Long-Term Performance of Epoxy Adhesive Anchor Systems (2013)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

D-1 A P P E N D I X D Anchor Pullout Tests—University of Stuttgart

D-2 ANCHOR PULLOUT TESTS—UNIVERSITY OF STUTTGART This section presents the test program conducted at the University of Stuttgart Intsitüt für Werkstoffe im Bauwesen (IWB) to investigate the effect of various parameters on the sustained load performance of three adhesive anchor systems. Test apparatus This section describes the test apparatus used for the short-term and sustained load (creep) tests. Short-Term Test Apparatus The testing apparatus for the short-term test used a 3.5” diameter x 0.04” thick Teflon PTFE (Polytetrafluoroethylene) confining sheet with a 1” diameter hole in the middle placed under a circular 1.7” thick steel confining plate with a clearance hole of 0.8” (Figure 1). The confining sheet was used to correct for any surface irregularities in the concrete. Figure 1: Test specimen with PTFE sheet and confining plate installed. Figure 2: Transducer mount on top of the test specimen. Figure 3: Tripod on top of the test specimens.

D-3 The transducer mount was placed on top of the test specimen before the tripod for the hydraulic ram and the load cell was installed (Figure 2). The tripod consisted of an upper triangular steel plate connected to a lower circular steel plate by three M24 threaded rods at a distance of 14” (Figure 3). Figure 4: Hydraulic ram and load cell on top of the tripod. Figure 5: Coupler installed between loading rod and anchor. A LUKAS Model LZOH 10/50-20 22-kip hydraulic ram and a HBM model C6 45-kip load cell were attached on top of the tripod, using self-centering steel adapters. A M20 threaded loading rod was passed through the ram and load cell and was secured at the top with a washer

D-4 and a nut (Figure 4). At the bottom, the loading rod was connected to the coupler which was connected to the anchor (Figure 5). The coupler provided a non-rigid connection between the anchor and the loading rod by utilizing axial spherical plain bearings at all connections except the connection to the loading rod. The coupler also allowed for the positioning of the linear transducer directly on top of the anchor (direct measuring). Transducer mount. Figure 6 shows the transducer mount on top of the test specimen with the linear transducer installed. Figure 6: Transducer mount. Figure 7: Transducer mount and transducer installed. The transducer was clamped into an aluminum cross-beam that was mounted on a steel ring with two threaded rods. The threaded rods penetrated the ring and served as feet for the mount. With an additional threaded rod, the steel ring worked as a tripod. The mount (without cross-beam and transducer) was placed directly on the concrete surface before the tripod of the hydraulic ram and the load cell was installed.

D-5 Figure 7 shows the transducer mount after the installation of the tripod. The cross-beam with the transducer was adjusted, fixed to the mount, and locked with two size M10 nuts. Two coil springs were placed on the locknuts and connected to two levers attached to the tripod. The springs would push the mount downward to the concrete surface keeping it in position without transferring vibrations or horizontal loads from the test rig to the transducer during loading. Except for the springs, the transducer mount had no contact to the rest of the test rig. Standard Sustained Load (Creep) Test Apparatus The testing apparatus for the sustained load (creep) tests used the same Teflon PTFE (Polytetrafluoroethylene) confining sheet. Instead of the steel confining plate that was used in the short-term load test apparatus, a two-part confining plate was used for the sustained load tests to make the installation of the test specimens easier. The dimensions of the confinement sheet and plate were unchanged. The equipment for sustained load testing of bonded anchors at the IWB, University of Stuttgart, was developed by IWB personnel in 2008. Two different types of heating chambers were developed. Figure 8 shows the large heating box with two back-to-back heating chambers. Each heating chamber contained three single test rigs. There were six large heating boxes installed at the IWB with a total number of 36 test rigs. Figure 9 shows the small heating chamber that contained a single test rig. There were 26 small heating chambers installed at the IWB.

D-6 Figure 8: Illustration of the large heating chamber (containing three test rigs). Figure 9: Small heating chamber (containing a single test rig). To apply the sustained loads, large packages of disc springs were used (Figure 10). They provided low spring stiffness, which minimized the loss of load when the anchors displaced. The disc springs were manufactured by Schnorr GmbH, Sindelfingen. The item numbers of the types used are 021 400 (6” x 3” x 0.31,” max. ~ 20 kips) and 021 350 (6” x 3” x 0.23,” max. ~ 11 kips). The spring characteristics are shown in Figure 11.

D-7 Figure 10: Disc-spring package. Figure 11: Disc-spring characteristics. A spring package usually consisted of at least 28 disc springs. Before the packages could be used for the tests, they were loaded for several days with the required test load to avoid any relaxation effect of the springs during testing. The M12 diameter anchor was connected to the M20 diameter loading rod by means of the same non-rigid coupler as in the static load test apparatus. Linear-pots were used to measure displacement in the same configuration as in the static load test apparatus.

D-8 Figure 12: Loading system. Figure 13: Loading system installed. A special loading system was developed to apply the loads to the spring packages (Figure 12). Two LUKAS Model LFC 23/11 (50 kip) hydraulic rams were placed on top of a Burster Model 8526 22-kip load cell. It was designed to avoid any effects to the applied loads from deformations of the test rig and to avoid any risk to the operator in case of failing during the loading process. No loss of load had to be taken into account due to unloading of the loading system. During loading, the load cell measured the transfer of force from the spring to the anchor. Once the desired load was achieved, the nut between coupler and anchor was tightened and the pressure in the rams was released. Specimen preparation The test specimens consisted of three parts; the concrete test member, the adhesive, and the anchor rod.

D-9 Concrete Test Member The concrete test members for the short-term tests and the sustained load (creep) tests were poured in steel cylinders with an 8” inner diameter, 6” height, and a wall thickness of ¼” (Figure 14). No reinforcement was used. In addition, 30” x 30” x 6” concrete slabs were cast from the same batches for additional tests for test series 05, 06, and for extras. The pour dates and number of test blocks produced in each series are listed in Table 1. Figure 14: Typical concrete test specimen. Table 1: Concrete pour details. Concrete Mix Pour Date Number of Cylindrical Test Members Number of Sustained Load Test Members Notes A April 07, 2010 85 6 - B September 17, 2010 100 3 - The concrete for all mixes was batched, mixed, and placed by Friedrich Rau GmbH & Co. KG, Ebhausen, according to DIN EN 206-1. Concrete with round river gravel without any admixtures was specified with a mean compressive strength between 4000 and 6000 psi during testing. All of the materials were batched by weight. After mixing the concrete was placed into the steel cylinders with shovels and vibrated on a vibration table. Due to size of the concrete mixer, the concrete for the research project was made in two batches. The concrete mix designs are included in appendix G.

D-10 Following the pour, the concrete was cured according to DIN EN 206-1 for 28 days. Concrete compressive strength was determined by testing the cubes in general accordance with DIN EN 206-1 on a Toni-Technik model 1515 compression machine at the laboratory of the IWB, University of Stuttgart (Figure 15). The average compressive strength for each series is presented in Table 2 and Table 3. Figure 15: Compression machine. Table 2: Concrete series US-A average compressive strength. Concrete Series Pour Date Average Compressive Strength (psi) 7 day 28 day 41 day 82 day 86 day 462 day1 A April 07, 2010 3902 - 5279 5656 6280 5787 1Due to the unexpected long test period, the last group of four test samples of the first batch had to be used for compression testing when series S5 and S6 were started. The compressive strength at the end of the project has to be estimated (e.g. according to Weber, 1979).

D-11 Table 3: Concrete series US-B average compressive strength. Concrete Series Pour Date Average Compressive Strength (psi) 7 day 28 day 538 day B Sept. 17, 2010 3031 4279 6193 Adhesive The same three adhesives identified earlier were used. The three adhesive products were stored at the laboratory of the IWB. Because it was not possible to environmentally control the whole laboratory the adhesives had to be conditioned to the specified setting temperature prior to every installation. This conditioning was done in a Noske-Kaeser Model KSP 502/40 H climate chamber at the laboratory. Anchor Rods Size M12 threaded rods and nuts were used as specified in ISO 1502. The steel grade was 12.9, which corresponds to 174 ksi ultimate strength and 157 ksi yield strength. The rods were galvanized to prevent rusting and to ensure nearly identical surface properties for all tests, even if the batch or the manufacturer changed during the project. They were delivered by Ferdinand Gross GmbH & Co. KG of Leinfelden-Echterdingen. The anchor rods were cut to a length of 6.7” from 39” stock and their ends ground. The bottom end of the anchor was ground to a 45° cone (Figure 16) in order to fit into a centering guide placed at the bottom of the drilled hole (except for test series 5 and 6). Prior to installation, the rods were cleaned with acetone and allowed to air dry. Figure 16: Anchor showing 45° cone to fit into centering guide. Instrumentation Measurement Displacement. A direct measurement of the anchor displacement was measured with Novotechnik model TRS25 potentiometric linear transducers.

D-12 Load. The tension in the anchor was measured indirectly as a compressive reaction of either the hydraulic ram or the compression spring in the test apparatus. For the short-term tests, the load was measured by a HBM Model C6 45 kip load cell connected to the HBM Model Spider-8 data acquisition system. The data logging was done with a PC using DIAdem 10. For the sustained load (creep) tests, the loads during the loading process were measured by a Burster Model 8526 22 kip load cell. Once the load was applied to the anchor, the loading system including the load cell was removed. After removing the load cell, it was not possible to monitor the loads that were actually applied to the anchor without disturbing the tests. However, the stiffness of the spring packages was chosen to limit the loss of load to 2% for an anchor movement of 0.04.” Temperature. The temperature of the test specimens could not be measured directly during the tests. Due to the small specimen size, a hole could not be drilled without affecting the load distribution inside the specimens. To guarantee that the required temperatures were reached before loading transfer, the required conditioning times were determined in advance using a thermocouple-equipped anchor set into a standard test specimen. This temperature calibration specimen was connected to the HBM Model Spider 8 data acquisition system and calibrated using a Testo Model t110 digital thermometer, calibrated on October 08, 2009 and October 24, 2011. The data logging was performed with a PC running DIAdem 10. Ambient air temperature in each test chamber was measured and controlled by GEFRAN Model 400-DR-1 temperature controllers. The temperature sensors attached to the controllers were Electrotherm Type K2RS PT100 sensors. All heating systems, consisting of sensor, controller, and heating elements were calibrated using the temperature calibration specimen. It was not possible to monitor the temperature of each chamber. The function of the temperature controllers were checked periodically with a calibrated Testo Model t110 digital thermometer. Humidity. The relative humidity within the test chambers could not be measured and controlled. Time. Time was measured using the computer’s internal clock.

D-13 Instrument Calibration Displacement in short-term tests. The Novotechnik Model TRS25 potentiometric linear transducer used in the short-term tests was calibrated on May 26, 2010, using the Mitutoyo Gauge Block Set No. BM3-32-1/PD, calibrated on April 2008. Displacement in sustained load tests. The Novotechnik Model TRS25 potentiometric linear transducer used in the sustained load test could not be calibrated as the creep displacements that occur in sustained load tests were below the accuracy guaranteed by the manufacturer. The accuracy was also affected by the increased temperature. Therefore all measured creep displacements could only be judged qualitatively. All transducers used for long- term testing were checked for proper functioning before each test. The measurements were adjusted for variations in power supply voltage and normalized to a 9 volt power supply. Load. The Burster Model 8526 22 kip load cell of the loading system for the sustained load tests was periodically calibrated against one of three HBM Model C6 45 kip load cells that were used in the short-term tests and for the loading of the disc-spring packages in the sustained load tests. The HBM load cells were calibrated on October 28, 2009 & October 17, 2011 (ID no.: KMD006), December 11, 2009 & December 13, 2009 (ID no.: KMD009) and July 12, 2010 (ID no.: KMD010) at the MPA Stuttgart (Material Testing Institute University of Stuttgart) according to DKD standards (Deutscher Kalibrierdienst). The spring packages were not calibrated as all loads were applied using a calibrated load cell. For determining the loss of load due to anchor movement the spring constants provided by the manufacturer were used. Compressive Strength. The compressive strength of concrete cubes was tested on a Toni-Technik Model 1515 compression machine, calibrated on September 11, 2008 and September 13, 2010 at the MPA Stuttgart (Material Testing Institute University of Stuttgart) according to DKD standards (Deutscher Kalibrierdienst). Temperature. For the conditioning of the test specimens prior to the installation of the anchors and for conditioning at elevated test temperature in the short-term test, the Noske-Kaeser Model KSP 502/40 H climate chamber was used, calibrated on August 06, 2008 and August 18, 2010. For the periodical checking of the temperature of the heating chambers, the Testo Model t110 digital thermometer was used, factory calibrated on October 08, 2009 and October 24, 2011 according to DKD standards (Deutscher Kalibrierdienst). The temperature sensors in the test

D-14 chambers were not calibrated separately but in combination with their controller and heating elements, using an original test specimen with a thermocouple-equipped anchor installed. This temperature calibration specimen was calibrated against the Testo Model t110 digital thermometer. Humidity. There were no humidity sensors installed. Environmental control Standard Temperature The adhesive was stored at the laboratory of the IWB without special air conditioning. The temperature was 73°F ±9°F (23°C ±5°C). Prior to installation, the test specimens, the anchors, and the adhesives were conditioned in the Noske-Kaeser Model KSP 502/40 H climate chamber (Figure 17). The chamber had a temperature range of -40°F to 356°F (-40°C to 180°C). Elevated Temperature The Noske-Kaeser Model KSP 502/40 H climate chamber was used for conditioning of the test specimens prior to every installation and for elevated testing temperatures in short-term tests.

D-15 Figure 17: Climate chamber. Data management and acquisition Generally a Microsoft compatible computer was used for the data acquisition. For short- term testing, National Instruments data acquisition software DIAdem10 was installed with special drivers for the HBM Model Spider-8 data acquisition system. The measured values included load and displacement. For sustained load testing, the Measure Foundry 5 data acquisition software from Data Translation was used together with the Data Translation DT9803 USB-connected measuring device. A special setup was developed at the IWB to automatically acquire and log the data. Measured values included the power supply voltage (9 volt), transducer voltage, and time. The voltage of the transducers represented the relative position of the transducers, whereas 0V represented the minimum transducer position and 9V represented the maximum transducer position. No calculations were performed before logging. Generally the logging interval was 10 minutes.

D-16 Due to minor fluctuations in the 9 volt power supply, the Data Translation program recorded the power supply voltage with each data reading and the readings of the transducer positions were appropriately adjusted to a normalized 9 volt power supply. Data Acquisition Software for Short-term Tests For short-term testing, DIAdem 10 was used, published by National Instruments (Figure 18). A load versus displacement curve was displayed on the screen for real-time feedback. Load, displacement, and time readings were recorded at a frequency of 5 Hz and stored as a Microsoft Excel file for analyzing. Figure 18: Screenshot of NI Diadem 10.2 data acquisition program. Data Acquisition Software for Sustained Load (Creep) Tests For sustained load testing, Measure Foundry 5 was used, designed especially for the data acquisition devices from Data Translation. It used a graphical programming interface that gave access to every function of the measuring device and let the user build customized setups. Since most of the test chambers were not located in the IWB laboratory, the tests could not be observed

D-17 daily. Therefore it was decided to build a very robust setup that only triggered the data acquisition of the measuring device in a 10 minute interval and wrote the transferred data as a simple ASCII-file to the hard disk (Figure 19). Further analysis was done in a second process using Microsoft Excel. Figure 19: Sustained load test setup, built with Data Translation Measure Foundry 5 (screen shot). Installation procedure The standard installation procedure is described below which was followed for test series 2, 3, 4, 5, 6, 8, 10 and 11. Standard Baseline Installation Procedure All anchors were installed according to the MPII. The holes were created with a 0.55” (14mm) carbide tipped concrete bit as specified by the manufacturer and a Hilti Model TE36 hammer drill. A drilling jig (Figure 20) with a depth stop was used to ensure that the holes were

D-18 drilled perpendicular to the surface of the concrete and to the correct depth. The holes were drilled 0.6” deeper than then embedment depth to allow for the placement of a centering guide at the bottom of the hole. The spoil at the concrete surface was removed with a vacuum prior to cleaning the hole. The holes were cleaned according to the MPII, which generally included blowing with oil-free compressed air, brushing with a steel brush provided by the manufacturer, and then blowing again with compressed air until no dust was discharged from the hole. Durations and numbers of brushing/blowing cycles varied by manufacturer, but for each case the holes were cleaned according to the MPII. To prevent dust from blowing into the operator’s mouth and eyes, an adaptor for the vacuum (Figure 21) was used to capture the dust ejected from the hole when blowing with compressed air. This adaptor attached to the vacuum hose and allowed the compressed air nozzle to be easily inserted and removed. Figure 20: Drilling rig and hammer drill. Figure 21: Vacuum adaptor.

D-19 Prior to installation, a centering tool (Figure 22 and Figure 23) was inserted into the hole and a concentric circle was drawn on the surface of the concrete to aid in centering the anchor. A plastic centering guide (Figure 24) was placed in the bottom of the hole. Figure 22: Centering tool. Figure 23: Centering tool inserted in hole. The adhesive products were dispensed with a manufacturer supplied cartridge gun. According to the MPII, several squeezes of adhesive were discharged and disposed of before dispensing into the holes to ensure that the adhesive was of uniform color and consistency indicating that it was properly/thoroughly mixed. The anchors were wiped clean with acetone and allowed to air dry. The anchor rod was rotated counterclockwise and jiggled while it was installed in the hole until the tip of the anchor seated into the centering guide at the bottom of the hole. Excess adhesive was wiped from the surface and the centering ring coated with wax to prevent adhering with the adhesive (Figure 25) was placed over the anchor and the anchor was centered and plumbed vertical. The anchors were left undisturbed during the specified gel/working time and the adhesive was allowed to cure for seven days prior to conditioning.

D-20 Figure 24: Centering guide with anchor. Figure 25: Anchor with centering ring. Exceptions to the Standard Baseline Installation Procedure All tests of series 5 and 6 were conducted without using a centering guide. All of the tests of series 10 and 11 were installed at 32°F (0°C). Specimen conditioning Upon completion of the 7 day adhesive curing period, the temperature conditioning started. Usually the conditioning took 18 hours. Immediately after conditioning the short-term tests were conducted and the sustained load tests were loaded. Testing procedure The standard testing procedures for the short-term and sustained load (creep) tests are described below which were followed for test series 2, 3, 4, 5, 6, 8, 10, 11 Short-Term Test Procedure Immediately prior to testing, the test specimen was removed from the climate chamber and exposed to normal ambient temperature. To avoid non-admissible loss of temperature the installation of the test apparatus described below was finished within 120 seconds. A 0.04” thick PTFE confining sheet and 1.7” thick circular steel confining plate were placed over the anchor. The measuring mount for the linear transducer was placed on top of the concrete surface with the anchor centered in the middle of the base ring. The tripod, the LUKAS Model LZOH 10/50-20

D-21 22-kip hydraulic ram, the HBM Model C6 45-kip load cell and the loading rod were placed together on top of the confining plate. The coupler was attached to the anchor and connected to the loading rod. Finally the cross bar with the transducer installed was attached to the measuring mount and the coil springs were installed to keep the measuring mount in position. The hydraulic ram was connected to an electrically operated hydraulic pump installed inside a measuring cabinet. The transducer and the load cell were connected to a HBM Spider 8 data acquisition system installed inside the measuring cabinet, connected to a PC running DIAdem under Microsoft Windows-Vista. The software was initialized with the appropriate sensor parameters (calibration factors, etc.) and checked for proper functioning. The anchors were loaded at a constant pump rate (i.e., displacement-controlled). The pump rate was adjusted to get failure within 60 to 180 seconds as specified in the ETAG001. The DIAdem program automatically recorded the test data in a proprietary format. After finishing the tests, the data was exported to a MS Excel spreadsheet. Standard Sustained Load (Creep) Test Procedure The disc-spring packages were compressed using the same LUKAS LZOH 10/50-20 22- kip and HBM Model C6 45-kip load cells that were used in the short-term tests and placed into the test rig between two triangular steel plates (Figure 26). Both steel plates were aligned by the same size M30 treaded rods that passed through holes in each corner of the plates surrounding the disc-spring package. Both triangular steel plates could be locked with M30 nuts in any desired vertical position along the M30 treaded rods. To compress the spring packages, the upper steel plate was locked and the lower steel plate pushed upwards. When the desired load was reached, the lower steel plate was locked and the pressure in the hydraulic ram released. In the following paragraph, the set of triangular steel plates with the compressed disc spring package fixed in between is referred to as the “spring frame”. After the spring frame was adjusted, a third triangular steel plate was installed to the top of the test rig and fixed with M30 nuts in the same way. From above, the same measuring mount that was used in the short-term tests was attached upside-down to the third steel plate. The test specimen was placed upside-down on top of the third steel plate in the same manner, with the PTFE confining sheet and the two-part steel confining plate placed in between. The coupler was

D-22 installed to the anchor as described in the short-term tests. The loading rod was passed through the spring package and the upper end was connected to the bottom of the coupler. A nut was attached to the lower end of the loading rod and hand tightened, bearing against the lower triangular steel plate of the spring frame. Figure 26: Illustration of the test rig with a vertical cut. Finally the cross bar with the transducer installed was attached to the measuring mount and the coil springs were installed pushing the measuring mount against the concrete surface Test specimen Spring frame Loading rod Disc spring package Coupler Anchor Cross bar Transducer

D-23 keeping it in position during testing. After installation, the temperature was raised to the test temperature. Once the conditioning period elapsed, the loading system was attached to the lower end of the loading rod bearing against the lower steel plate of the spring frame. The load was applied to the anchor with a hand operated hydraulic pump. Once the preload force of the spring package was reached, the M30 nuts that supported the lower steel plate of the spring frame were loosened and screwed downwards before the nut at the lower end of the loading rod was screwed upwards against the lower steel plate of the spring frame and hand tightened. Finally the pressure was released from the hydraulic rams and the loading system was detached. During the loading procedure, the load was permanently observed using the Burster Model 8526 22-kip load cell (which is an integrated part of the loading system) and the HBM Spider 8 data acquisition system. Exceptions to the Standard Baseline Testing Procedure All of the tests of series 3 were conducted at 120°F (49°C) and those of series 4 were conducted at 70°F (21°C). All of the tests of series 8 were conducted in moist concrete during service. The specimens were watered for 24 hrs. Immediately after the watering process, the test specimens were put into plastic bags to prevent them from drying. The anchors were guided through small holes in the plastic bags so that the loading equipment could be attached to the anchors as usual. After heating up the specimens and loading the anchors, the specimens could be checked anytime and rewetted if necessary through the mouth of the bag. All of the tests for series 10 were conducted at 32°F (0°C). Post Test Procedure After sustained load failure occurred the anchors could not be extracted from the test specimens by pulling them out without destroying the remaining mortar shell that surrounds the anchor. Instead the anchors were extracted by splitting the test specimens as follows. The concrete cylinder was pressed out of the surrounding steel ring using a hydraulic ram. A 1” diameter hole was drilled into the concrete at a distance of 0.8” from the anchor. With a special wedge that is usually used to generate cracks in concrete slabs, the concrete cone was split in half. Usually the anchor could be extracted now with some gentle strokes of a hammer perpendicular toward the head of the anchor. Once the anchor was separated from the concrete the actual failure could be determined.