A laboratory investigation on bonding properties of dowels in concrete roads

Available online at www.rilem.net Materials and Structures 38 (August-September 25) 721-728 A laboratory investigation on bonding properties of dowels in concrete roads M. Löfsjögård Swedish Cement and Concrete Research Institute, Sweden Received: 23 August 24; accepted: 17 January 25 ABSTRACT Inclined dowel bars and unmovable bars may cause pavement cracking in the vicinity to the bar ends. The aim of the investigation is to study if there are any differences in bonding properties due to dowel material, coating or diameter of the dowel. Steel dowels with different coatings and dowels made of composite material are tested. The maximum draw-out force for a draw-out travel of 1.5 mm is measured. The test is repeated four times and ends with a final cycle to establish the constant force needed for a draw-out travel of 5 mm. Steel dowels with bituminous coating show the lowest initial draw-out force. The draw-out force increased 2 to 3 times with a diameter increase of 5% for steel dowels with plastic coatings. For composite dowels the comparing result showed an increase of draw-out force 2 to 5 times with an increase in diameter with one third. The results from the repeating test for several cycles showed that the draw-out and push-back force were almost the same for all dowels. However, for the dowels with bituminous coating a higher push-back force was needed compared to the draw-out force. It should be noted that the testing speed could affect the results, especially for dowels with bitumen. 1359-5997 25 RILEM. All rights reserved. RÉSUMÉ Des goujons d assemblage inclinés ou inamovibles peuvent engendrer des fissures au voisinage de leur extrémité. L objectif de cette étude est de déterminer si les propriétés d adhésion dépendent du matériau utilisé pour le goujon, de son revêtement ou de son diamètre. Des goujons en acier avec différents revêtements ainsi que des goujons en matériau composite sont testés. La force maximale correspondant à une extraction de 1.5 mm est mesurée. Le test est répété quatre fois et est suivi d un cycle final pour déterminer la force constante nécessaire à une extraction de 5 mm. La plus petite force initiale d extraction est obtenue pour les goujons en acier avec un revêtement bitumineux. Pour les goujons en acier avec revêtement plastique, la force d extraction augmente de deux à trois fois lorsque le diamètre augmente de 5%. Pour les goujons en composite, la force d extraction augmente de deux à cinq fois lorsque le diamètre augmente d un tiers. Les tests comportant plusieurs cycles ont montré que les forces d extraction et de rétraction étaient pratiquement similaires pour tous les goujons. Cependant, pour les goujons avec revêtement bitumineux, la force de rétraction était supérieure à celle d extraction. Il doit être noté que la vitesse avec laquelle les essais étaient réalisés pouvait affecter les résultats, particulièrement pour les goujons avec revêtement bitumineux. 1. INTRODUCTION The plain jointed concrete pavement (PJCP) is the most common type of concrete road pavement in Sweden. In order to reduce tensile stresses, limit cracking and limit joint movements the pavement is jointed every 5 meter [1-3]. Traffic loads cause stresses in the slabs and for un-dowelled slabs the stresses are approximately twice as great at the slab edges as in the middle of the slab [3]. With dowels the edge stress is distributed between two adjacent slabs and the principal function of the dowels is to transfer the load from one slab to the other. Omitting dowels in joints will reduce the load bearing capacity and if water is penetrated into the joint a pumping action can occur between the concrete pavement and the sub-grade [1-3]. The results could be step formation between the slabs and/or cracks in the concrete slabs. Using dowels leads to a more durable pavement system and a reduction of the overall pavement deflection [4]. The dowels often consist of smooth steel bars with a length of 6 mm. Dowels used for roads usually have a Editorial note The Swedish Cement and Concrete Research Institute (CBI) is a RILEM Titular Member. 1359-5997 25 RILEM. All rights reserved. doi:1.1617/14148

722 M. Löfsjögård / Materials and Structures 38 (25) 721-728 Fig. 1 - Principle figure of dowel between two concrete slabs. diameter around 25 mm, but there exist smaller ones too, with a diameter of around 17 mm. The diameter of the dowel is selected with regard to the slab thickness, [2]. The dowels are placed at mid-thickness in the slab, spaced about 3 mm apart, and at right angle to each joint. The dowels can be placed in the concrete by insertion and vibration in the fresh concrete or above the subbase in advance. Fig. 1 shows a principle figure of the dowel placement in the slabs. To prevent bonding between the dowel and concrete and to protect the dowel against corrosion, a coating is used on the dowel [1]. It is important that the coating lasts on the dowel. If the coating is not working properly the dowel will get stuck in the concrete resulting in cracks in the middle of the slab or at the cross section where the dowel ends [1]. Steel dowels with coatings of epoxy, plastic (polyethene) or bitumen are the most common type of dowels used for road constructions. Today there also exist dowels made of composite materials. However, as far as the author knows, composite dowels have not been used in concrete roads, at least not in Sweden. One advantage with them is that they are solid, i.e., they have no coating that could fall off. Eisenmann et al. [5] describe an investigation with the purpose of determining the draw-out forces for a draw-out travel of.5 mm. The draw-out test was repeated several times showing that some of the draw-out resistances varied widely with time. The coating must stay on the dowel in order to allow the dowel to be drawn out and pushed back several times without changing the resistance. According to [6], the test speed was 1 mm/minute (maximum displacement 5 mm). After draw-out, the dowels were pushed back with the same speed to its initial position and then the test was repeated. The aim of the joint is mainly to allow longitudinal translations due to shrinkage and thermal movements. The translation of concrete pavements is usually uni-directional both on an annual basis and a daily basis. During its service life, e.g. 4 years, the concrete pavement may be exposed to more than ten thousand 24-hour thermal cycles. Hence, the demand of the dowel is high. Eisenmann s test included steel dowels with coatings of bituminous and plastic films and with variation in dowel diameter. In the study it was stated: an ideal coating should have the lowest possible initial draw-out resistance and this value should not increase in further draw-out tests. Fig. 2 shows the results from Eisenmann s test. Fig. 2 - Results from Eisenmann s investigation on draw-out forces for different types of dowels, after [3]. The effect of sloping or misfit of dowels has been studied in [4, 7]. In [7], steel dowels of length 5 mm and diameter 26 mm were tested with a slope variation from to 4 mm. The dowels were coated with bitumen and RILSAN, with coating thickness.6 mm and.3 mm, respectively. The results show that plastic coated dowels can be installed at greater angles than those with bitumen. For plastic dowels, the maximum slope can be up to 2 mm and for bitumen, 1 mm, referred to a dowel length of 5 mm. Eom et al. [4] confirm that the misfit of dowels can affect the performance of the dowels in the concrete slabs negatively. Several finite element studies of the load transfer of dowels in concrete pavements have been performed, see for example [4, 8, 9]. A literature review shows that dowels have been used for load transfer since the beginning of 19. A thorough overview of the most relevant work regarding dowels in concrete slabs is presented in [9]. A laboratory investigation has been performed at the Swedish Cement and Concrete Research Institute (CBI) in Stockholm regarding bonding properties of dowels. The background to the study is experiences of two 7-year-old concrete road pavements in Sweden. These roads were repaired a couple of years ago and between old slabs and new

M. Löfsjögård / Materials and Structures 38 (25) 721-728 723 slabs, dowels were installed. However, after one or two years, cracks occurred in the slabs and there were also damage at the joints. The joint damage is likely to be linked to the dowels. The reason could be the bonding properties of the dowels used or that the dowels were not mounted perpendicular to the joints. The objective of the test is to investigate the difference in bonding properties regarding steel dowels with two different diameters and different coatings as well as dowels made of composite material (vinyl ester) with two different diameters. 2. LABORATORY TESTS Table 1 - Concrete mixture Swedish concrete grade K5 kg/m 3 Cement, Std Degerhamn 43 Aggregate 8 mm, Underås 855 Aggregate 8-16 mm, Underås 4 Aggregate 16-32 mm, Underås 4 Air entraining agent.42 Plasticiser, 92 M 3 Water (totally) 18 W/C-ratio.4 2.1 Material data The chosen concrete mixture is a road concrete that is used for repair of old concrete roads and therefore easy to work with. The mix proportion was slightly modified to suit the aggregate used in these tests, see Table 1. The cement used in the test, Std Degerhamn, is a low alkali-sulphate resistant cement suitable for civil engineering structures (CEM I 42.5 BV/SR/LA). 2.2 Test specimen A total of 2 test specimens were tested, including five different types of dowels with variation in dowel and coating materials as well as dowel diameter. The moulds were made of plywood and designed in order to secure the dowel in the right position, Fig. 3. The misfit of the dowels must be minimised, since it could affect the results of the tests, cf. Section 1. The test specimen consisted of a concrete block 2 2 25 mm and the dowel was placed in the middle with a length of 2 mm into the concrete block, Fig. 4. For the test, steel dowels with plastic or bituminous coating and dowels made of composite material with no coating were used, Table 2. For the composite dowels, the composite material it self act as a coating. 2.3 Mixing and pouring of concrete The concrete was mixed at CBI with an Eirich Intensive Mixer R 9 T with a capacity for mixing 1 liters of concrete. A total of three batches were produced. The concrete was poured into the moulds after the mixing and then subsequently consolidated on a vibrator table. The concrete was then covered with plastic sheeting for five days and water-hardened the first three days. After five days the plastic sheeting was removed and the moulds were placed in a conditioned room with a temperature of +2 C and a relative humidity of 5 % until the day of testing. The moulds were not removed until just before testing since they were protecting the dowels and made it possible to move the test specimens in a secure way that would not affect the dowels before testing. Concrete was also poured into test cubes for determining the 28-day strength of the concrete. 2.4 Testing The tests were carried out on 2 test specimens. The Fig. 3 - Mould with dowel. Fig. 4 - Test specimen. bonding properties of the dowels were tested in a MTS Material Test System 81 machine. The specimens were placed in a steel box, especially designed and constructed for this test, in the testing machine, Fig. 5. The tests were performed according to the following test procedure: 1. The dowel is drawn-out 1.5 mm with a speed of.5 mm/s. 2. The dowel is then pushed back 1.5 mm to mm with a speed of.5 mm/s. Step 1 and 2 are then repeated 4 times. 3. The dowel is drawn-out 5 mm with a speed of.5 mm/s. With the chosen test procedure, the author aimed to simulate, as close as possible, the behaviour of the dowel in real life within a limited budget. A computer measured the draw-out forces and push-back forces continuously during the whole test. 3. TEST RESULTS 3.1 Consistency and temperature The consistency of the concrete was determined by carrying out a slump test when the concrete was mixed. For

724 M. Löfsjögård / Materials and Structures 38 (25) 721-728 Name PY26-1 PY26-2 PY26-3 PY26-4 PY17-1 PY17-2 PY17-3 PY17-4 PB17-1 PB17-2 PB17-3 BB17-1 BB17-2 BB17-3 CV25-1 CV25-2 CV25-3 CV19-1 CV19-2 Table 2 - Dowel specification Diameter incl. coating (mm) Material Coating Coating thickness (mm) 26.4 Steel Polyethene, yellow 1.6 17.4 Steel Polyethene, yellow 1.6 17.4 Steel Polyethene, black 1.6 16.2 Steel Bitumen.2 24.8 Vinyl ester 2 Vinyl ester 3 19. Vinyl ester 2 Vinyl ester 3 CV19-3 1 = according to the manufacturer, there is no difference between the material except colour. 2 = a composite material. 3 = the composite material itself acts as a coating. Fig. 5 - Test specimen in its test box in the testing machine. the three batches, the slump measurement was 9, 85 and 95 mm. The air content was determined for the three batches as 7.8, 7.6 and 7.6 % using the Pressure method (Swedish Standard SS 13 71 24). During the tests, the temperature was around + 2 C. 3.2 Compressive strength The compressive strength was determined by compressing a total of six concrete cubes in a compression press (Swedish Standard SS 13 72 1). The average compressive strength of the concrete after 28 days was 49 MPa. 3.3 Initial draw-out force Table 3 shows the result from the initial draw-out period. In Table 3, it can be observed that an increase in dowel diameter increases the draw-out force. This holds for both steel dowels with polyethene coating and dowels made of composite material. According to the manufacturer of the steel dowels, there should be no difference between the dowels with polyethene of different colour. However, the results show a difference between the two polyethene coatings. Before drawing any conclusions about that, more tests have to be made in order to determine the difference statistically. The bonding between concrete and dowel should be as low as possible and the test results show clearly that steel dowels with bitumen coating have the lowest bonding of all tested dowels. Compared to steel dowels of the same diameter and a plastic coating, the steel dowels with bitumen only need a draw-out force that is approximately 1 % of the force needed for the steel dowels with plastic coating (yellow and black). The steel dowel with bitumen also shows better results than the composite dowels. The results from the tests presented in this paper are lower than those presented by Eisenmann et al. [5], cited above in Section 1. For the bituminous dowels the results were considerable lower. One explanation can be the type of bitumen on the dowels and also the speed of the draw-out force. The Eisenmann investigation used a test speed of Table 3 - Measured maximum draw-out force for the initial draw-out travel of 1.5 mm Dowel Name Maximum draw-out force (kn) Position for max. draw-out force (mm) Steel with yellow polyethene ø = 25 mm Steel with yellow polyethene ø = 17 mm Steel with black polyethene ø = 17 mm Steel with bitumen ø = 17 mm Composite vinyl ester ø = 25 mm Composite vinyl ester ø = 19 mm PY26-1 PY26-2 PY26-3 PY26-4 PY17-1 PY17-2 PY17-3 PY17-4 PB17-1 PB17-2 PB17-3 BB17-1 BB17-2 BB17-3 CV25-1 CV25-2 CV25-3 CV19-1 CV19-2 CV19-3 14.8 14.5 14.3 15.8 14.9 5.6 3.4 5.3 6.9 5.3 4.7 7.5 7.9 6.7.4.61.44.48 4.7 5.2 5.9 5.3 1.4 1.5 1. 1.3.54.38.65.46.18 1.5.17.25 1.5.35.34.9.6.15.38.18.3.1.6.4

M. Löfsjögård / Materials and Structures 38 (25) 721-728 725 1 mm/min that can be compared to the test speed used in the test presented in this paper,.5 mm/s (or.3 mm/min). Comparing the same dowel diameter, a higher test speed should result in a higher draw-out force and that has been observed here. Due to a limited budget, it was not possible to further investigate the effect of the test speed in this study. Comparing composite dowels with steel dowels show that composite dowels have a draw-out force that is one third of the force for steel dowels (valid for 25 mm dowels). The force is normally considered to be proportional to the envelope surface of the dowel. However, comparing different diameters (envelope surfaces) for steel dowels with yellow polyethene show that the draw-out force increases 2 to 3 times while the diameter is only increasing with 5%. For the composite dowels an increase in diameter with one third gave an increase in draw-out force of 2 to 5 times. The test budget has unfortunately been too limited to make additional tests and other investigations necessary to explain this difference. It is important to study if the measured forces are high enough to cause any damage in the concrete slabs, usually in forms of cracks. The maximum measured force in the test, 15 kn, gives an average tensile stress in the slab of approximately.25 MPa which is considerably lower then the tensile strength of a concrete classified as Swedish concrete grade K5 (28 day concrete compressive strength = 5 MPa), even at early stages of the concrete hardening. This means that no damage would occur in the slabs. The cracking of the repaired concrete slabs described in Section 1 are therefore likely to have another cause, but finding it was not a part of this paper. 3.4 Cyclic behaviour Fig. 6 shows the cyclic behaviour for two of the dowels tested. In Appendix, all the diagrams of the tested dowels are presented. The cyclic behaviour gives a good image of how the dowel is performing in the concrete during the exposure of loads and temperature changes in the slabs. The push-back force is due to friction between the concrete and dowel. Shown in Fig. 6 and in Appendix, repeating the tests for several cycles, the forces needed to draw-out and push-back the dowel is almost the same for all cycles except for the dowels with bituminous coating. They seem to need a higher push-back force than draw-out force. One explanation can be the difference in testing speed, the pushback is done by a speed that is ten times faster than the speed of the draw-out force. (When analysing the test results for cycles 2 4, the author noticed some difficulties in separating the draw-out force from the push-back force since it is difficult to see where they both start and end.) The results indicate that the dowel and coating are working as they are supposed to do since the forces do not change by repetition. This should, however, be investigated further. The last draw-out to a distance of 5 mm is performed in order to investigate the force needed to draw out the entire dowel from the concrete. The hypothesis is that there is a constant force needed and that this constant force will not increase with increased draw-out distance. The results show that the constant force is not increasing by increased draw-out distance, meaning that the hypothesis that the force needed is constant is still valid. 3.5 Inspection Two specimens were cut into halves in order to inspect the coating of the dowels after the testing of draw-out force. The two dowels were coated with polyethene and the inspections show that the coating is not damaged (Fig. 7). 4. FURTHER RESEARCH For further research, the following items are of special interest. The observed difference between dowels with yellow and black polyethene coating needs further investigation. Especially since the manufacturer states that there is no difference in the material except colour. Further investigation ought to be made regarding the effect of different test speeds and how the test speed affects different types of coating materials. The repeated tests 15 1 5-1 -15 PY 26-2 -5 1 2 3 4 5 BB 17-2 8 6 4 2-2 1 2 3 4 5-4 -6-8 Fig. 6 - Typical cyclic behaviour (observe that the figures have different scales). Fig. 7 - Photo of specimen cut in halves showing that the coating is still intact after the test.

726 M. Löfsjögård / Materials and Structures 38 (25) 721-728 should be done by another test procedure that makes it easier to separate the draw-out force from the push-back force, e.g. by making a stop before the push-back starts. The use of composite dowels for concrete road pavements needs further research, especially their performance during shear forces. The long-term performance and durability of the coating in the concrete is interesting for further investigations. 5. CONCLUSIONS Pavement cracking in the vicinity to the bar ends can be caused by inclined dowel bars and unmovable bars. A laboratory investigation of the bonding properties is presented in this paper and from the test the following conclusions can be drawn. The lowest values of the initial draw-out force were obtained for the dowels with bituminous coating. However, it should be noticed that the speed of the draw-out force is likely to have an affect on the results especially for dowels with bitumen. An increase in the envelope surface (diameter) with 5 % increased the draw-out force 2 to 3 times for the steel dowels with plastic coating. The comparing results for the composite dowels showed that an increase in diameter by one third gave an increase in draw-out force of 2 to 5 times. The composite dowels show lower forces than steel dowels with plastic coating for the same dowel diameter. The measured draw-out forces are causing stresses that are lower than the tensile strength of the slabs meaning that no damage, e.g., cracks, will occur. Therefore, the cracking at the repaired slabs must most likely have another cause. Repeating the test for several cycles shows that the drawout and push-back forces were almost the same for all cycles and dowels except for the dowels with bituminous coating. A higher push-back force was needed compared to the draw-out force, which could also be an effect of the test speed. Inspections show that the plastic coating on steel dowels is intact after testing. ACKNOWLEDGEMENTS The Swedish National Road Administration, Cementa AB and the Swedish Agency for Innovation Systems (through the doctoral project Relationships between functional properties of concrete roads ) are acknowledged for financing the study. REFERENCES [1] Petersson, Ö., Swedish design method for concrete roads (Svensk metod för dimensionering av betongvägar), Bulletin 16, Licentiate Thesis (Department of Structural Engineering, Royal Institute of Technology, Stockholm, 1996) [In Swedish]. [2] Jeuffroy, G. and Sauterey, R. (series editors), Cement Concrete Pavements (A.A. Balkema Publishers, Brookfield, 1996). [3] Eisenmann, J. and Leykauf, G., Dowels in Concrete Pavements, Information brochure (Testing Office for Rural Highway Construction of the Technical University of Munich, 198?). [4] Eom, I-S., Parsons, I.D. and Hjelmstad, K.D., A finite element study of load transfer between doweled pavement slabs, 4 th International Workshop on Design Theories and their Verification of Concrete Slabs for Pavements and Railroads, September 1-11, 1998, Bucaco, Portugal (CROW Record 19, 1998) 249-26. [5] Esienmann, J., Lempe, U. and Deischl, F., Investigation of the influence of dowel type on the transverse load transmission capacity and draw-out resistance (Untersuchung von unterschiedlich ausgebildeten Querfugen hinsichtlich der Querkraftübertragung und des Gleitwiderstandes der Dübel). Forschung Strassenbau und Strassenverkehrstechnik (22) (1976) 11-28 [In German]. [6] Leykauf, G., Personal contact (Prof. Lehrstuhl und Prüfamt für Bau von Landverkehrswegen, Techniche Universität München, 23). [7] Eisenmann, J. and Leykauf, G., The influence of misfit of the dowel on the draw-out resistance (Auswirkung von schräg liegenden Dübeln auf den Ausziehwiderstand). Strasse und Autobahn (11) (1977) [In German]. [8] Scarpas, A., Ehrola, E. and Judycki, J., Simulation of load transfer across joints in RC pavements, 3 rd International Workshop on design and evaluation of concrete pavements, September 29-3, 1994, Knumbach, Austria (CROW Record 14, 1994). [9] Ioannides, A.M. and Korovesis, G.T., Analysis and design of doweled slab-on-grade pavement system, Journal of Transportation Engineering 118 (6) (1992) 745-768. APPENDIX - Diagrams of the tested dowels 15 1 5-1 -15 15 1 5-1 -15 15 1 5-1 -15 PY 26-1 -5 1 2 3 4 5 PY 26-2 -5 1 2 3 4 5 PY 26-3 -5 1 2 3 4 5

M. Löfsjögård / Materials and Structures 38 (25) 721-728 727 15 1 5-5 -1-15 PY 26-4 1 2 3 4 5 PY 17-1 8 6 4 2-2 1 2 3 4 5-4 -6-8 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 PY 17-2 1 2 3 4 5 PY 17-3 1 2 3 4 5 PY 17-4 1 2 3 4 5 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 PB 17-1 1 2 3 4 5 PB 17-2 1 2 3 4 5 PB 17-3 8 6 4 2-2 1 2 3 4 5-4 -6-8 8 6 4 2-4 -6-8 BB 17-1 -2 1 2 3 4 5 8 6 4 2-2 -4-6 -8 BB 17-2 1 2 3 4 5

728 M. Löfsjögård / Materials and Structures 38 (25) 721-728 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 BB 17-3 1 2 3 4 5 CV 25-1 1 2 3 4 5 CV 25-2 1 2 3 4 5 CV 25-3 1 2 3 4 5 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 8 6 4 2-2 -4-6 -8 CV 19-1 1 2 3 4 5 CV 19-2 1 2 3 4 5 CV 19-3 1 2 3 4 5