Concrete Material Properties

The research utilized two distinct 30 MPa concrete mixes: a low shrinkage concrete (LSC) and a standard normal weight concrete. The LSC was employed specifically for specimen G1 during Phase 1 of the short-term push-out tests, while the normal weight concrete was used for all other specimens in that phase. The precise mix designs for both concrete types are detailed in Table A1-2 (this table is not provided in this excerpt but would be found in the complete document). This initial selection of concrete types already highlights a key variable influencing the subsequent structural tests. The choice to use both types of concrete demonstrates a deliberate effort to investigate how differing shrinkage characteristics would affect the overall structural performance of the timber-concrete composite elements under load. The subsequent analysis across various phases of testing would then seek to determine the extent of the influence of the chosen concrete mix on the behavior of the composite material.

A comprehensive suite of tests was performed to characterize the mechanical properties of the concrete mixes. Table A1-4 summarizes the results of tests conducted on the concrete used throughout Phases 1, 2, and 3, encompassing both short-term and long-term push-out tests. Similarly, Tables A1-5 and A1-6 provide corresponding summaries for Phase 4 (short-term beam tests) and Phase 5 (long-term beam tests) respectively. These tables (again not included here but part of the full document) are crucial in establishing a baseline understanding of the concrete behavior, which is a fundamental component of the broader study on the overall timber-concrete composite performance. Noteworthy is the calculation of the concrete's elastic modulus using Equation A1-1 from Clause 5.2.3 of NZS 3101: Part 1 (SNZ, 2006). This calculation specifically considered the varying densities across different concrete mixes, emphasizing the importance of accounting for these variations when determining the overall material response. The use of established standards like NZS 3101 underscores the rigorous approach to the study’s methodology.

The assessment of concrete properties relied on three primary testing methods. First, cylinder compression tests (using 100 × 200 mm cylinders) were performed according to NZS 3112: Part 2 (SNZ, 1986a), with a minimum of three cylinders per test at 7, 28 days, and on the day of the push-out tests. These tests directly evaluated concrete compressive strength and other related material properties. Secondly, slump tests were carried out following NZS 3112: Part 1 (SNZ, 1986b), providing a measure of concrete workability and consistency. This aspect is significant as it provides information on the ease with which the concrete could be handled and placed during construction and directly influences the quality of the final product. Finally, drying shrinkage tests were performed using prisms, as specified in NZS 3112: Part 3 (SNZ, 1986c), providing data on the dimensional changes in the concrete over time. This data is vital because it shows how shrinkage will affect the structural integrity of the complete assembly over time. The selection of these three methods demonstrates a comprehensive analysis aiming to account for the variability and complex behaviors of concrete in a real-world structural application.

The LVL (Laminated Veneer Lumber) joists in this study were subjected to combined bending and tensile stresses. The inherent difference in timber strength distribution under these two loading types was explicitly acknowledged. Determining the LVL's mean strength required careful consideration of these distinct loading behaviors. The data used to determine LVL strength came from Carter Holt Harvey (CHH) factory production tests. These tests, conducted in Auckland, New Zealand, focused on 95d × 63w mm LVL cross-sections taken from entire billets. Since the actual LVL dimensions used in the composite beam specimens were 400d × 63w mm, a size effect correction was applied to the factory data using a strength reduction factor developed by CHH for use in the USA. This correction accounts for the known influence of the size of the timber member on its load-bearing capability. The methodology clearly demonstrates a commitment to using accurate and relevant data for a reliable strength assessment of the LVL joists within the constructed specimens.

Data for tensile strength was also sourced from CHH factory production tests, this time from their US operations. A significant aspect of the tensile strength determination was the dependence on the length of the tested specimens, with strength decreasing as the length increased. A specific relationship was proposed to account for this length effect, involving a strength reduction factor applied to the tensile strength values. This highlights a common challenge in material characterization where variations in geometric properties influence the resulting strength. The detailed procedures and calculations involved are likely found in more detailed supplementary data, as only the resulting strength is explicitly mentioned here. It's crucial to note that the tensile stress was found to be consistent across different specimens. As a result, a mean value of f(M/N) = 39.4 MPa was adopted for all beam specimens. This average value simplifies the analysis while acknowledging the relatively consistent tensile behavior observed. The use of CHH data demonstrates the practical approach of leveraging readily available, industry-standard test results to inform the study. It reinforces the relevance of the research findings to actual construction scenarios.

The final step in the LVL strength determination was calculating the mean combined bending and tensile strength. Based on the consistent tensile stress observed across specimens, an average value of f(M/N) = 39.4 MPa was utilized. Consequently, a mean combined bending and tension strength, f_mean, of 39.4 MPa was assumed for all LVL components tested within the composite beam specimens. This consolidation of bending and tensile strength into a single mean value allowed for a simplified yet robust representation of the LVL's overall load-bearing capacity within the composite system. This approach effectively integrates the findings from both bending and tensile strength analyses to generate a practically applicable single metric for LVL strength. The significance of this single value lies in its use as a key parameter in the subsequent structural analyses of the timber-concrete composite members. The 39.4 MPa value serves as a fundamental input for several calculations and modeling aspects detailed in later sections of the document.

Short-term symmetrical push-out tests formed a significant part of the research, conducted in two phases between late 2006 and 2008 at the University of Canterbury, New Zealand. The first phase involved 30 specimens representing 15 different connection types (A1 to C2, and D1 to H4), with two specimens per type. The second phase significantly expanded the testing, utilizing 36 specimens: 30 for short-term tests and 6 for a concurrent long-term test. This phase focused on three connection types (triangular notched coach screw – T; 300 mm rectangular notched coach screw – R; toothed metal plate – P), each with nine specimens, plus three additional triangular notched specimens tested in the weak direction (TT). Each phase of testing took approximately two months to complete, highlighting the significant time investment involved in the fabrication and testing of these specimens. The two-phase approach allowed for an iterative refinement of the testing methodology and a focused investigation into a subset of connection types which proved particularly insightful during the first phase of testing.

Constructing the push-out test specimens was a labor-intensive process. The creation of each specimen involved several steps: formwork preparation, LVL cutting and notching, coach screw insertion, slab formwork assembly, reinforcement work, and concrete casting. The construction process for each phase took roughly two months, with two skilled workers dedicating 8 hours a day, 5 days a week to the task. This translates to a total of 640 man-hours invested in constructing just 30 symmetrical push-out test specimens. The manual nature of much of this work, coupled with the use of only small power tools, underscores the complexity and resource commitment associated with each specimen. The high labor investment emphasizes the significant effort required for this experimental approach and indicates a high level of detail and precision applied throughout. The lengthy fabrication period also suggests that the careful preparation and execution of each step were crucial for the reliable generation of meaningful experimental data.

The push-out test specimens were carefully designed to represent practical connections in timber-concrete composite structures. For example, the formwork for the concrete slab utilized 45b × 90d mm sawn timbers for edges and 17 mm thick plywood for the base, interconnected by splices and screws. Notches were cut using a bandsaw, and 28 mm diameter holes were made for instrumentation. A 6 mm threaded rod with coupling nuts was embedded for potentiometer mounting during testing. The formwork boxes were then painted with two coats of acrylic paint to prevent water absorption and reduce drying shrinkage and cracking. These design features reflect consideration of real-world construction practices and constraints and aim to provide a realistic simulation of load-bearing behavior. This focus on the details of formwork construction emphasizes the meticulous preparation required to achieve accurate and reliable results within the study. The choice of materials and construction methods, therefore, are not incidental but are designed to mimic the conditions found in actual timber-concrete structures.

The push-out tests were conducted within the Structures Laboratory of the Department of Civil and Natural Resources Engineering at the University of Canterbury. A Universal Testing Machine served as the primary testing apparatus. Instrumentation was critical for precise data acquisition. Specifically, 50 mm potentiometers (±0.4% accuracy) – P1, P2, P5, and P6 – measured relative slips within the connections, with potentiometer P3 dedicated to measuring horizontal slip. Measuring horizontal slip was crucial for observing potential separation between the LVL and concrete, which would be indicative of a failure of the connection. The initial tests showed that without restraint, significant separation could occur due to bending moments induced by the load. To address this and to more accurately simulate the in-service conditions (where shear is the dominant loading), steel straps were strategically used to restrain the specimens during the testing process. Calibration of all instruments, including load cells and data acquisition boxes, was meticulously performed to ensure data accuracy. A Universal Data Logger (UDL) software was utilized to record the load-relative slip relationships for each channel of every test.

Each specimen, weighing between 100 and 150 kg, was carefully positioned on a custom-made steel test platform using an overhead crane. The platform, placed on the Universal Testing Machine's rail, ensured that only the concrete flanges were supported to avoid interference. A 20 mm thick steel plate was positioned atop the LVL web for even load distribution. The dimensions of this plate were carefully matched to the specimen’s geometry. The test platform was equipped with a stopper knob to prevent eccentricity during positioning under the loading ram of the testing machine. The careful positioning of the specimen under the loading ram emphasizes the importance of minimizing any potential sources of error or variability in the loading procedure. The attention paid to these details highlights the significant effort taken to control testing parameters and ensure accurate measurement of the response of the specimens. This setup reflects an understanding that accurate load application is paramount for generating reliable results from experimental testing.

The connection push-out tests followed a specific loading regime adhering to EN 26891 (CEN, 1991). The load was applied in shear at a rate of 0.2F_est kN per minute (F_est being the estimated strength). This incremental loading approach aimed at achieving reliable and accurate results. The loading process involved an initial load to 0.4F_est, held for 30 seconds, followed by unloading to 0.1F_est and holding for another 30 seconds, before finally loading the specimen to failure or to a maximum slip of 20 mm. This initial load-unload cycle played a crucial role in the testing procedure and served the purpose of eliminating any internal friction and pre-existing slack within the connection. The initial phases of the loading regime were thus designed to achieve a consistent starting point for all specimens and mitigate any potential influence of these factors on the final recorded results. The UDL software automatically plotted the load-relative slip data, providing a real-time visualization of the specimen's response during testing. The use of the standard EN 26891 further enhanced the reliability and comparability of the results obtained from these tests.

The long-term push-out tests aimed to determine the creep coefficient of different timber-concrete connections over an extended period. Three connection types were selected for this phase of the research: (1) Triangular notch 30°_60° 137l × 60d coach screw φ 16 – T; (2) Rectangular notch 300l × 50d coach screw φ 16 – R; and (3) Toothed metal plate 2 × 333l staggered – P. Three test frames were specifically designed and constructed for this purpose, beginning with planning in September 2007 and construction in January 2008, completed by late April 2008. Testing commenced May 19, 2008. In addition to the loaded specimens, three unloaded 'dummy' specimens of each connection type were also set up to monitor any displacement caused by the specimens' self-weight, fluctuations in temperature, and changes in relative humidity. This setup ensures a more comprehensive understanding of the factors influencing long-term behavior and the creep characteristics of the different connections.

The design of the test frames was driven by the need to apply sustained loads over a long duration. The frames incorporated existing concrete blocks (1000b × 1000w × 330h mm, weighing 823 kg each or 8.07 kN) as weights. The service load on each push-out specimen was defined as 0.3F_max, representing the quasi-permanent part of the serviceability design load (F_max being the maximum strength determined from the short-term push-out tests). The length of each frame was calculated based on the concrete block weight to ensure the desired force was applied to the specimen. The 400 mm distance from the load application point to the pivot point was carefully chosen to optimize the specimen's position within the frame. The frames themselves were constructed using universal column 200 UC 46.2 kg/m for the top horizontal member, and single or double parallel flange channels 200 PFC 22.9 kg/m for the other horizontal, vertical, and stabilizing members. This design reflects a robust engineering approach focused on precise load application and control over environmental factors affecting the long-term behavior of the connections.

The construction of the frames began with cutting 6m-long universal columns and parallel flange channels to the required lengths, according to the drawings (Figures A8-2 to A8-4 detail the frame designs, and Figures A8-5 and A8-6 show connection details). Assembly took place in a garage near the University of Canterbury due to logistical and handling considerations. The confined space and the weight of the components required careful planning. Two specialized lifting frames—one high frame with pulleys and a chain block, and one low frame with rollers—were constructed to facilitate the assembly process (Figure A8-7 illustrates the use of these frames). The assembly of all three frames with specimens took eight days in May 2008. The handling of the 100-150 kg specimens required two people due to their irregular shapes. Steel plates were used to level the specimens and eliminate eccentricities in the setup. The detailed description of the construction process emphasizes the considerable logistical challenges encountered, and the robust design of the test setup and its adaptation to the environmental conditions.

Three timber-concrete composite (TCC) beams were constructed and tested under sustained load to study long-term behavior. The construction and setup of a 600 mm wide TCC beam with a single LVL and a 300 mm rectangular notched coach screw connection are detailed in Figures A9-1 to A9-3. Figures A9-4 to A9-7 illustrate the construction and setup of a 1200 mm wide TCC beam with a double LVL and a metal plate connection. These different beam configurations allowed for a comparison of the long-term performance under sustained load for varying beam designs and connection types. The construction details highlight the methods and materials used to assemble these complex composite structures, providing insights into the practical aspects of building these types of beams and allowing for a thorough understanding of their mechanical and structural features. This detailed documentation of beam assembly is critical for reproducibility and ensuring the reliability of the experimental data gathered.

Precise measurement of displacements and environmental conditions were crucial for the long-term beam tests. 30 mm potentiometers (±0.7% accuracy) measured displacements at midspan and at the supports of each beam (Figure A9-8(a)). Relative humidity and temperature were monitored using HIH-4000 Series humidity sensors and LM-35 temperature sensors, respectively (Figure A9-8(b)). A high sampling rate was employed: every minute during the initial concreting process and the first 24 hours, and hourly thereafter. All sensors and potentiometers underwent calibration before use. Each sensor was assigned a specific channel in a data acquisition box (Figure A9-8(c)), which was connected to a computer for automated data recording (Figure A9-8(d)). This comprehensive instrumentation setup demonstrates a commitment to high-quality data collection, ensuring that even subtle changes over time would be captured and analyzed for a thorough understanding of the long-term behavior of the beams under sustained load conditions. The detailed description of this data logging system highlights the importance placed on accurate and continuous data acquisition for this type of long-term experimental analysis.

The long-term beam tests were conducted in an unheated, uncontrolled environment within a garage. While sheltered, the garage lacked wall insulation, exposing the beams to natural temperature and humidity variations. The beams were subjected to sustained loading conditions for an extended duration, allowing for a thorough investigation of their long-term response and material properties. This un-controlled environment introduces a notable variability, which reflects the challenge of simulating real-world structural exposure conditions under long-term loading. The choice to conduct the test in this environment represents a compromise between the need for environmental control in a laboratory and the desire to assess structural performance under more variable and realistic scenarios. The findings from this study will account for this aspect of the environmental variability, making the results relevant to practical applications.

The design span tables presented aim to provide practical guidance for engineers designing semi-prefabricated LVL-concrete composite floors. These tables offer a quick reference for determining safe live loads (in kN/m²) for various spans. The tables focus on M-section modules with a total width of 2400 mm (Figure A10-1, not shown here), a common configuration in composite flooring systems. The analysis incorporates three distinct connection types: (1) A 300 mm long rectangular notch reinforced with a 16 mm diameter coach screw (R-300); (2) A triangular notch reinforced with the same coach screw (T); and (3) Two 333 mm long toothed metal plates (P) (Figure A10-2, also not shown). The selection of these three connection types allows for a comparison of their respective load-bearing capacities and suitability for different design scenarios. The span tables provide practical design data based on the experimental results presented in the preceding sections, translating the research findings into a readily usable format for engineering design.

The span tables specifically address three connection types, each with its own set of characteristics influencing its performance within the composite floor system. The R-300 connection employs a rectangular notch reinforced with a 16 mm diameter coach screw, which represents a common and relatively simple connection detail. The T connection uses a triangular notch, also reinforced by a 16 mm diameter coach screw. The P connection, on the other hand, is different, using two 333 mm long toothed metal plates. These connections, therefore, represent a range of connection detail complexity and their corresponding structural performance and load-bearing capabilities. The inclusion of these diverse connection types emphasizes the adaptability of the presented design data and provides options suitable for various design preferences and constraints. Each connection type’s performance characteristics are determined via the experimental testing and analysis presented elsewhere in the document.

Beyond the span tables themselves, additional design considerations are mentioned, underscoring the comprehensive approach to composite floor design. The document states that vibration has been checked for a 1 kN point load at mid-span, with the resulting deflection being less than 1 mm. This verification ensures that the design is appropriate and accounts for the dynamic loads and the resulting deflection that may arise due to these dynamic loadings. Importantly, the document acknowledges that the effect of inelastic strains in the concrete has not been considered in the analysis presented. This omission is significant and emphasizes a limitation of the presented design tables. The document notes that this may cause excess deflection in the long term, depending on factors such as environmental conditions, concrete type and thickness, and the type of connection used. This statement highlights the potential need for further refinement of the design approach for extended service life.