road and drainage channel. Most of the streets built to date had an integral side drain also constructed with geocells. Recent experi- ence with the use of the geocell for concrete linings of canals showed that there is no water loss through the cells. In many of the urbanizing communities, there is hardly any space for additional side drains, and a drain in the center of the road is a suitable solution. Instead of pro- viding a camber that takes the water off the road, the cross-section is sloped to the center. Since there is no water loss into the pavement structure, there is no danger of the pavement wetting up. ECONOMIC ANALYSIS The economic viability of the plastic cell slab pavement was com- pared with that of traditional flexible and block pavements (2). The structures used were for an access street in a wet area designed to carry less than 0.2 million standard axles. The in situ soaked CBR was taken as greater than 7. The suburb had steep slopes necessitating the use of an all-weather pavement. Typical construction costs in built-up urban areas were used, but it should be noted that wide fluctuations in costs can occur between different sites. For any specific project, local costs should be used. It was assumed that no maintenance would be executed, and that the residual value of the pavement in all three cases was the same. This means that only the construction costs were compared. It was concluded by Visser (2) that the plastic cell pavement is highly competitive with the flexible pavement design, and more economical than block paving. The major benefit is, however, that no excavation is required for the plastic cell pavement, and that pedestrians and light vehicles can use the facility shortly after construction. CONCLUSIONS AND RECOMMENDATIONS The aim of this paper was to present the design catalogue for geo- cell pavements for low-volume roads, and to demonstrate the suit- ability of the catalogue by evaluating streets that have been in service for more than 5 years. It was found that this pavement had load-spreading capacity when built directly on the in situ material for low-volume streets. The catalogue of pavement structures was developed from the results of a laboratory study and field trials. An evaluation of eight road projects that had been in service for more than 5 years supported the structural design, but also showed that construction control was important to ensure good performance. An economic analysis showed that it was competitive with other pavement types normally used on access streets. The plastic cell pavement is suited for construction by small busi- nesses, as relatively little equipment is required and most of the con- struction takes place with labor. The construction process also holds potential for helping small businesses to develop. It is recommended that these principles be put into practice and adapted to suit local circumstances. Further research will include the application of the geocell pave- ment in industrial applications, and the tailoring of the designs to Visser and Hall 127 constraints in urbanized areas, such as incorporating the stormwater drain in the center of the road. ACKNOWLEDGMENTS The assistance of final-year students A. J. de Lange, J. L. Erasmus, and J. F. van Graan is gratefully acknowledged. Reinhold Amts- büchler and Hamish Laing of Blue Circle Cement provided valuable assistance in the construction of the field trials and the testing of the materials. REFERENCES 1. Addis, B. J. Fultonâs Concrete Technology.Portland Cement Institute, Midrand, South Africa, 1986. 2. Visser, A. T. A Cast In Situ Block Pavement for Labour-Enhanced Construction. Concrete Beton,No. 71, Feb. 1994. 3. Horak, E., P. Paige-Green, L. R. Sampson, and A. T. Visser. Guidelines for the Design and Maintenance of Low-Volume Residential Streets in Developing Communities.Transportek Report DPVT-C32.2, CSIR, Pretoria, South Africa, 1988. 4. Structural Design of Segmental Block Pavements for Southern Africa. Draft UTG2. Committee of Urban Transport Authorities, Pretoria, South Africa, 1987. 5. Structural Design of Interurban and Rural Road Pavements.Committee of State Road Authorities, TRH4, Pretoria, South Africa, 1985. 6. Structural Design of Urban Roads.Draft UTG3, Committee of Urban Transport Authorities, Pretoria, South Africa, 1988. 7. Traffic Loading for Pavement and Rehabilitation Design.Draft TRH16. Committee of State Road Authorities, Pretoria, South Africa, 1991. 8. Guidelines for Road Construction Materials.Committee of State Road Authorities, TRH14, Pretoria, South Africa, 1985. 9. Cundill, M. A. MERLINâA Low-Cost Machine for Measuring Road Roughness in Developing Countries. In Transportation Research Record 1291,Vol. 2, TRB, National Research Council, Washington, D.C., 1989, pp. 106â112.
Comments
Report "Flexible Portland Cement Concrete Pavement for Low-Volume Roads"