Strengthening.com - CFRP Strengthening and Load Testing of a Parking Garage: A Case Study

The Ashford Perimeter Parking Garage is a unique structure that supplies approximately 1,000 parking spaces for a commercial office rental property in Atlanta, Georgia. The 6-story building (approximately 350,000 sq. ft.) was built in 1983, and the garage was added in 1984. The garage is an independent structure that consists of four parking levels and is connected to the building by a link bridge. The parking levels are arranged in a terraced “cascade” configuration and each level is accessed via an external on-grade roadway. There are no internal structural ramps (see Figure 1). The garage deck is 8 in. thick, 2-way, cast-in-place concrete, post-tensioned, flat slab supported by cast-in-place concrete columns with drop panels. The typical bay size is 27 ft. in the E-W direction, and 25 ft. or 35 ft. in the N-S direction; there are also cantilever spans at the terraced end levels two through four.

During the construction of the garage, it was discovered that the post-tensioning system in the column strip in the E-W direction (27 ft. spans) of the slab was inadequate (approximately 30% to 40% deficiency). A remedial repair was developed by the original designer of the garage that consisted of a heavily-reinforced (nine #6 bars), 3 in. thick and 3 ft. wide Gunite beam that was added to the underside of the slab between the drop panels (see Figure 2). The construction was substantially completed in 1984. In late 1984, the first delamination between some of the Gunite beams and the underside of the slab was noticed. During a non-destructive testing (NDT) investigation that followed (pulse velocity test), it was concluded that approximately 30 percent of the Gunite beams may have more than 40% delamination. In June 1989, a Gunite beam debonded from the underside of the fourth level slab and fell to the slab below. Initial conclusions related the delamination to inadequate bond that was associated with poor sandblasting of the underside of the and to carbonation found on the underside of the slab that was most likely caused by the use of unvented heaters during installation of the Gunite beams.

After the last change in ownership (1999), a specialty engineering firm retained by a new building owner recommended a new and more comprehensive NDT inspection utilizing impact echo. Of the 133 beams tested, 105 showed some delamination, and, of these, 19 had 33% or more delamination. According to the available structural drawings, the garage slab was constructed using ½ in. diameter, low-relaxation, 7-wire-strand, unbonded tendons with an effective stress of 175 ksi. For the current investigation, the structural floor was modeled using finite element modeling software. In addition to the dead weight of the structure, an additional 2 psf dead load was added to account for the electrical hardware attached to the underside of the slab. A reduced live load of 30 psf per 1991 Southern Building Code was assumed; a 50 psf live load, without reduction, is required by the current building code.

The results of the FE analysis for the factored load condition are summarized in the table below. The calculated strengths and demand moments are for the full width of a bay. The redistribution moment values reflect adjustments allowed by ACI 318-99 to account for the hinging in redundant structures. Moment redistribution was used to eliminate or reduce the negative-moment region deficiencies. As shown in Table 1, the finite element analysis of the floor slab indicated that the slab is overstressed at mid-span by 83% in the E-W direction and 56% in the N-S direction (35 ft. span). The slab is also overstressed in the N-S direction at the supports by 10%, which was assumed to be within acceptable limits.

Direction, Span	Location	Demand Moment (ft - k)	Redistributed Moment (if necessary)	Strength (ft-k)	Comment
E-W, 27 ft	Mid-Span	137	201	110	83% overstress
E-W, 27 ft	Face of Capitol	-67	N/A	-149	OK
E-W, 27 ft	Column Face	-360	-296	-269	10% overstress
N-S, 25 ft	Mid-Span	60	N/A	209	OK
N-S, 25 ft	Face of Capitol	-67	N/A	-192	OK
N-S, 25 ft	Column Face	-288	N/A	-389	OK
N-S, 35 ft	Mid-Span	294	327	209	56% overstress
N-S, 35 ft	Face of Capitol	-127	N/A	-223	OK
N-S, 35 ft	Column Face	-429	-396	-396	OK

REMEDIAL PROGRAM

Carbon fiber reinforced polymer (CFRP) reinforcement was considered for flexural strengthening of the deficient positive moment regions using the recommended guidelines of ACI Committee 440 (2000). These guidelines have limits to ensure that the CFRP strengthened structure will not collapse in case the CFRP reinforcement is lost due to fire or vandalism. This structure was found to be slightly exceeding this limit. However, it was decided that a CFRP strengthening is warranted if the performance of the upgraded member is verified by in-situ load testing.

Externally Bonded Carbon FRP Reinforcement

The design of CFRP strengthening was based on the premise that the structure should not collapse if the CFRP is lost due to fire or vandalism. This requirement was deemed met if the strength of the unstrengthened structure is adequate for 1.2 (dead load) + 0.85 (live load). The results of the analyses indicate that these limits are exceeded and the calculated overstresses in the E-W and N-S directions. However, if a satisfactory performance of the slab without CFRP is verified by load testing, a relaxation of these limits can be justified.

The CFRP reinforcement was designed for flexural strengthening of the deficient positive moment regions. Strengthening of the negative moment regions was not considered (small deficiencies). Two strengthening systems were considered for flexural upgrade, FRP pre-cured laminates and FRP fabric applied using the wet-lay-up procedure, as shown in Figure 3. The laminates are 2 in. wide and 0.05 in. thick with an ultimate strength (ffu) of 406 ksi, and modulus of elasticity (Ef) of 23,900 ksi. The CFRP sheets are 0.0065 in. thick with an ultimate strength (ffu) of 550 ksi, and elastic modulus (Ef) of 33,000 ksi. Summary of the design requirements is given in Table 2.

Table 2. CFRP design summary

Location (slab direction/span)	No. of CFRP Strips	Width of two-layer CFRP sheets (in.)	Length (ft)*
E-W, typical 27 ft bay	6	48	14
N-S, typical 35 ft bay	8	72	22

*Centered on mid-span

In-Situ Load Testing

The scope of load testing included studying the performance of unstrengthened slab sections, the influence of the Gunite beams that were properly bonded, the performance of the slab after the Gunite beam is removed, and slab behavior after strengthening with surface bonded FRP composites. It was decided that seven load tests were required to achieve these goals. These test included testing one 27-ft span with a Gunite beam, two 27-ft spans with Gunite beams removed, two 27-ft span strengthened with CFRP composites (after the Gunite beams has been removed), and two 35-ft spans without strengthening. All tests were done on the slab of the garage second level.

A typical in-situ load-testing method consists of applying uniformly distributed loads to the structure in the form of water, sand, etc. These techniques do not provide adequate information for the engineer to evaluate the stability and response the structural components. In addition, these methods carry the risk of possible catastrophic failure during testing. An efficient, cyclic and economical load testing procedure was proposed to provide a non-destructive yet conclusive demonstration of the performance of structural components (CIAS 2000). The load testing procedure utilizes hydraulic jacks placed at strategic locations to induce the internal forces equivalent to those resulting from distributed loads. This method allows for better control of the applied load and maintains a strict control on safety as the loads can be removed very quickly. In comparison with typical load testing methods, the proposed method was faster, more economical, safer, and more conclusive.

The load testing equipment used consisted of two 30-ton hydraulic cylinders and a hydraulic pump for applying the load, eight linearly variable differential transducers (LVDTs) for measuring deflections, and a load cell for measuring the applied load (see Figure 4). The load tests were “push-type” tests, in which the jacks react against the two floors above, using the floors’ dead weight as a counterweight (see Figure 5). Post shores, each with a minimum axial capacity of 10,000 lbs., were installed above the one being tested to engage the two floors above. Also, two shoring towers with a minimum axial capacity of 40,000 lbs. per tower were placed on the level below the one being tested to support the slab in the remote possibility of failure. Deflections are acquired by measuring the relative displacement of the slab in reference to the floor below. The LVDT layout is shown in Figure 5. Strain gauges were also used to measure the strain in FRP reinforcement at each level of applied load. Figure 6a shows a hydraulic jack and load cell used during testing. Figures 6b and 6c show an LVDT and a strain gage, respectively, used during the tests. For each load test, the intended maximum load is attained through six load cycles. Each load cycle consists of a minimum of 4, approximately equal, load steps (see figure 7), followed by at least 2 steps to unload the structure. Each load step is maintained for approximately 2 minutes. During this time the mid-span deflection of the structure is monitored for stability. If the deflection begins to increase with a constant or dropping load, the system has past the elastic threshold and the test is halted. The peak load for each successive cycle is gradually increased to finally equal the maximum test load.

The magnitude of the applied loads were calibrated to create forces in the slab mid-span similar to those due to the uniformly applied design loads load. This included the dead load of the structure and a 30 psf live load uniformly applied over the entire garage bays. The load tests, however, were based on the application of two point loads. These point loads were calibrated to create forces in the slab (throughout the span) similar to those resulting from a uniformly applied load. Although every effort was made to minimize the differences between the two load conditions, it was impossible to create the same loading condition. Therefore, the applied test loads were designed to be conservative. Additionally, the behavior of the garage slab was continually monitored during the tests to ensure that the desired response is achieved.

Test Results

In Test 1 (27 ft span with Gunite beam), a total of 10 load cycles where applied to the span. The span was loaded to the service load level in the first six cycles. In the following four cycles, the maximum load was increased to achieve 85% of the ultimate load condition. The slab performed satisfactorily based on all above-described criteria, even at the ultimate load levels. Figure 8 shows the load deflection curve for Test 1. The deflection shown is measured at mid span, and the load is the recorded load at the jacks. Only 7 cycles were applied for Test 2 (27-ft span, Gunite beam removed). Again, the span was loaded to threshold load level and then pushed to 85% of ultimate. The slab performed well through all 6 cycles to the threshold load level. On the seventh cycle, in an attempt to reach the ultimate load level, the span became inelastic and the test was halted. Figure 9 shows load deflection curve for test 2.

Due to the non-linear behavior of the slab at Test 2, the 27-ft span at Test 3 (27-ft span, Gunite beam removed) was only tested to the threshold load levels. It performed satisfactorily. Tests 4 and 5 (35-ft spans, unstrengthened) had 8 load cycles, first loading to threshold level and then to ultimate. The slabs performed well through all 8 cycles and to the ultimate loads. The tests show that the CFRP strengthening is not needed. Figure 10 shows the load-deflection envelope at mid span (ultimate load level) for the two tests. Tests 6 and 7 were performed on 27-ft span slabs, tested in Tests 2 and 3, after strengthening. Tests 6 and 7 had the standard six load cycles with the maximum load applied corresponding to 85% of the ultimate level. Both slab spans performed satisfactorily. Figure 11 shows load deflection curves at mid span for Test 6.

Discussion of Test Results

As expected, 27-ft slab spans with Gunite beams were stiffer than spans without Gunite beams, with or without CFRP. Figure 12 shows a comparison of test results (at maximum loads) for Tests 1, 2, and 3. As shown before, the 27-ft span with Gunite beam performed well at ultimate load levels. The span without Gunite beam that was loaded to near the ultimate load level did not. The other span without Gunite beam performed well, but it was not loaded to the same load level.

Figure 13 shows a comparison of load-deflection envelopes for 27-ft spans with Gunite beams (Test 1) and without Gunite beams but strengthened with CFRP (Tests 6 and 7). As shown, all three tests performed well at the same load levels (curves are “tight” and predominantly linear). Note the difference in stiffness (slope of the curves) between the three tests. Based on the tests results, the CFRP strengthening of 27-ft spans where delamination of Gunite beams was found was warranted and it performed satisfactorily.

Figures 14 and 15 show comparisons of load-deflection envelopes for two 27-ft spans tested without Gunite beams, before and after CFRP application. The curves on Figure 14 (Tests 2 and 6) show that, even though the unstrengthened slab had cracked and behaved nonlinearly near ultimate loads, once the CFRP was placed, the same span performed satisfactorily.

SUMMARY AND CONCLUSIONS

The load tests showed that strengthening of E-W garage spans with delaminated Gunite beams is warranted; these spans did not perform satisfactorily. Conversely, the 27-ft spans with Gunite beams that are not delaminated were found to perform well. The load tests further showed that the 27-ft slab spans repaired with CFRP performed well. There were no significant differences between the performances of spans strengthened with two different CFRP systems. Based on the test results, additional nineteen 27-ft slab spans, where delamination of Gunite beams was evident, were strengthened with CFRP. Further monitoring of the performance of Gunite beams is required, and additional CFRP strengthening will be done periodically. The performance of the existing typical spans in the N-S direction (35 ft spans) was found to be acceptable without any repairs. Structural deficiencies of the typical second-level end bays were repaired using additional steel framing supports. The steel framing is used instead of CFRP due to the fact that larger structural deficiencies were found in the end bays than acceptable for CFRP repair. Large deflections of cantilevered end bays were mitigated with insertion of additional steel supports.

REFERENCES

CIAS (2000), Guidelines for Rapid Load Testing of Concrete Structural Members, Appraisal Report, Concrete Innovations Appraisal Services, Report # 00-1, ACI International, P.O. Box 9094, Farmington Hills, MI 48333-9094.

ACI 440 (2000), (Draft). Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures. American Concrete Institute, P.O. Box 9094, Farmington Hills, MI 48333.