Experimental Program

5.2.1 Beam Details

A total of four prestressed concrete beams (PC beam) were taken out for tests, consisted of two beams post-tensioned type (herein after abbreviated as PC-O) and two beams pretensioned type (PC-R). All beams have the identical cross-section of 150 x 300 mm and length of 2400 mm, but bar arrangements and cover depth are varied according to the type beam designated. Cover depth of 30 mm for PC-O beams and 35 mm for PC-R beams. Cross-section of the beams is described in Fig. 5.1.

Fig. 5.1 Cross-section of PC beams

The materials for making concrete are high early strength Portland cement. The properties of the aggregate are presented in Table 5.1 and mix proportion of concrete was designed with water/cement ratio (w/c) and sand/total aggregate ratio (s/a) of 40.7% and 37.0%, respectively.

The composition of material in unit weight (kg/m³) for water, cement, sand and gravel was 167, 410, 640 and 1175 respectively (5.3), (5.5)

. The concrete was designed with a slump target of 5 ± 1 cm and air content about 3 ± 1%.

Prestressing tendons were round wires of 2Ø2.9 mm diameter for PC-R beams, and round bars of 17 mm diameter for PC-O beams. Yield strengths were 1148 and 1795 MPa for the wire

a. Bonded post-tensioned beam (PC-O) b. Pre-tensioned beam (PC-R)

77

and bar respectively. Deformed bars of 10 mm diameter in the PC beams were embedded as stirrups with spacing of 100 mm.

Table 5.1 Summary of aggregates (5.3), (5.5)

Aggregate Specific Gravity Fineness Modulus

Fine-river sand 2.25 2.84

Coarse-crushed stone 2.75 6.63

After placing of concrete, the beams were moisture cured for one day and demoulded.

Then PC beams were steam cured (maximum temperature 60^oC) for about 10 hours, followed by curing in air until the start of exposure. Prestress was introduced by the pre-tensioning method for PC-R beam and post-tensioning method for PC-O beam. Effective prestress was about 12.7 to 13.2 MPa for PC-R and 13.7 MPa for PC-O at the bottom fiber of the beam immediately after its introduction. After post tensioning, cement paste was grouted between the bars and the sheath. Also, the ends of each beam were covered with a pad of cement mortar to prevent corrosion of the ends of strands or end anchorages.

5.2.2 Exposure Condition

All beams have been exposed to real marine environments at the Sakata Port for 20 years, which is located in north-west of Japan (38^o56’N, 139^o47’E) and faces the Sea of Japan. The annual average temperature is around 11.9^oC. The minimum temperature between December and March reaches below zero mostly every day, which could cause freezing and thawing action. Furthermore in winter time, the daily maximum wind velocity is more than 25 m/s which produces much splashing.

The beams were placed in the tidal just in front of a caisson-type quay wall two months after placing of concrete, where the beams have been subjected to wet and dry conditions alternately due to tidal action.

Table 5.2. Summary of PC beam specimens ^(5.5)

Type Cover Depth

(mm)

Cracking and Loading Condition

Effective Prestress (MPa) PC-O-1

30 B

13.7

PC-O-2 B

PC-R-1

35 B

12.7-13.2

PC-R-2 A

78

The beams can be categorized into two conditions in term of pre-cracking, i.e. ‘A’ and

‘B’. Condition A refers to pre-cracked, for which beams were pre-cracked by bending moment until 65% of the ultimate bending moment. After cracking, the bending moment was released, and no continuous load was applied during exposure. Condition B refers to no pre-cracks.

Summary of the beam specimens were presented in Table 5.2.

After 20 years of exposure, the beams were transferred to the laboratory, stored at a constant temperature and sheltered from the rain over 15 years as shown in Photo 5.1. Then beams were taken out for laboratory tests. All tests were conducted in conformity with JIS, JSCE and ASTM.

Photo 5.1 PC beam specimen

5.2.3 Test Item and Method a. Material testing of beam

In order to quantify the degree of deterioration of concrete materials, compressive strength, static modulus of elasticity, chloride content, pore volume and pore size distribution, carbonation depth, monitoring of the rebar condition and external appearance of the beams were observed. First, after cleaning all surface of specimen, the existence of rust stain, air bubble, chips, etc. was observed. For pre-cracked specimens, those cracks were checked and the crack width was measured. It was also observed that a new crack has occurred or not. Moreover, all specimens were photographed in detail.

Then specimens were investigated for corrosion of steel bar by using half-cell potential and anodic polarization curve by the contact method. The half-cell potential was measured at an interval of 200 mm at four strands with various cover depth, stirrups and steel bar. Whereas

79

polarization curve was only measured at three points (20 cm from both ends and in the center of specimen) in the four-strands, stirrups and steel bar. The measurement was conducted with the silver/silver chloride electrode after 1 hour of pre-wetting.

To obtain the mechanical properties of the specimens, ultrasonic pulse velocity (UPV) was conducted by the direct method with an interval of 200 mm and path length of 150 mm.

After completion of the ultimate load test, several cores of 50 mm in diameter were taken from each beam in areas that was predicted un-cracked during bending test for compressive strength, carbonation depth, porosity and chloride content measurements. Chloride content was measured at certain depths of core samples, taken from two points of beam. Then porosity of the mortar matrix was measured on a depth of 10 - 20 mm of core samples. In addition, carbonation depth was also evaluated on the whole surface of beam and freshly cut surface of core samples after spraying a 1% phenolphthalein solution.

Finally, beam specimens were crushed by using a hammer and concrete drill, and then all strands and steel bar were removed for further observation. Firstly, the photograph of steel bar was taken, followed by observation of changes in stirrups shape after bending test and situation that describe the propagation of rust were carried out. In addition, the ratio of portion surface area that has been rusted corresponding to all surface area of steel bar was measured.

b. Bending test

Bending test as described in Fig. 5.2 was carried out in a three-point loading with a span of 2100 mm. All PC beams were arranged in the testing machine in the same manner. The instrumentation layout was also similar for all PC beams with the exception of the strain gauging at crack location. Each beam was positioned in the 1800 KN capacity universal test machine. Load was applied to the specimens through the load cell and a full width distribution plate as shown in Photo 5.2.

Fig. 5.2 Test configuration

150 350 350 350 350 350 350 150

150 700 700 700 150

5010010050

LVDT LVDT LVDT LVDT LVDT

Load cell

Strain gauge

150

300

30 50 5050

100

Note: Attack  gauge (wide crack gauge) on the bottom of beam Mechanism: 1. Loading until crack, then release.

2. Attack strain gauge on the bottom of crack than loading again until failure

80 5.2.4 Test Procedure and Instrumentation

Each beam was tested in three separate phases for bending test. In the first phase of testing, load was applied to create and locate a series of flexural cracks to instrument with strain gauges, crack gauges and displacement transducers (linear variable differential transformer or LVDT). The second phase was used to determine the decompression load and first cracking load in each beam, based on strain and displacement measurements of crack openings for the crack identified and instrumented in the first phase of testing. Phase 3 of each test involved the loading until the ultimate failure.

Photo 5.2 Test setup

The beams were loaded to cause flexural cracking, and the location of these cracks be marked, so they could be located and instrumented after load was removed from the beams.

Before loading, the beams were instrumented with five 60 mm strain gauges. Three gauges were placed on one side to measure the strain distribution through the depth of the member and two gauges were placed on the compression side of member. Then three 3 mm  gauges were mounted on the bottom side to measure the crack width. In addition, five 50 mm LVDT were mounted on the bottom surface of the beam with a distance of 350 mm to measure the deflection at the specified distance. And also a sensor connected to the AE system mounted at the surface of beams to detect internal crack and prestressing strands failure.

At the location of the first crack detected, strain gauge and/or  gauge were instrumented.

The strain gauges were mounted on the surface of the concrete beam where crack was identified and measured decompression strain in the extreme fiber of the specimen as load was applied.

81

Measurements of crack opening with LVDT were also used to determine the decompression load. In a similar manner to the strain gauge method, the decompression load was determined by studying the load vs. crack opening measurement plot. This plot typically showed an increasing amount of crack opening displacement per unit load after crack began to open. The load that corresponds to this change in displacement rate was taken as the decompression load ^(5.2).

The 5 kN inspection interval was typically used when increasing to any new load level to allow for visual inspection for cracks in the bottom side near mid span. Initial cracking was predicted at a load of approximately 90 and 85 kN for PC-O and PC-R beams, respectively. The load was then increased until a change in curve linearity of load vs. displacement appeared and held constant while all cracks identified and marked.

After the initial crack was detected, the beams were completely unloaded and prepare for the decompression load and ultimate load test. Overall, a total of two or three load cycle, with periodic pauses to inspect for cracks to ensure that all cracks was identified in the tensile region.