Design Guide for Precast UHPC Waffle Deck Panel System, including Connections

CHAPTER 1. UHPC AND WAFFLE DECK SYSTEM

Introduction

The combination of aging infrastructure, increasing numbers of structurally deficient or obsolete bridges, and continuous increases in traffic volume in the US demands rapid improvements to the Nation's bridge infrastructure with an emphasis on increasing bridge longevity. The increased emphasis on work zone safety and user costs associated with traffic delays, as well as quality and environmental impacts of the construction process, require development of technologies and structural details suitable for rapid construction. In consideration of these challenges, the Federal Highway Administration (FHWA) has been promoting accelerated bridge construction (ABC) methods using prefabricated bridge elements.

In the context of ABC, precast concrete deck panels are being utilized more and more by several State departments of transportation (DOTs) for both bridge deck replacements and new structures to reduce construction time. ⁽¹⁾ Previous studies have shown that the use of prefabricated full-depth precast concrete deck systems can accelerate the construction and rehabilitation of bridge decks significantly, extend service life, lower life-cycle costs, and minimize delays and disruptions to the community.^(2,3,4)

However, transverse connections used previously between precast bridge deck panels have exhibited various serviceability challenges due to cracking and poor construction of connections._(5,6) Therefore, it is imperative that durable and efficient field connections be developed to implement precast deck panels successfully in practice. These connections could utilize high-performance materials such as ultra-high-performance concrete (UHPC). In addition to these materials adhering well with precast components, the connection should be detailed to prevent cracking and leakage along the connection interfaces between precast elements.

UHPC is a newly developed concrete material that exhibits high compressive strength, dependable tensile strength, and excellent durability properties including very low permeability. The superior structural characteristics and durability of UHPC are perceived to provide major improvements over ordinary concrete and high-performance concrete (HPC) bridges in terms of long-term structural efficiency, durability, and cost-effectiveness. Hence, the construction of new bridges and renewal of aging highway bridges using UHPC has been explored in terms of improving construction efficiency, enhancing bridge performance, and reducing maintenance and life-cycle costs.

Previous use of UHPC for bridge applications (mostly in bridge girders) in the US has proven to be efficient and economical. (See references 7, 8, 9, and 10.) This guide focuses on a new bridge deck application.

A prefabricated UHPC waffle deck system with field-cast UHPC connections was developed as part the FHWA Highways for LIFE (HfL) program by combining the advantages of UHPC with those of precast deck systems. An integrated experimental and analytical study was performed to evaluate the performance of the precast UHPC waffle deck system and UHPC connections under laboratory and field conditions.

This study found that the UHPC waffle deck system performed extremely well under service and fatigue loading conditions. In addition, the ultimate capacity tests revealed that the UHPC waffle deck system has significantly higher capacity than the required design level capacity, suggesting potential improvements to the design of the UHPC waffle deck system and reduction in the construction costs. The benefits of the UHPC waffle deck system, along with a summary of the experimental and field studies, are presented later in this chapter.

Given the success of the precast UHPC waffle deck system and increased interest in full-depth precast deck panels for ABC, the goal of this guide is to increase the awareness, improve efficiency, and broaden the use of UHPC waffle deck systems for new and replacement bridges.

This guide provides the technical and practical information necessary to allow future bridge owners to consider the use of UHPC waffle slabs in a wide variety of bridge types. Detailed recommendations for the design, detailing, and construction of a full-depth precast UHPC waffle deck system are presented. An introduction to the UHPC waffle deck system, material properties of UHPC, and design guidelines are presented in the next sections.

Ultra-High-Performance Concrete

UHPC is defined worldwide as concrete with a compressive strength of at least 22 ksi. ⁽¹¹⁾ In recent years, Lafarge North America has been marketing Ductal, a form of UHPC with steel fibers, which regularly achieves compressive strengths of 24 to 30 ksi. UHPC is an advanced, highly engineered, cementitious material consisting of typical portland cement, fine aggregate made of sand, silica fume, crushed quartz, steel fibers, super plasticizers, and high water reducers. The typical composition of UHPC is shown in table 1.

Table 1. Material composition of typical UHPC mix.⁽¹²⁾

Material	Amount, lb/yd3	Percent by Weight
Portland Cement	1200	28.5
Fine Sand	1720	40.8
Silica Fume	390	9.3
Ground Quartz	355	8.4
Super plasticizer	51.8	1.2
Accelerator	50.5	1.2
Steel Fibers	263	6.2
Water	184	4.4

A few notable differences in the UHPC composition when compared to HPC are the lack of typical coarse aggregate, addition of steel fibers, high proportions of cementitious materials, and low water/cement ratio. The use of powder and well-graded components helps in achieving a high packing density of the UHPC constituents and leads to significantly improved mechanical properties such as increased compressive strength and considerable tensile strength as compared to HPC and normal-strength concrete (NSC). In addition, UHPC demonstrates a very dense cement matrix creating a low permeability concrete, which greatly enhances its resistance to corrosion and degradation.

In precast environments, UHPC is commonly subjected to heat treatment at 194 °F at 95 percent humidity conditions to accelerate the full development of its strength and durability properties. However, this is not a requirement. Ambient curing of UHPC is also appropriate depending on the constraints set forth by the specific application. For example, the field-cast UHPC connections between the precast elements are not typically heat cured. More details about the curing conditions and its influence on UHPC material behavior are presented in the material properties section below.

The use of steel fibers in UHPC improves the material's ductility as well its tension capacity. In addition to the advantages realized from its superior mechanical and durability properties, with the use of super plasticizers in the mix design, UHPC displays a self-consolidating/self-leveling behavior, and this allows it to be placed in the plant and field conditions with little to no vibration, reducing construction costs.

The properties of UHPC suggest the potential to improve the overall economy of construction projects significantly. The high compressive strength of UHPC allows designers to select smaller sections for members, decreasing dead load on the structure and improving overall structural efficiency. UHPC also displays rapid early strength gain, which, along with its suitability for precast/prestressed applications, can contribute to reduced construction times. UHPC's appreciable tensile and shear capacity due to the inclusion of steel fibers may often lead to total or partial elimination of conventional mild steel reinforcement. Finally, the superior durability characteristics of UHPC should contribute to an increased service life and reduced maintenance costs compared to conventional concrete structures in nearly all applications. Therefore, in the current-day concrete technology, UHPC can arguably be considered a nearly ideal structural material. However, its use has not reached widespread acceptance in industry due to a lack of design guidelines and the cost of the material when compared to NSC.

For designers to accept and feel comfortable designing UHPC structures, there is a need to optimize structural shapes and to develop a comprehensive understanding of UHPC material behavior. Therefore, this guide draws upon previous research on UHPC material properties and provides recommendations for the use of UHPC in developing durable bridge decks.

UHPC in Federally Funded Projects

The use of UHPC in Federal-aid highway construction projects is subject to the Buy America provisions found in 23 U.S.C 313 and 23 C.F.R. 635.410. These provisions pertain to the obligation of funds to a project wherein steel is used and require that the steel be manufactured in the United States. The FHWA's Buy America regulatory policy allows for a minimal use of foreign steel as defined in 23 CFR 635.410(b)(2). The policy also allows a waiver from the Buy America requirements where the requirements would be inconsistent with the public interest, or where domestic steel or iron products are not produced in sufficient and reasonably available quantities which are of satisfactory quality.

UHPC commonly includes steel fiber reinforcement that is subject to these provisions. As of May 2013, some steel fiber suppliers have expressed an interest in producing Buy America compliant steel fibers. These suppliers are working to develop and demonstrate the performance of fibers with the appropriate concrete manufacturers. If a State has a project involving the use of UHPC, the State should work with the concrete manufacturer to determine if there are any steel fiber suppliers who meet the material specifications. The steel fiber must be certified by the steel manufacturer that it complies with the Buy America requirements.

If a State determines that domestic steel fibers are not available, the State may determine if the quantities involved meet the minimal use provision which allows use of foreign steel if the total cost of all non-domestic steel used on a project falls below the threshold set within the regulations. If the quantities do not meet the minimal use provision, the State must submit a waiver request to FHWA's division office.

Material Properties

This section summarizes state-of-the-art research regarding the material behavior of UHPC under various loading conditions. In addition, recommendations for structural properties of UHPC suitable for usage in design practice are presented.

Compressive Strength

Compressive strength of the material is one of the critical parameters needed for design of concrete structures. The characteristic compressive strength of UHPC is defined in a similar fashion as defined for NSC, by its 28-day strength. The applicability of the standard ASTM compression testing method to evaluate the compressive behavior of UHPC has been studied by several researchers.^(12,13) Based on this research, it was found that a standard ASTM test using 3-inch by 6-inch cylinders with appropriate surface preparation and at a load rate of 150 psi per second can be utilized for compression testing to determine the characteristic compressive strength of UHPC.⁽¹²⁾

UHPC has very high compressive strength compared to NSC or typical HPC used in current-day practice due to the high-density matrix and the absence of coarse aggregate. Based on the current research on UHPC, the characteristic 28-day compressive strength (fc' ) of UHPC ranges from 18 to 33 ksi, depending on the type of curing process. (See references 12, 13, 14, and 15.)

A steam-cure treatment for 48 hours at a temperature of 194 °F is typically used to achieve the full compressive strength, especially in precast environments. An FHWA study completed on more than 1,000 compression test samples showed that the average compression strength of UHPC from Lafarge North America is 18.3, 28, 24.8, and 24.8 ksi for air, steam, delayed steam, and tempered steam curing conditions, respectively.⁽¹²⁾

A recent study conducted by the Michigan DOT reported compressive strengths of 23.9 ksi and 30.5 ksi for air-cured and steam-cured specimens, which are higher compared to previous studies.⁽¹³⁾ In addition, it was found that under steam-curing conditions, the characteristic strength of UHPC was reached after 3 days of casting (which included 2 days of steam curing). For air-cured specimens, the UHPC strength gradually increased with time, reaching strengths of 14 ksi and 19 ksi after 3 days and 7 days, respectively.

Furthermore, a recent FHWA study investigating the effects of curing temperature on compression behavior of non-steam-cured UHPC suitable for field-cast applications found that UHPC achieved a compressive strength of 22.5 ksi to 24.5 ksi at 28 days when curing temperatures of 50 °F, 73 °F, and 105 °F were used.⁽¹⁶⁾

Based on the summary of the above-mentioned studies and the compressive strength reported in real-world UHPC bridge projects (see references 7, 8, 9, 10, and 17), the following recommendations related to the compressive strength of UHPC are proposed:

Recommendations

• The compressive strength of UHPC depends on the curing conditions (air-cured, steam-cured, delayed steam-cured, etc.)
  
  
• The steam-cure treatment for 48 hours at 194 °F and 95 percent humidity accelerates the strength gain of UHPC. Hence, steam curing of UHPC products is encouraged when they are produced in precast plants.
  
  
• The characteristic compressive strength of UHPC shall be taken conservatively as 24 ksi at the end of curing for steam-cured conditions and 18 ksi at the age of 28 days for air-cured conditions.
  
  
• For steam-cured or heat-treated bridge elements, the compressive strength of UHPC can be taken to be equal to fc' after 3 days from curing for capacity estimations. Therefore, these elements can be used without the 28-day waiting period, as is the case with conventional concretes.
  
  

Compressive Stress-Strain Behavior

The stress-strain behavior of the material is essential to estimate the neutral axis depth, reinforcement strains, and capacity at service and ultimate limit states. The compressive stress-strain behavior of UHPC was established in numerous concrete compressive cylinder tests conducted by several researchers and is shown in figure 1a. (See references 12, 13, 16, 18, and 19.)

The FHWA study notes that the strain value corresponding to the peak compressive strength is about 0.0035 and 0.0041 for air- and steam-cured specimens, respectively. ^(12,16) In addition, unlike NSC, the measured stress-strain relationship for UHPC was found to be linear up to 80 to 90 percent of the peak stress for both curing conditions (see point A in figure 1a). However, according to Sritharan et al., heat-treated UHPC exhibited linear elastic behavior up to failure, corresponding to a compressive strain of 0.0032.⁽¹⁸⁾

The post-peak behavior of the UHPC is dependent on the volume percent of fibers in the mix and orientation of the fibers. Hence, for design purposes, the stress and strain values can be obtained by using appropriate safety factors on the observed values. Consistent with the French, Australian, and Japanese design recommendations for UHPC, the stress-strain curve shown in figure 1b is recommended for design.^(20,21,22) The critical values in this figure are also consistent with some of the design guidelines proposed by Graybeal.⁽²³⁾

Recommendations

• The stress-strain behavior of UHPC in compression shall be taken as a trilinear curve, as shown in figure 1b.
• In the absence of test data, for design purposes, the maximum compressive strain should be limited to 0.0032.

Tension Stress-Strain Behavior

The use of steel fibers results in dependable tension capacity for UHPC. Consequently, the tension capacity of UHPC can be utilized in design at ultimate limit state. This is in contrast to the design of members using normal concrete, where the concrete tension capacity is ignored after cracking. The tensile strength and post-cracking behavior of UHPC depends on the strength, quantity (e.g., volume by percentage), length, and orientation of steel fibers, which effectively prevent or delay opening of concentrated cracks. The tensile strength is also influenced by the type of curing treatment (steam- or air-cured) provided for UHPC members.

An investigation conducted by the FHWA examined four different methods to evaluate the tensile behavior of UHPC including the flexural prism test, split cylinder test, Mortar Briquette ("dog-bone") test, and direct tension test.⁽¹²⁾ Although all four of the test methods provided realistic tensile cracking strengths, the results varied by 0.5 ksi depending on the test method. Therefore, the study conservatively recommended the cracking tensile strength of UHPC to be taken as 1.3 ksi and 0.9 ksi for steam-cured and untreated (i.e., air-cured) conditions, respectively.

As a follow-up to a series of compression and flexural tests, a set of direct tension tests were conducted on large steam-cured dog-bone specimens, which produced tension parameters comparable to those resulting from the FHWA study.⁽¹⁸⁾ The tensile stress-strain behavior established from these dog-bone tests, which have been used successfully in characterizing the flexural response of a UHPC full-scale bridge girder, tapered H-shaped piles, and waffle deck panels, is shown in figure 2a.

Based on the back analysis of large-scale UHPC I-girder tests under flexure and shear, Graybeal proposed a conservative approximation for the UHPC tensile stress-strain behavior for estimating the ultimate capacity of the UHPC sections.⁽²³⁾ Accordingly, UHPC under tension can be assumed to behave in an elastic-perfectly plastic fashion with a post-cracking capacity of 1.5 ksi for strains below the pullout strain of 0.007 (see figure 2a)

In a recent study on tension behavior of UHPC, Graybeal and Baby used a direct tension test method using dog-bone shaped test specimens of different sizes.⁽²⁴⁾ These tests characterized the tension behavior of the UHPC for different curing conditions and steel fiber quantities. This study noted that the tensile response of UHPC consists of the following four phases: elastic behavior, inelastic cracking, straining in discrete cracks, and single crack localization. Results from these tests also confirmed that the tensile response of UHPC can be represented with an elastic-perfectly plastic response for design purposes as suggested in figure 2b.

When comparing the different studies, variations in the post-cracking behavior of the UHPC can be noted, which is influenced by the dispersion and orientation of the fibers in the test specimens and, most importantly, by the location and length of gauges and how the strain was characterized if measured over a long gauge length. In consideration of the reported experimental responses and recommendations including those adopted by others, the tension model as shown in figure 2b is recommended for use in the design of precast waffle deck systems. The recommended model is motivated by maintaining simplicity and achieving satisfactory characterization for the waffle deck panels at the serviceability and ultimate limit states.

Recommendation

• The stress-strain behavior of UHPC in tension shall be taken as a bilinear curve as shown in figure 2b.

Modulus of Elasticity

UHPC displays linear elastic behavior in both compression and tension up to certain strain limits, as was shown in figure 1. Tests by Bonneau et al. showed the elastic modulus of UHPC without fibers is 6,700 ksi compared to 7,100 ksi with a 2.0 percent steel fiber content, an increase of only about 6.5 percent due to the presence of fibers.⁽²⁵⁾ According to Graybeal, standard heat treatment increases the elastic modulus of UHPC by 23 percent from 6,200 ksi to 7,650 ksi.⁽¹²⁾

Many equations have been used to define the relationship between the elastic modulus and the compressive strength of concrete. The following equations have been developed specifically for UHPC based on the available experimental data:^(17,18,20)

The comparison of these equations with the experimental data is shown in figure 3. It is clear from the figure that the equation proposed by Graybeal predicts the Young's modulus more accurately than other equations.⁽¹²⁾

Recommendation

• The elastic modulus of UHPC can be obtained using . In the absence of exact concrete strength, a value of 7,500 ksi can be used for design purposes.

Density

The density of UHPC is slightly higher than that of HPC or NSC due to its very compact microstructure. The average reported value for the density of UHPC mixes from 17 published mix descriptions was approximately 157 lb/ft³.⁽²⁶⁾ A unit density of 155 lb/ft³ was suggested by studies done in the US.^(12,13)

Recommendation

• The unit weight of the UHPC shall be assumed to be 157 lb/ft³ for dead load estimations.

Coefficient of Thermal Expansion

UHPC tends to exhibit a higher coefficient of thermal expansion (CTE) than NSC. This may be attributed to the fact that UHPC contains a comparatively high volume of cementitious materials (relatively high CTE) without any coarse aggregate (low CTE values). The FHWA and Michigan DOT studies recommended a value of about 8.2 x 10-6/°F for CTE.^(12,13) The Michigan study found that this value can be used regardless of the age of the concrete once thermal treatment has been completed. For comparative purposes, the expected CTE for NSC is about 6.0 x 10-6/°F and for HPC is between 4.0 - 7.3 10-6/°F. ^(27,28)

Recommendation

• For thermally treated UHPC, a CTE value of 8.2 x 10-6/°F is recommended for design.

Chloride Penetration Resistance

Concrete degradation in bridge decks is accelerated by the penetration of chloride ions from deicing agents into the concrete from the top, leading to the corrosion of the reinforcement. Therefore, chloride ion penetration resistance is one of the critical parameters that dictate the durability of a system and concrete cover requirement for reinforcement. UHPC exhibits very low to negligible permeability when compared to NSC due to its high-density matrix and low water/cement ratio. The ASTM C1202 standard, commonly known as the rapid chloride ion penetrability test, can be used to estimate the chloride resistance of UHPC.

In the study by Graybeal, the chloride penetration resistance of specimens receiving any of the four curing regimes at 56 days and those receiving any form of heat treatment at 28 days achieved a Negligible rating (< 100 coulombs).⁽¹²⁾ Only untreated specimens at 28 days did not receive this rating; those specimens averaged passing 360 coulombs of charge, resulting in a Very Low permeability qualification. Similarly, Michigan DOT reported negligible chloride penetrability for all specimens tested in their study.⁽¹³⁾

Graybeal used another test procedure, known as the chloride ponding test, to determine the level of migration of chloride ions into the UHPC over 90 days. ⁽¹²⁾ According to these findings, the chloride ion content was extremely low for all curing regimes. The average chloride content for different curing regimes was less than 0.00312 lb/ft³ and, in most cases, the average was 0.00125 lb/ft³. All of these values are below the minimum accuracy threshold for the test method, indicating that the volume of chlorides that penetrated into the UHPC is extremely low.

A summary of average values of various durability parameters for UHPC, HPC, and NSC is presented in table 2. In addition, the durability properties of UHPC, HPC, and NSC are compared in graphical form in figure 4.

Table 2. Durability properties of UHPC compared to HPC and NSC.

Parameter	UHPC	HPC		Normal Concrete
		Value	Ratio to UHPC	Value	Ratio to UHPC
Salt Scaling Mass Lost (28 cycles)	0.010 lb/ft2	0.031 lb/ft2	3.0	0.31 lb/ft2	30
Chloride Ion Diffusion Coefficient	2.2×10-13 ft2/s	6.5×10-12 ft2/s	30	1.2×10-11 ft2/s	55
Chloride Ion Penetration Depth	0.04 in.	0.32 in.	8	0.91 in.	23
Chloride Ion Permeability Total Charge Passed	10 – 25 coulombs	200 – 1000 coulombs	34	1800 – 6000 coulombs	220
Carbonation Depth (3 years)	0.059 in.	0.16 in.	2.7	0.28 in.	4.7
Reinforcement Corrosion Rate	4×10-7 in./yr	9.8×10-6 in./yr	25	4.7×10-5 in./yr	120
Abrasion Resistance Relative Vol. Loss Index	1.1 – 1.7	2.8	2.0	4.0	2.9
Resistivity	53.9 kW·in.	37.8 kW·in.	0.70	6.3 kW·in.	0.12

Recommendations

• For precast UHPC deck panels, the chloride penetration resistance can be assumed to be 100 coulombs (negligible).
• For the deck panel connections, which are not typically heat-treated, the chloride penetration resistance can be assumed to be 400 coulombs (very low).

Concrete Cover and Spacing

Concrete cover is the distance from the surface of the concrete to the surface of the reinforcing bars and strands embedded in the concrete. Ensuring sufficient concrete cover is critical not only for the durability of concrete structures but also for the development of the bond strength along the primary reinforcement. Concrete cover requirements depend on environmental conditions and the chloride penetration resistance of the concrete used for design. Recognizing UHPC's high resistance to chloride penetration, several international codes for UHPC, Gowripalan and Gilbert recommended a decreased concrete cover.⁽²¹⁾ The Japan Society of Civil Engineers (JSCE) and the Australian code for UHPC require a minimum of 0.75 inches of concrete cover for uncoated reinforcement.^(21,22)

Research by Tuchlinski et al. suggested that the UHPC cover for prestressing strands designed to develop the full strength of the strand may be reduced compared to normal concrete.⁽³⁵⁾ Based on their research, Tuchlinski et al. recommended a concrete cover of 1.5 times the strand diameter for UHPC to ensure that the full strength of the strand can be developed. In addition, a clear cover of 0.84 inches was used successfully on 0.6-inch-diameter strands in full-scale UHPC bridge girder tests at Iowa State University (ISU).⁽⁹⁾

The Tuchlinski et al. research also recommended center-to-center spacing for prestressing strands in UHPC to be at least 3.0 times the nominal strand diameter.⁽³⁵⁾ For research completed at ISU, a clear spacing of 3.3 to 4 times the strand diameter has been used successfully.^(9,26) Based on experience, a local precaster recommended that a minimum spacing of at least 1.5 inches should be used to ensure flow of UHPC freely through the section during casting.

Recommendations

• The minimum concrete cover for unprotected mild reinforcement in UHPC shall be 0.75 inches.
• The minimum cover concrete for prestressing strands shall be 1.5 times the strand diameter.
• The reinforcement including strands shall be designed with a minimum clear spacing equal to 3.25 times the diameter of the reinforcement or 1.5 inches, whichever is greater.
• Unlike the current American Association of State Highway and Transportation Officials (AASHTO) requirements, for decks exposed to tire studs or chain wear, there is no need to increase the cover requirement for UHPC because the abrasion resistance of the UHPC is significantly higher than that of NSC.

Bond Strength

The bond between UHPC and reinforcement (rebar or strand) is important in determining development and transfer lengths. Given that UHPC has such drastically different mechanical properties from conventional concrete, the use of the current AASHTO guidelines is conservative.

In a recent study by the FHWA, the development length of mild steel reinforcement embedded into UHPC was investigated through a series of tensile pullout tests on #4, #5, and #6 bars cast into 16-inch-diameter UHPC cylinders.⁽³⁶⁾ It was found that embedment lengths of 2.9 inches (= 5.8d_{b #4}), 3.9 inches (= 6.24d^b#5), and 4.9 inches (= 6.53d_b#6) were sufficient to fracture the #4, #5, and #6 bars, respectively. Thus, to develop the yielding of the mild steel reinforcement (f_y = 60,000 psi), the embedded length required is about two-thirds of the measured values.

With the limited data available, the anchorage length required for mild steel reinforcement to develop its yield strength in UHPC can be assumed to be 6db (db = diameter of reinforcement) at the minimum. While further research is needed to fully characterize anchorage behavior of reinforcement in UHPC and take into account the effect of fatigue loading on the bond development length, the completed system test suggests that meeting the recommended minimum anchorage length is expected to provide sufficient anchorage of reinforcement in waffle deck systems.

Recommendations

• The development length required for mild steel reinforcement in waffle deck systems to develop its yield strength in UHPC shall be no less than 6d_b, where d_b is the diameter of the reinforcement in inches.
• The development length required for mild steel reinforcement in waffle deck systems to develop its ultimate strength in UHPC shall be no less than 9d_b

Shrinkage Behavior

UHPC exhibits relatively large shrinkage behavior in comparison to normal concrete due to high cementitious material content; it is therefore more susceptible to cracking under restrained conditions. A study by Graybeal found that UHPC exhibited rapidly occurring, large-value, early-age shrinkage strains.⁽¹²⁾ The shrinkage strains in heat-treated UHPC were nearly 850 microstrains during the curing period. Untreated UHPC also exhibited shrinkage strains beyond 790 microstrains.

Although the shrinkage strain in UHPC is higher than in NSC, shrinkage in UHPC takes places at an early age. In fact, heat-treated UHPC does not exhibit any shrinkage in the post-treatment period. In the absence of treatment, specimens reached 95 percent of ultimate shrinkage at the age of 2 months.⁽¹²⁾

UHPC will crack in tension at strains lower than these shrinkages (see tensile behavior), and it is therefore important to mitigate or eliminate shrinkage restraints in UHPC structural members during casting.

Recommendations

• For heat-treated UHPC, the shrinkage strain can be taken as 850 microstrains during the curing period and zero microstrains after the heat treatment.
• For untreated UHPC, shrinkage strains up to 850 microstrains will develop over a period of 60 days.

Poisson's Ratio

In the recent study by Michigan DOT, the Poisson's ratio for UHPC under all curing conditions and ages was found to be consistently between 0.20 and 0.21, which is in line with typical concrete and the value given by the Association Française de Génie Civil.^(13,20) Although a value as low as 0.13 was reported in the literature, the higher values have been reported consistently in other research as well.^(26,37)

Recommendation

• The Poisson's ratio for UHPC material shall be taken as 0.20.

System Configuration and Benefits

Similar to the typical full-depth precast deck systems used in current practice and developed in previous research, the waffle deck system consists of a series of UHPC waffle deck panels that are full-depth in thickness (as required by the structural design) and connected to the supporting girders with robust and efficient connections. More details about the connections are presented in chapter 3. The UHPC waffle deck panel system consists of the following (see figure 5):

• Precast UHPC waffle panels with shear pockets to accommodate the shear connections between the waffle panels and supporting girders.
• Transverse panel-to-panel connections that connect two adjacent waffle deck panels forming a joint parallel to the bridge transverse axis.
• Longitudinal panel-to-girder connections that connect the waffle deck panel and the center girder of a specific bridge (joint will be present when the deck panel lengths are less than the full width of the bridge and the joint will run parallel to the bridge longitudinal axis).
• Some type of overlay (e.g., asphalt) may be used to improve the pavement rideability as needed by the specific requirements from the owner.
• In situ UHPC material used to fill the panel-to-panel connections and shear pockets in the field.

A UHPC waffle deck panel consists of a slab cast integrally with concrete ribs spanning in transverse and longitudinal directions. This system is similar to the two-way joist system used in the building industry.

A transverse strip along the deck panel acts a T-beam, distributing wheel load effects to the adjacent bridge girders. The longitudinal ribs help in distributing the wheel load to the adjacent panels through the panel-to-panel connections. The reinforcement needed to resist the wheel loads is provided in the ribs along both directions. The spacing of the ribs in both directions is determined based on the girder-to-girder spacing, panel dimensions, and minimum detailing requirements for panel-to-panel connections. Additional details about the spacing and reinforcement of the panel are presented in chapter 2.

Due to the excellent structural properties of UHPC, the waffle deck system for a given thickness has the same or higher capacity and is 30 to 40 percent lighter than a comparable solid precast full-depth panel made of NSC. The decreased weight of the UHPC panel has significant benefits. For example, it can increase the span length and girder spacing, improve bridge ratings when used for deck replacement projects, and reduce seismic, substructure, and foundation loads when compared to solid precast deck panel systems. The presence of the steel fibers in UHPC and very minimal shrinkage of UHPC after steam curing of the precast elements decreases the reinforcement requirements.

The panels can be designed to act as composite or non-composite with the supporting girders. However, to improve structural efficiency and cost-effectiveness, it is recommended to design the UHPC waffle deck system to act compositely with the girders for resisting design loads. The length and width of the deck panels are typically determined by specific bridge geometry, type of bridge (concrete or steel girder), and transportation/handling limitations. Based on the current practice, the width of the panels (length in the direction of traffic or short dimension of the panel) may be varied between 8 and 12 ft.

Experimental and Field Testing Waffle Deck System

This section presents a summary of the laboratory testing done to evaluate the structural performance of the waffle deck panel and connections designed for a 60-ft-long replacement bridge in Wapello County, Iowa. With experimental verification of the structural behavior in the laboratory completed, the UHPC waffle deck system was implemented in the replacement bridge project for the first time in the US. The observations from the field testing of the bridge are also summarized in this section.

Laboratory Testing

Based on the previous research completed at the FHWA on structural behavior of a prestressed UHPC waffle deck system, a UHPC precast waffle deck system with conventional mild steel reinforcement was developed by the Iowa DOT and ISU for ABC purposes. ⁽¹⁷⁾ The UHPC waffle deck panel was designed for a two-lane, single-span replacement bridge in Wapello County, Iowa. Figure 6 shows the cross-section details for the bridge. A full set of design drawings is provided in the project final report. ⁽³⁸⁾

The waffle deck panel was 8 inches thick and was designed to resist current AASHTO load requirements. This resulted in Grade 60 #6 (d^b = 0.75-in., where d^b is diameter of the bar) and #7 (d^b = 0.875-in.) mild steel reinforcement as top and bottom reinforcement, respectively. All of the reinforcement was provided along panel ribs in both directions. The plan view of the waffle deck panel showing the reinforcement and rib spacing is shown in figure 7a. More detailed information regarding the cross-section and reinforcement locations is presented in chapter 2 and in Aaleti et al.⁽³⁹⁾

To make the UHPC waffle deck panels fully composite with the prestressed concrete girders, three different connections were utilized: shear pocket connection, longitudinal connection, and transverse connection. The shear pocket connection consists of a shear hook from the girder extended into a shear pocket in the waffle deck panel, with the shear pocket filled with in situ UHPC (figure 7b).

The longitudinal connection between the waffle panel and girder was formed by tying dowel bars from the panels with shear hooks from the girder, using additional longitudinal reinforcement, and then filling with in situ UHPC (figure 7c).The transverse connection between the UHPC waffle deck panels contained dowel bars from the panels tied to additional transverse reinforcement, with the gap between the panels filled with in situ UHPC (figure 7d).

Test Setup and Instrumentation

For the experimental investigation, a waffle deck region between two adjacent girders, as identified in figure 6, was chosen. Two waffle deck panels (UWP1 and UWP2), 8 ft long by 9 ft, 9 in. wide, were fabricated using a commercially available, standard UHPC mix from Lafarge North America. The waffle deck panels were cast upside-down for ease of construction. The details about the construction of waffle panels are presented in Aaleti et al. ₍₄₀₎ The setup used for the UHPC waffle deck system test was designed to replicate the critical regions of the field structure closely and is shown in figure 8.

The UHPC deck panels were supported on two 24-ft-long prestressed concrete girders, which were 7 ft, 4 in. apart and simply supported at the ends on concrete foundation blocks. The connections between the two deck panels, as well as those between the panels and the girders, were then cast using UHPC mixed in the laboratory at ISU.

Several string potentiometers and strain gauges were used to monitor the performance of the waffle deck system during testing. The instrumentation details are shown in figure 9. A 10-inch by 20-inch steel plate attached at the loading end of a ±55 kip fatigue hydraulic actuator was used to simulate a truck wheel load on the panel.

Summary of Experimental Results

The performance of the UHPC waffle deck system, including the UHPC in situ connections between precast elements, was examined using nine different tests and a single-wheel truck load simulated using a hydraulic actuator. The following two different locations were chosen to apply the load along the centerline between the two girders: center of the deck panel and center of the panel-to-panel transverse joint. The details of the completed load tests are summarized in table 3.

Table 3. Sequence and details of the tests conducted on the waffle deck system.

Test	Test Description	Location	Maximum Load
1	Service load test panel-2 (UWP2)	Center of the panel	1.33a x 16 k = 21.3 k
2	Service load test on transverse joint	Center of the joint	1.75b x 16 k = 28 k
3	Fatigue test on the transverse joint	Center of the joint	28 k (1 mil cycles)
4	Overload test of transverse joint	Center of the joint	48 kips
5	Fatigue test on the panel-1 (UWP1)	Center of the panel	21.3 k (1mil cycles)
6	Overload test of the panel	Center of the panel	40 kips
7	Ultimate load test on panel UWP1	Center of the panel	160 kips
8	Ultimate load test on the transverse joint	Center of the joint	155 kips
9	Punching shear failure test on UWP1	Between transverse ribs	155 kips

This section focuses only on the results from service load tests and ultimate load tests, while the results and observations from the overload tests and fatigue tests are presented in Aaleti et al.⁽⁴⁰⁾

Panel and Joint Service Load Tests

In the panel service load test, a maximum load of 21.3 kips, representing the AASHTO truck service load plus 33 percent impact, was applied at the center of panel UWP2 (at rib TR2 and between the girders, see figure 9). In the joint service load test, a maximum load of 28 kips, approximately representing the AASHTO truck service load plus 75 percent impact, was applied at the center of the transverse joint. The load-deflection curves for both cases are shown in figure 10. In both cases, a linear relationship was observed between the load and deflection. Maximum deflections of 0.034 inches and 0.022 inches were measured during the panel and joint service load tests, respectively.

The peak strain recorded in the bottom reinforcement of the center rib, running in the transverse direction, during the panel service load test was only 375 microstrains, or 18 percent of the yield strain. The strain variation along the length of the bottom reinforcement, in the transverse rib TR2 of panel UWP2, and the variation of normalized bottom reinforcement strains at the center of the transverse ribs at the peak load are shown in figure 11a.

Figure 11. Graphs. Measured strain distribution along the transverse rib in the center of the panel and normalized strains at the center of the transverse ribs along the longitudinal direction under service load conditions.

Figure 11b illustrates that, for an applied load P at the center of the panel, the transverse rib TR2 provides 70 percent of the resistance. The adjacent ribs on either side of TR2 (i.e., TR1 and TR0, and TR3 and TR4) provide 10 percent and 5 percent of total resistance, respectively.

Panel Ultimate Load Test

The ultimate load test was carried out to investigate the adequacy of the precast deck system and its connections under ultimate load conditions. The ultimate load referred to in this study was arrived at based on the recommendations from Iowa DOT personnel. A total load of 160 kips, equivalent to 10 times the AASHTO truck service load, was applied at the center of panel UWP1. The load-deflection curve established at the center of this panel during testing is shown in figure 12.

The panel exhibited a linear force-displacement behavior response up to 80 kips. A maximum deflection of 0.82 inches was measured at the center of panel UWP1 (i.e., at the center of transverse rib TR2). The peak strain measured in the bottom reinforcement of transverse rib TR2 was about 1600 microstrain, which was about 76 percent of the yield strain of the reinforcement. A significant amount of cracking was observed on all of the transverse ribs (TR1, TR2, and TR3) and both of the longitudinal ribs (LR1 and LR2) of panel UWP1. The maximum crack width measured along the transverse rib TR2 in UWP1 was 0.08 inches.

Joint Ultimate Load Test

A total load of 155 kips, equivalent to about 10 times the AASHTO truck service load, was applied at the center of the transverse joint. The load-deflection curve established at the center of the panel-to-panel joint is shown in figure 13a. The peak strain measured in the bottom reinforcement of transverse rib TR2 was about 1475 microstrain, or about 70 percent of the yield strain of the reinforcement. At the end of the test, a large number of cracks were formed in transverse ribs of the joint (see figure 13b). The maximum load applied was controlled by the shear cracking initiation in the prestressed girders.

Punching Shear Failure Test

In this test, a wheel load was applied at the center of the waffle deck cell bounded by transverse and longitudinal ribs TR2, TR3, LR1, and LR2. Load was applied at increments of 5 kips on the waffle deck panel using a 200-kip actuator. The 10-inch by 20-inch plate at the loading end of the actuator was replaced with a 6-inch by 8-inch steel plate to force punching shear failure in a region bounded by the ribs in the panel.

As the loading increased, a large number of radial cracks in the top surface and flexural cracks in the ribs were formed. The measured load-displacement response at the center of the cell is shown in figure 14a. The crack pattern on the bottom surface of the waffle deck was as expected for a typical punching shear failure and is shown in figure 14b. The waffle deck failed suddenly at a maximum load of 154.6 kips, leaving a 6-inch by 8-inch hole (the same size as the steel plate placed at the top of the deck) on the top surface at the center of cell. The punching shear failure surface had edges sloped at approximately 45 degrees, as shown in figure 14c. The measured average punching shear strength was about 1.068 ksi, which is equivalent to . The measured punching shear failure capacity is nearly 2.3 times the estimated value using the equation recommended by Harris and Roberts-Wollmann.⁽⁴¹⁾

Field Testing

The constructability of the UHPC waffle deck system and structural performance of its critical connections and panels were investigated under service, fatigue, and ultimate loads using large-scale laboratory tests as described in the previous section. The results and observations from the laboratory tests were used to design the demonstration bridge on Dahlonega Road in Wapello County, Iowa.

This bridge replacement project was used to demonstrate the deployment of the UHPC waffle deck panel technology from fabrication through construction and to evaluate the performance of the UHPC waffle slab deck under true service conditions. The Dahlonega Road Bridge over Little Cedar Creek was opened to traffic in November 2011 and was field tested in February 2012.

For the field evaluation of the bridge system, live load was applied by driving a heavily loaded, standard dump truck across the bridge along predetermined paths. The total weight of the truck was 60,200 lb, with a front axle weight of 18,150 lb and two rear axles weighing roughly 21,000 lb each. The truck configuration with axle loads is shown figure 15.

Seven load paths were used to evaluate the field performance of the waffle deck panel system and its connections. The outer edge of the loading truck was located 2 ft from each barrier rail in load paths 1 and 7. In load paths 2 and 6, the truck edge was along the centerline of each respective traffic lane, causing maximum movement in the deck panel. Load paths 4 and 5 were 2 ft to either side of the bridge centerline for the outer edge of the truck, and load path 3 straddled the centerline of the bridge. A schematic of the load paths is shown in figure 16.

For static tests, the truck was driven across the bridge at very slow speed (< 5 mph). Each load path was traversed twice to ensure repeatability of the measured bridge response. For dynamic tests, the truck speed was increased to 30 mph to examine dynamic amplification effects.

Taking advantage of the bridge symmetry about its longitudinal and transverse centerlines, two UHPC waffle deck panels along the length of the bridge were selected for instrumentation. One of these panels was located near the mid-span, and the other was located adjacent to the south abutment.

Surface-mounted strain gages were used on each panel and their adjacent UHPC deck joints to quantify deformations and identify the likelihood of cracking under service loads. The locations of these strain gages were selected carefully to coincide with critical locations on the panels and deck joints where stress and strain would likely be extreme.

In addition to the strain gages on the deck panels, 13 surface-mounted strain gages and 5 string potentiometers were attached to the girders to characterize the global bridge behavior, measure mid-span deflections, and quantify lateral live load distribution factors. Using two additional string potentiometers, deflections were also measured at the mid-spans of the deck panel located near the center of the bridge. Top and bottom girder strains were monitored for three of the girders at mid-span and at a section 2 ft from the southerly abutment. More details about the instrumentation can be found in Rouse et al.⁽⁴²⁾

The maximum strain in the UHPC waffle deck panel at the mid-span recorded during the static load testing was 95 microstrains. This value is well below the cracking strain for the UHPC. This behavior implies that there is no cracking in this deck panel, and it is responding elastically to the applied truck load. However, the maximum strain measured in the waffle deck panel adjacent to the abutment was 276.2 microstrains, which is slightly more than the cracking strain of UHPC. The maximum deck deflection recorded for the static loading for different load paths was 0.038 in., which is well below the allowable AASHTO deflection requirements. Overall, the bridge system performed as expected, confirming the excellent behavior of the waffle deck system.