A New Load Transfer Assembly for the Jointed Concrete Pavements

A New Load Transfer Assembly for the Jointed Concrete Pavements Alireza Zeinali, Kamyar C. Mahboub, Herbert F. Southgate Doctoral Candidate, Department of Civil Engineering, Raymond Bldg., University of Kentucky, Lexington, KY - (corresponding author). Phone: () -, E-mail: azein@uky.edu Professor, Department of Civil Engineering, Raymond Bldg., University of Kentucky, Lexington, KY -. Phone: () -, E-mail: kmahboub@engr.uky.edu Retired Highway Research Engineer, Sherwood Drive, Lexington, KY. Phone: () -, email: hfsouth@windstream.net Submission Date: // Number of Tables: Number of Figures: Text Word Count (including the abstract and references):, Total Word Count (with words added for each table and figure):,

Zeinali, Mahboub, Southgate ABSTRACT Jointed concrete pavements continue to suffer from joint failure as the main distress and rehabilitation issue. The shear stress in the vicinity of dowel bars which is resulted from heavy traffic loads is the major factor that contributes to initiation of microcracks and joint faulting. A new load transfer assembly is introduced in this paper to be used in place of the conventional dowel bars. The introduced load transfer assembly was approved by the United States Patent Office as a new invention. The assembly comprises a spine and a plurality of dowel bars projecting from the spine where the spine comprises an elongated, freely rotating hinge. Finite element modeling was employed to evaluate the effect of using the new apparatus on the reduction of shear stress in concrete slabs. The finite element model was solved under various loading conditions and subgrade support. The obtained results showed that the suggested system reduces the shear stress level in the concrete by to percent. This amount of reduction in the shear stress can significantly reduce the shear-induced cracking near the joint. Furthermore, using the new load assembly was shown to be more effective on shear stress level than stabilizing the subgrade. The new invention has the capability of addressing curling and warping induced stresses as well as horizontal movement due to contraction and expansion. Furthermore, the new system prevents the dowel bars misalignment during new construction, as well as expediting the process of retrofitting damaged joints.

Zeinali, Mahboub, Southgate BACKGROUND A major advantage of rigid pavements is the low deflection under traffic loadings due the high modulus of elasticity of their surface course. Typically the surface course of a rigid pavement is made of portland cement concrete (PCC) and regarding the effect of thermal expansion and shrinkage, most of the PCC pavements (PCCP) are built with separating joints between slabs in order to provide extra space for movement due to expansion and shrinkage. However, poor load transfer causes pumping, which is a major distress mode in PCCP. Large deflection at slab joints necessitates a load transfer system to bond the adjacent slab together as an axle load approaches the joint. Load transfer is crucial for the durability of jointed plain concrete pavements (JPCP). Poorly performing joints are the main cause behind most performance problems of concrete pavements such as faulting, pumping, and corner breaks. Havens () included results of Road Rater tests made on a concrete slab in good condition on I and on several concrete slabs on I with varying degrees of distress at contraction joints. The Road Rate was manufactured by Foundation Mechanics, Inc., El Segundo, CA. KY DOT Research Division had the only Road Rater model that was equipped with four velocity pickup sensors and the capability of varying the frequency of loading in cps intervals. The normal test conditions for the Road Rater consisted of a static load of lbs ( N) and a dynamic load of lbs ( N) peak-to-peak at a frequency of cps. For testing of concrete pavements, the meaning of absolute values was not important; rather the differential in readings was important. Low Road Rater readings resulted when the pavement was in contact with the aggregate base. Higher readings were associated with the condition of non-contact between the bottom of the concrete and the surface of the aggregate base. For test locations near the edge between the concrete and paved shoulder, previous tests at other locations showed that higher readings can also be associated with moisture under the edge of the pavement and/or some deterioration at the bottom of the concrete. At am, deflections at the ends of the slabs were relatively higher than at the mid-length of the slabs. This indicated that the exposed surface was attempting to contract relative to the base of the slab. In summary, the cool night air resulted in the pavement surface temperature to be cooler (contracting forces) than the bottom of the concrete that was still being heated by the base beneath (expanding forces). Conclusion: there are contracting forces at the top and tensile forces at the bottom resulting in the slabs curling upward at the ends. By noon, deflections at all test points were relatively equal indicating that the temperature distribution with depth in concrete was nearly equal due to heat from the sun. By about : pm, deflections at the mid-length of the slabs were higher than at both sides of a contraction joint, and very nearly equal across the lane width at the contraction joints. By pm, deflections revealed that the slabs were perched off the base but supported ONLY at the four corners of each slab. Thus, the slabs behaved similar to diving boards at swimming pools early in the morning (anchor points located at mid-length), fully supported at mid-day, perched across the lane width at the contraction joints by mid-afternoon, and supported only at the four corners by sundown. Interlock of the aggregates at the two sides of a joint used to be the sole means of transferring traffic load across the joint (). However, the efficiency of this kind of joints decreases drastically as the pavement ages and the joint face loses its friction. Dowel bars are used nowadays as the permanent load transfer tool over concrete joints. Dowel bars are typically

Zeinali, Mahboub, Southgate installed at the mid-depth of the slabs and coated with a substance to prevent bonding to the concrete and corrosion. As an axle load passes over a joint with dowel bars, highly concentrated compressive and shear stresses are induced in beneath and around the dowel bars. It has been shown that the bearing stress beneath the dowel bars is not critical. However, dowel bars with elliptical cross sections cause lower bearing stress than those with circular or rectangular sections (). In another research study, using Concrete-filled, glass fiber-reinforced polymer dowels was suggested to reduce the bearing stress (). Although the proposed methods can improve the performance of dowel bars, the issue of shear-induced cracking is yet unresolved. These cracks initiate at the interface of the dowel bar and concrete as the result of repeated high shear stresses imposed by traffic loading. Shear-induced microcracks grow at a downwards to the subgrade and upwards to the pavement surface. At the end these cracks form sizeable spallings on top and bottom of concrete slab and the joint practically loses its load transfer abilities. Misalignment of dowel bars is one of the construction related issues of the concrete pavements. The dowel bars that are not installed exactly perpendicular to the joint face constrain the expansion of the slabs and cause tensile stress in the concrete. One of the suggested solutions to the issue is to use dowel bars with square cross section and attach a strip of a flexible material to its sides. This gives the dowel bars some extra flexibility in case of misalignment (). Since the joint faulting is a relatively common distress in JPCPs, repairing the damaged joints has become a vital part of rehabilitation of concrete pavements. In a full-depth repair of a deteriorated joint, the damaged part of the adjacent slabs is wholly replaced by new concrete. The newly poured concrete is attached to the old concrete cut faces through a number of new dowel bars. A research on long-term performance of repaired joints showed that they are highly likely to show the same cracking after the retrofit (). A new load transfer assembly is proposed in this paper to use with the joints of jointed concrete pavement. The introduced system and apparatus was approved by the United States Patent Office as a new invention in year (). The analysis showed that the shear-induced cracking is reduced by using the load transfer assembly. Other issues regarding the construction of concrete pavements were also considered in the design of the new system such as dowel bar misalignment, retrofit application, and expansion of concrete slabs. INTRODUCTION The load transfer assembly introduced in this paper is a new apparatus which would accommodate movement between adjacent slabs of a JPCP due to thermal expansion and traffic loads while transferring the load from one slab to another. The apparatus comprises a dowel bar having a first end, a second end and a freely rotating hinge provided along the middle point of the dowel bar. As shown in Figure, any number of desired dowel bars can be fabricated together to generate an assembly of dowel bars. The hinge at the mid-span of the dowel bars itself comprises two tubes with different radii in a way that the inner tube can slide and rotate inside the outer tube. This hinge provides a rotational degree of freedom for all of the dowel bars whereas the vertical wheel loads are transferred to the adjacent slab efficiently.

Zeinali, Mahboub, Southgate Dowel Bars Top Ridge Inner Tube Outer Tube FIGURE New load transfer assembly Two thin and narrow ridges may also be mounted to the top and bottom side of the outer tube. As shown in Figure, these ridges are used during the construction of the pavement to keep a flexible spacer in place between two slabs. The spacers are made of an impermeable and flexible material in order to provide adequate space for thermal expansion of slabs during the summer time. In order to assure a free slip condition, at least one end of the dowel bars must be embedded in a polymer tube or sleeve as illustrated in Figure. In addition, the contact area of the dowel bar and its sleeve must be coated with a friction reducer material to provide a free slip conditions for the design life of the pavement. Flexible Spacer Sleeve Dowel Coated with Friction Reducer Plastic Cap FIGURE Details of the flexible spacer, slip sleeve and plastic cap for the load transfer assembly

Zeinali, Mahboub, Southgate In addition to traffic loads, temperature curling of the concrete slabs has a deteriorating effect on growth of shear-induced cracks. As a result of temperature differential at the top and bottom of concrete slabs, they tend to curl upward or downward based upon the time of the day and changes in air temperature. The curling imposes a high bending moment on dowel bars and as a consequence the shear stress in the concrete rises considerably. A research study showed that adding an open-graded asphalt rubber friction course over the concrete reduces the temperature fluctuation of the slabs between day and night (). The design of the load transfer assembly provides free rotation for the dowel bars at their mid-span. As a result, no bending moment is generated in the dowel bars when the concrete slabs curl. This can relieve the shear stresses due to temperature curling as well as increasing the durability of the concrete joints. FINITE ELEMENT MODELING Finite Element Model Finite element modeling was employed to evaluate the effect of using the new load transfer assembly on behavior of concrete pavements at transverse joints. The software ANSYS was used to generate and solve the finite element models. A ft (. cm) standard lane of a concrete pavement was modeled. The distance between the transverse joints was assumed to be ft (. cm). The dimensions and material properties of the modeled concrete slabs and dowel bars are shown in Table. Due to the characteristics and geometry of the model, the element SOLID of the ANSYS library was used to generate the finite element mesh of the concrete slabs. This element is defined by eight nodes with three degrees of freedom for each node. Since all the angles of the concrete slab were right angles, no prism or tetrahedral elements were used in the mesh. All the elements in the concrete slab had equal dimensions in order to minimize the analysis error for the different loading and load transfer conditions. The element BEAM was utilized to model the dowel bars across the transverse joints. BEAM is a three dimensional elastic beam which is defined by two points and six degrees of freedom at each point. This uniaxial element is capable of modeling torsion, bending, tension, and compression for beams. A liquid foundation was also assumed as the subgrade of the concrete pavement. The liquid foundation was modeled by CONTAC element. This element only has the capability to support compression. When the two contacted surfaces are separating from each other, this element does not impose any tensile force. As the results, there was no tensile bonding between the concrete slab and subgared. The stiffness of the element was adjusted to model the assumed California bearing ration (CBR) of the subgrade. The finite element model was solved for different loading positions and magnitude. The load of a dual tire axle was modeled at three different positions on the slab as follows (Figure ): middle slab (equal distances from two joints of the slab), over joint (the tires contact area equally distributed on both slabs), joint edge (the whole contact area of the tires are placed on one side of the joint).

Zeinali, Mahboub, Southgate TABLE Dimensions and Material Properties for the Finite Element Model Property Magnitude Concrete Slab Width ft (. cm) Length ft (. cm) Height in. (. cm) Density. lb/in ( kg/m ) Modulus of Elasticity psi ( Mpa) Dowel Bars Diameter. in. (. cm) Length in. (. cm) Distance in. (. cm) Modulus of Elasticity psi ( MPa) For every axle positions mentioned above, the finite element model was solved for three different axle loads as follows: kip (. kn), kip (. kn), and kip (. kn). An in. (. cm) traffic wander space was assumed between the slab edge and dual tire. For every combination of axle load and position, the model was generated with three subgarde CBRs: %, %, %. Over-Joint Middle Slab Joint Joint Edge Dowel Bars FIGURE Different axle positions for the finite element model In the primary evaluations, the finite element model was consisted of concrete slabs. The results showed that the effect of a single loading axle does not reach beyond the adjacent slabs. In order to save in the calculation and processing time, the main models were all generated with three concrete slabs and the loading axle was modeled on the middle one.

Zeinali, Mahboub, Southgate It should be noted that the finite element model was generated based upon the worst-case scenario, and hence the calculated maximum shear stresses are based upon this scenario. For example, the load transfer between slabs offered by aggregate interlock at the joint was assumed to be zero. Additionally, the opening of the slab joint was assumed to be large enough to allow slab rotation without any joint interface contact. These scenarios exposed the new load transfer assembly to the most severe loading conditions without any assistance from neither the slab-joint interface nor the road shoulders. Postcalculations and Analysis The finite element model was run with different loading and subgrade support conditions. The shear and compressive stresses and strains at the critical nodes were then determined through the postcalculation procedures and the results were recorded. Figure shows the calculated shear stress in the joint face at its horizontal centerline. The finite element model for this figure was run with an kip axle load located right at the edge of the joint. A weak subgrade support with a California bearing ratio (CBR) of % was assumed in this model and conventional dowel bars were used in the model as load transfer device. Each major gridline in the chart represents the location of a dowel bar. The origin of the horizontal axis in Figure is assumed to be at the center point of the dowel bar where the shear stress reaches its maximum in the joint face. This dowel bar is placed exactly beneath the centerline of a dual tire wheel width and it is referred to as critical dowel bar henceforth in this paper. As seen in Figure, there was a high concentration of shear stress in the vicinity of the dowels bars beneath the axle tires. This could result in initiation and growth of shear cracks in these areas. (mm) -,,, Joint Face x Critical Dowel Subgrade CBR =% Dowel Bars in. S XY (psi) - - (kpa) - - - - - Horizontal Distance from the Critical Dowel (in.) FIGURE Shear stress in the joint face for an kip axle on the joint edge and conventional dowel bars NOTE: x= means center-point of the critical dowel bar section on the joint face

Zeinali, Mahboub, Southgate Figure shows the variation of shear stress along the critical dowel bar with an kip (. kn) axle located at three different locations as follows: joint edge, over the joint and middle slab. The three chart lines in Figure represent three axle positions on the slabs. The origin of the horizontal axis in this chart is assumed to be at the interface of the critical dowel bar and the joint face. As can be seen in the figure, the highest level of shear stress was imposed in the joint face and when the axle load was located right at the edge of the joint. For this loading condition, it was assumed that the axle load was entirely applied on one slab and the load was transferred to the adjacent slab by the means of dowel bars. Since the axle load was equally applied on both adjacent slabs in the over-joint loading condition, the dowel bars did not transfer any load between the two faces of the joint. As a result, the shear stress level was considerably lower than in the joint edge loading condition. Likewise, an axle located at the middle of a slab did not impose a considerable amount of shear stress in the vicinity of the joint. S XY (psi) - (mm) Dowel Bar Joint x Joint Edge Loading (kpa) Axle Load = kip Subgrade CBR= % Dowel Type: Ordinary Dowel Material: Steel Dowel-Slab Slip Condition: Unbound Axle Position Joint Edge Over-Joint - - Midddle Slab Horizontal Distance from Joint (in.) FIGURE Shear stress along the critical dowel bar for different loading positions NOTE: x= means center-point of the critical dowel bar section on the joint face The finite element model was also run for the new load transfer assembly with dimensions and loading conditions identical to the model for the conventional dowel bars. In order to model the hinge action of the apparatus, a rotational degree of freedom was generated at the mid-span of each dowel bar. In other words, it was assumed that each dowel has its separate hinge. Consequently, the apparatus was modeled with the worst case scenario. A comparison between using the new load transfer assembly and ordinary dowel bars in terms of their effect on the resulting shear stress is showed in Figure. The chart depicts the variation of shear stress

Zeinali, Mahboub, Southgate along the critical dowel bar for two axle positions. As can be seen in the figure, using the new dowel assembly results in a percent reduction in shear stress in the vicinity of the transverse joint. The shear stress for the over-joint axle position is shown in Figure as well. The analysis showed that using the new dowel bar assembly does not have any negative impact on the load transfer efficiency of the joint in any of the loading combinations. (mm) Dowel Bar Joint x Axle Load = kip Subgrade CBR = % Dowel Material: Steel Dowel-Slab Slip Condition: Unbound S XY (psi) - - Horizontal Distance From Joint (in.) - - (kpa) Axle Position- Dowel Type Joint Edge - Ordinary Joint Edge- Hinged Over-Joint- Ordinary Over-Joint- Hinged FIGURE Shear stress along the critical dowel bar for the new load transfer assembly NOTE: x= means center-point of the critical dowel bar section on the joint face The effect of using the new load transfer assembly instead of the conventional dowel bars on the resulting shear stress is shown in Figure for various axle loads. The origin of the horizontal axis of this chart was assumed to be at the joint face. As seen in Figure, the shear stress along the critical dowel bar was reduced for all the axle loads as well as the magnitude of the negative shear stress at the end of dowel bar and inside the concrete. Further analysis showed that the same amount of reduction took place in the shear stress in the whole joint face area. The stress analysis of the finite element models showed that the maximum resulting shear stress in the transverse joint area varies linearly by changes in the axle load. As presented in Figure, for a jointed concrete pavement placed on a subgrade with % CBR, the maximum shear stress raises by. psi (. kpa) on average, for every kip (. kn) increase in the axle load. Whereas, using the new load transfer assembly reduces the slope of the maximum shearaxle load chart to. psi/kip (. kpa/kn).

Zeinali, Mahboub, Southgate S XY (psi) - (mm) Dowel Bar Joint Horizontal Distance from Joint (in.),, FIGURE Shear stress along the critical dowel bar for different axle loads and dowel types x - - (kpa) Subgrade CBR = % Dowel Material: Steel Dowel-Slab Slip Condition: Unbound) Axle Load - Dowel Type kip- Ordinary kip- Ordinary kip- Ordinary kip-hinge kip-hinge kip-hinge (kn) Dowel Material: Steel Dowel-Slab Slip Condition: Unbound Maximum S XY (psi) y z Critical Dowel Dowel Bars Axle Load (kip) FIGURE Maximum shear stress for different subgrade CBR and axle loads (kpa) Subgrade CBR - Dowel Type CBR=% - Ordinary CBR=% - Hinged CBR=% - Ordinary CBR=% - Hinged CBR=% - Ordinary CBR=% - Hinged

Zeinali, Mahboub, Southgate The effect of subgrade support on the load transfer efficiency of the joint was also evaluated. Figure depicts the effect of subgrade support on the maximum shear stress in the concrete slab for both ordinary dowel bars and new load transfer assembly. The results showed that utilizing the new load transfer assembly is more effective on reducing the shear stress than increasing the subgrade CBR from % to %. The subgrade modulus is typically increased in the filed by the means of chemical stabilizers. The stabilization of the subgrade is a costly and time consuming process. In addition to increasing the durability of the concrete pavements, using the new load transfer assembly also benefits the pavement construction process economically. Critical Shear Stress According to ACI code (), for non-pre-stressed concrete members subjected only to shear and flexure stresses, the member shear capacity V c is related to concrete compressive strength in the form of the following equation: V c = f c b w d () where V c = shear strength of the concrete, lb f c = compressive strength of the concrete, psi b w = width of the member, in. d = distance from extreme compression fiber to centroid of tension reinforcement, in. Therefore, the permitted shear stress for concrete slabs (assuming a uniformly distributed stress with a typical concrete compressive strength of f c = psi (. MPa) to psi (. MPa)) would approximately be: v c = to = to psi ( to kpa) TABLE Calculated Maximum Shear Stress in Concrete Slab for a Single Axle Load Applied at the Joint Edge Subgrade CBR Single Axle Load, kip (kn) Maximum Shear Stress, psi (kpa) Ordinary Dowel Bars New Load Transfer Assembly % (.) (.) (.) (.) (.) (.) (.) (.) (.) % (.) (.) (.) (.) (.) (.) (.) (.) (.) % (.) (.) (.) (.) (.) (.) (.) (.) (.)

Zeinali, Mahboub, Southgate The maximum calculated shear stress from the finite element analysis is presented in Table for various loading and subgrade support conditions. This table shows that the load transfer assembly may make a significant contribution to moderating the maximum shear stresses. A comparison between the maximum shear stresses in Table and the shear strength of the concrete slab shows that shear failures are highly likely to happen around the joints. Since the magnitude of the induced shear stress is close to the maximum shear strength of the concrete, reducing the shear stress level by to percent may improve the durability and performance of the load transfer joints considerably. This reduction in the stress level can be achieved by designing the pavements with the load transfer assembly in place of the ordinary dowel bars. CONSTRUCTION AND REFTROFIT Misalignment of the dowel bars which is a common problem during the construction of concrete pavements can be prevented by using the suggested apparatus. For a new road construction, the entire assembly is transported to the construction location and placed at the specified point using the dowel baskets. Due to the integrity of the dowel bars and the spine tube, the bars are entirely restricted to stay at their initial positions during the process of construction and the dowel misalignment is avoided. Ease and simplicity of the installation of the introduced apparatus makes it a practical and durable alternative for retrofitting the deteriorated concrete joints. First, the damaged area of the pavement requiring replacement is cut and removed. Next, horizontal holes are located and drilled in exposed faces of the remaining roadway. The holes must be drilled perpendicular to the cut face so that the proper alignment of the load transfer apparatus is assured. Next, epoxy or other suitable substitute adhesive is applied to the outer surface of the sleeves and the sleeves are immediately installed in the drilled holes. The dowel bar baskets are then properly aligned with the future joint. If not already installed, the inner tube and dowel bars are installed along with the flexible spacer. The outer pipe is then positioned on the baskets in alignment with the future joint. Figure (a) illustrates the retrofit of the damaged joint at this stage. Finally, the concrete for the repair patch is poured and allowed to cure according to construction specifications. Plastic cap can be installed on the tubes ends to assist in keeping fresh concrete from entering the hinge. Figure (b) is a schematic view of the final retrofitted joint.

Zeinali, Mahboub, Southgate (a) (b) FIGURE Retrofitting the damaged concrete joints with the load transfer assembly, (a) after drilling the dowel holes and placement of the assembly, (b) final retrofitted joint CONCLUSIONS A high amount of shear stress is imposed near the joints of concrete pavements as a moderate to heavy axle load passes over a joint. Due to the induced shear stress, occurrence of the shear cracks is highly likely for jointed concrete pavement. Employment of the new load transfer assembly at its worst scenario reduces the shear stress level in concrete by to percent.

Zeinali, Mahboub, Southgate Using the load transfer assembly results in a significant reduction in potential for shearinduced cracking of concrete pavements. The introduced load transfer assembly reduces the deteriorating effect of temperature curling of concrete slabs significantly. Dowel bar misalignment is avoided during the construction of concrete pavements by using the prefabricated load transfer assembly. Due to ease and simplicity of its installation, the load transfer assembly is a reliable alternative for damaged joints repair and retrofit. REFERENCES. Havens, J. H. The D-Cracking Phenomenon: A Case Study for Pavement Rehabilitation. Research Report. Division of Research, Bureau of Highways, Kentucky Department of Highways,, pp. -.. Nowlen, W. J. Influence of Aggregate Properties on Effectiveness of Interlock Joints in Concrete Pavements. Journal of the PCA Research and Development Laboratories, Vol., No.,, pp. -.. Porter, M., N. Pierson. Laboratory Evaluation of Alternative Dowel Bars for Use in Portland Cement Concrete Pavement Construction. In Transportation Research Record: Journal of the Transportation Research Board, No., Transportation Research Board of the National Academies, Washington, D.C.,, pp. -.. Murison, S., A. Shalaby, A. Mufti. Concrete-Filled, Glass Fiber-Reinforced Polymer Dowels for Load Transfer in Jointed Rigid Pavements. In Transportation Research Record: Journal of the Transportation Research Board, No., Transportation Research Board of the National Academies, Washington, D.C.,, pp. -.. Schrader, E. K. Solution to Cracking and Stresses Caused by Dowels and Tie Bars. Concrete International, Vol., No.,, pp. -.. Pierce, L. M., J. Uhlmeyer, J. Weston, J. Lovejoy, J. P. Mahoney. Ten-Year Performance of Dowel-Bar Retrofit: Application, Performance, and Lessons Learned. In Transportation Research Record: Journal of the Transportation Research Board, No., Transportation Research Board of the National Academies, Washington, D.C.,, pp. -.. Southgate, H. F., K. C. Mahboub, A. Zeinali. Load Transfer Assembly. U.S. Patent Cl. -. Filed September,, and issued June,,.. Belshe, M., M. S. Mamlouk, K. E. Kaloush, M. Rodezno. Temperature Gradient and Curling Stresses in Concrete Pavement with and without Open-Graded Friction Course. Journal of Transportation Engineering. Vol., No.,, pp. -.. ACI, -. Building Code Requirements for Structural Concrete (ACI -) and Commentary. American Concrete Institute, Farmington Hills, MI..