Additional Participants

Senior Personnel

Habib Dagher

William Davids

Douglas Gardner

Barry Goodell

Roberto Lopez-Anido

Michael Peterson

Stephen Shaler

Senthil Vel

Technician, Programmer

Bradley Weiland

Ben West

Project Period

May 2001-April 2004

Level of Access

Open-Access Report

Grant Number


Submission Date



Below are abstracts for the projects conducted by the 10 NSF supported REU students. Included with each abstract are the project supervisor and associated graduate students. The abstracts were written by the students. Structural Durability of FRP Drains in Bridge Decks Micah Florea, Tulane University Advisor: Roberto Lopez Anido Fiber Reinforced Polymers (FRP) are quickly becoming the material of choice to solve civil engineering challenges in corrosive environments. In the state of Maine, where deicing salts are generously applied to roadways every winter, conventional steel down spouts used to drain concrete bridge decks are quickly corroded and need constant replacing. Looking to lower maintenance costs, the Maine Department of Transportation has joined efforts with the University of Maine to design an FRP down spout to replace the degraded steel ones. Three identical prototypes were fabricated, embedded in a simulated concrete decking, and were submitted to a series of laboratory tests to ensure structural integrity and to optimize the design for longterm durability. The tests included an ice formation study, dual-ramp compression loading tests based on the AASHTO HS25 loading specification, 100,000 cycle fatigue tests, ultimate compression tests, and reverse push out tests. The prototype proved to be acceptable, but not optimal. The prototype design was slightly modified and the new design described in this report is approved as a safe, functional, and durable drain system to be fabricated and installed in bridge decks throughout the state of Maine.

Timber Guardrail Splice Connection Design Chad Gibson, University of Evansville Advisor: William Davids Participating Grad Student: Josh Botting Research into designing an aesthetically pleasing and structurally sound timber reinforced guardrail system at the University of Maine is being sponsored by the New England Transportation Consortium. The project specifies the use of glulam beams, as opposed to the use of solid sawn timbers. The composite design will feature a strip of pultruded E-glass epoxy bonded to the tension side of the guardrail and bonded steel splice connections that will attach the railings to posts. Through modeling with the BARRIER VII program, it is apparent that large tensile forces must be transferred between rail sections. The splice connection design will remain as the central focus of this paper as a result of a tendency for structural failure at these critical locations. From BARRIER VII the projected critical axial load is 40 kips. A splice connection must be designed to transfer the loading across the rail system and be appropriately evaluated in a tension test. The connection, which has been tested, is a steel plate bonded to the FRP and bolted to the splice plate. The critical component of the splice is the steel to FRP bond. In order to assess this bond a test setup has been designed. The results of the test proved comparable to the capacity of standard W-beam guardrails.

Extrusion of Wood-Plastic Composites Elizabeth Jordan, University of Connecticut Advisor: Douglas Gardner As an REU student the focus of my project was supposed to be the extrusion of wood-plastic composites involving nylon and pine wood flour. Unfortunately due to several miss haps with the equipment and machinery this project was not carried out. Sue to the fact that my original project did not pull through I have done a variety of things this summer and I have learned a great deal too. Although it may not sound like a lot I have also been able to gain an understanding of lab work and graduate school though the program this summer. This summer I have seen polypropylene and wood flour extrusion as well as masonite and polypropylene extrusion. Our most recent extrusion involving mixtures of polypropylene and wood flour actually came out very well. after successfully making the wood-four polypropylene wood composite we did a four point bending test on our samples. For this we used the Instron 8801 universal testing machine. With these tests we found the maximum loading strength of our boards and then plotted our results. we also used the instron machine to test some commercial boards. We did water soak tests for these. Our results from these boards were also plotted and recorded. By doing these tests I learned how to use the instron 8801 and how to plot and record data in excel. As stated earlier the woodtruder or Wt 94 was out of order for a good portion of the summer. While the machine was down I used the time to write work instructions and to begin a thermal expansion testing project. For the thermal expansion testing project I prepared samples and obtained a set up to test for the thermal expansion of wood-composites in relationship to time. Before I started my thermal expansion project and during the time that the extruder was still broken I wrote work instructions and helped other students with their projects. After all of this I have no real results to show other than the few tests that we just completed but I did gain a great deal of experience. Even though I did not get to finish my project I learned how research can be unpredictable and how problem solving a key aspect of it all.

Multifunctional Reinforcements for Conventional Wood Composites: Lateral Nail Test Katherine Jordan, University of Connecticut/Materials and Metallurgy Engineering Advisor: Ciprian Pirvu Conventional oriented strand board(OSB) is commonly used in construction. In certain areas that are prone to hurricanes, high winds, and high humidity, such as coastal areas, traditional OSB structures become weakened where they are connected to the wall studs. By desiging an effective edge protection system, moisture can be prevented from entering the wood compostie and causing deterioration. Also, the reinforcment should increase the dimensional stability and connector durability of the compostie. During this project, eight different types of reinforcement were tested for lateral nail holding strengh. These reinforcements were designed to prevent moisture from entering the wood composite at the connection joints and to increase the nail holding strength of the composite. In order to determine which reinforcements performed the best under both wet and dry conditions, a lateral nail test was performed on each type of sample under both wet and dry conditions. The samples were composed of Polyester Resin or Vinyl Ester Resin, Regular OSB and either, Glass Fabric, Aramid Fabric, Chopped Strand Fiberglass, or Fiberglass particles. Advanced OSB, 3/8!inch and 1/2!inch plywood as well as Regular OSB were used as control samples with no reinforcement added. Half of all the samples were soaked in a water bath for twenty-four hours. Each sample was then nialed to a stud and tested on a 22 Kip Instron machine. Together with the results of a nail pullthough test and the moisture content test, the most effective reinforcement can be chosen. From lookinag at the results of the Lateral Nail Test I noticed that Aramid Fabric tended to act well and the _!inch Plywood and Advanced OSB also performed well under wet and dry conditions. Further analysis is necessary in order to determine which method of reinforcement provides the greatest amount of reinforcement at the most reasonable price and with the most efficient application process.

Simulating the Failure and Load Deformation Response of Nailed and Bolted Connections Melissa Kahl, Syracuse University, Civil Engineering Advisor: Eric Landis The prospective applications of computer simulations have a tremendous potential for deciphering of real world situations. One such application would be to the failure of wood that is joined together with either a bolt or nail. If the details of this failure of the two dominate fasteners in wood construction were to be correctly predicted, then it would be possible to begin to take necessary steps to prevent this failure. This purpose of this project was to compare the load deformation curve and the aesthetic distortions of the wood produced by the computer simulation, to those produced by the data collected from a simple tension test of these joints. The computer simulation works by first creating a lattice of the wood specimen through the definition of several variables including the thickness of the wood, its width and height, and the number of nodes. The beams and verticals connecting the nodes where the bolt or nail is located are removed to represent the damage caused by these penetrations, creating a notch in the middle of the mesh. The bolted mesh is represented as a square hole in the mesh, as opposed to the nailed mesh that is represented with the beams around the hole curved to simulate the distorted grain surrounding the notch. Each node is then assigned the appropriate restraints. The nodes on the left border of the notch are given displacements to represent the force from the nail or bolt. The nodes on the right border of the specimen are all treated as fixed connections, excluding the end one which is treated as a pin. All of the other nodes are free to move. The experimental tension test was performed by inserting a bolt through a drilled hole in the wood and then through a piece of steel with a comparable predrilled hole. Steel was used to ensure that the wood failed on the side of the connection that was being recorded. Strain gauges were then attached to measure the extension of the wood. For the nailed connections, a concrete nail was used to hold the wood and the steel together and prevent slippage. All of the 20 tests for each connection were recorded with digital microscopy. The comparative results of the computer simulation and the tension tests were very similar. The typical computer simulation of the nailed and bolted connection predicted a generally linear extension vs. load curves that was very similar to the curves created from the tension test data. The visual deformations of the meshes when compared to the wood specimens were also alike with both models showing the portion of the specimen between the applied load and the wood border ‘pushed out’ of the wood.

Acquisition of the Elastic Constants of Small Diameter Timber Using Non-Destructive Ultra-Sonic Testing Katherine Kwasnik, Boston College/Physics Advisor: Michael Peterson The elastic constants of wood are an important factor in how the wood can be used. The constants that correspond to large diameter wood are well known, making it more desirable than small diameter wood, which has constants that are significantly different from those of mature wood and are not as well understood. Small diameter timber causes a fire hazard in National Forests, and must be thinned out in order for a healthy forest to thrive. However the small diameter trees that are thinned are not often used because the properties are not well understood, and thus it is difficult and costly to thin these forests. The purpose of this project was to use ultra-sonic testing to find the elastic constants of small diameter timber to facilitate the use of these trees. Twenty spherical wooden balls were made out of sitka spruce on a wood lathe and conditioned in an environmental chamber so that the results can later be repeated and confirmed. An apparatus was constructed that allows an ultra-sonic wave to be passed through the ball, and the ball to be turned so that the wave can pass through at different angles and in different incident planes. The transducers were connected to an oscilloscope, and a computer recorded the data. Each ball was turned through 25 angles in each of 4 incident planes. A lens was connected to a television to allow the angles to be accurately measured. This data was processed in a MatLab program to calculate the velocity of the wave in each of the angles. The velocities calculated are then used to calculate the elastic constants of the balls that were tested. The wave was expected to travel at different speeds different planes of symmetry, since wood is an aniostropic material. The velocity curves that were acquired with the data showed that, with a few inconsistencies due to the fact that wood is not a consistent material subject to environmental conditions, the velocities fell into three levels, confirming the fact that there are three planes of symmetry in the samples. This data indicates that the data collected will allow the elastic constants to be found using the method developed. The elastic constants could not be calculated due to an optimization problem in the MatLab program designed to calculate the constants, but the velocity curves support the expectation the elastic constants can be found from this data.

Improved Fluid Penetration in Permeable, Impermeable Woods and Composites: Analysis of the Role of Compressed Air in Pressure Processes Seth McDonald, King College/ Physics Advisor: Barry Goodell Participating Graduate Student: Ben Herzog Previous work has shown that compressed air pockets form in impermeable timber species during pressure treatments. We have hypothesized that venting to atmospheric pressure will allow the compressed air to be released from wood resulting in increased apparent permeability. Our initial work to test this hypothesis has been done with both aqueous dyes in wood, and with a model system using liquid resin to penetrate highly permeable E-glass fabric sandwiched, or bound, between layers of impermeable plastic wood. Aqueous experiments included those with non-vented and single vent samples of 2” x 6” x 2’ long southern yellow pine where venting was to the atmosphere. Samples were submersed in containers of aqueous dye prior to the application of pressure. The applied pressures were 60 and 120 psi for periods of 30, 45, 60 minutes. This round of experiments included full cell processes. For the liquid resin experiments with E-glass, billets of 1” x 6”-2’ plastic lumber were ported with one vent, two vent, and five vents, or were left unvented. The samples were immersed in resin, and the pressures again were 60 and 120 psi for a period of 2 hours. Venting did not appear to increase the dye weight retention in the southern pine. However it did improve penetration of resin in the glass fabric samples. The largest increase in resin infusion appeared to occur when only a single-vent was applied. Increasing additional vents marginally increased resin infusion but this limited increase may have been due to the small size of the samples. Clamping number, pressure, and position all affected the infusion of resin. Although venting did not improve weight retention in the southern pine wood, the penetration data were not examined in this work. Future work should therefore explore how fluid penetration in wood is affected by venting to relieve air compression. A study of the effects of clamping pressure on infusion is also recommended.

Buckling of Wood-Plastic Composite Columns Ryan Vignes, University of Iowa/Mechanical Engineering Advisor: M. Asif Iqbal Due to the navy’s extensive ownership of wooden marine structures, it is extremely interested in enhancing docks and piers to better withstand the hazards of their aqueous environment. Wood-plastic composites as possible replacements for conventional wood in structurally demanding applications have recently become an area for extensive research. Along with woodplastic composites’ increased ability to endure an aqueous environment by resisting fungal decay and marine borers, the composite is also a practical solution to help improve environmental concerns such as rapid deforestation. Since the composite is composed mostly of wood fibers embedded in a plastic matrix, less wood is required for a structural member, resulting in a diminishing demand for logging. Additionally, wood-plastic composites do not require the application of the dangerous and hazardous chemicals used when treating conventional wood. In addition to these attributes, the composite material can be handled and assembled in the same manner as conventional wood; no modification of current construction equipment is necessary [Dagher 01]. Before wood-plastic composite columns can be used in a structural capacity, their reactions and ability to support loads must be evaluated. To perform this evaluation, the material properties must be determined. Column reactions will then be analyzed both experimentally and analytically with ANSYS, a finite element analysis package.

Post Peak Deformation of Wood Loaded Parallel to the Grain Donald Waller, University of Central Florida/Civil Engineering Advisors: Eric Landis Fracture mechanics is based on the study of the stress and displacement at the fracture process zone. This zone is the region around the tip of the crack. For many years wood has been an interesting subject to study in the category of fracture mechanics. Brittle fracture (unstable crack growth) has been known to commonly occur in wood specimens. This projects goal was to determine the complete tensile response of a notched wood specimen. There were no ASTM specifications for this type of experiment so they were determined by pretesting. A dog-boned spruce specimen was determined from the pre-testing and was loaded until fracture at two different rates using the Instron 8801. The control mechanism for our experiment was the crack mouth opening displacement (CMOD). Strain gauges were attached to each side of the specimen and used as the output of the CMOD. The larger of the two was used as the control of the closed loop. The purpose of the closed loop is to keep the CMOD at the rate desired. Eighteen specimens were tested at a rate of 0.05 mm/min. Two data points were collected every second and were used to formulate the load vs. strain graphs. For the most part each the graphs showed unstable crack growth and nothing new was discovered to differentiate the results from earlier experiments done by other researchers. Modifications were made and fifteen more specimens were tested. The rate was changed to 0.02 mm/min and the machine was manually tuned. Automatic tuning was used in the earlier testing and showed a deviation of 1mm in some parts. The first five specimens were tested at the new rate and during the test the strain gauges were unable to find the desired signal and started vibrating. It was too slow for the machine so the final 11 specimens were tested at the original rate of 0.05 mm/min. The same unstable crack growth occurred for the remainder of the tests. The suggestion I have for future study of this subject is to use a different strain gauge setup. Strain still needs to be the control but the extensometers didn’t perform the way we needed them to for us to achieve our desired goal of stable crack growth. Future study, with some modifications and the right tools, has the potential for determining the complete tensile response of a wood specimen.

Active Vibration Control of a Carbon Composite Beam, Using Piezoelectric Actuators and Sensors Simon Weiss, Brown University/Mechanical Engineering Advisor: Senthil Vel “Smart Structures” are mechanical systems that utilize lightweight, energy efficient electronics to actively dampen out mechanical vibrations, where passive damping units are not suitably effective or too cumbersome for the design application. Active Vibration Control (AVC) using piezoelectric actuators (PZT) has applications where weight is a factor, and thus lightweight, stiff composite structures are in use. Space and Aeronautic systems can benefit from PZT AVC because both applications rely on the accuracy and precision of lightweight structures. To determine the relative effectiveness of Active Vibration Control using Piezoelectric (PZT) actuators and sensors, an experimental setup of a cantilevered, thin carbon fiber beam, three PZT actuators (one disturbance actuator, and two active damping actuators) and a PZT sensor, and computer interface and control algorithms was created. The effectiveness of various simple active control algorithms was compared to free vibration of the beam, and a simple PZT-resistor damping system by measuring the frequency response of the system, time-response after excitation, and comparing the calculated damping coefficients. Two methods were used to show the effectiveness of the AVC. The time response functions of the beam show decrease in vibration magnitude with respect to time after momentary disturbance (manually tapped). Because of the method used to begin vibrations in the beams, the initial magnitudes from test to test are not necessarily consistent, but the data still provides useful qualitative and quantitative information. As can be seen from the graphs, it takes the model without damping 8 seconds to reach 10% its initial vibration magnitude, while the AVC damped version takes less than 1 second. The frequency response functions (FRF) of the beam with and without Active Vibration Control (AVC) shows how the AVC affects the first and second vibration modes. As is apparent from the slopes of the mode peaks from AVC OFF vs. AVC ON, the severity of slope and magnitude of the vibration is substantially greater without Active Vibration Control. The test results and other information acquired this summer show the effectiveness of active vibration control using PZT material relative to passively and free-damped systems.

Files over 10MB may be slow to open. For best results, right-click and select "save as..."