Hybrid steel fibre-glass fibre rebar segment design Feb 2020

In recent years, the adoption of fiber reinforced concrete (FRC) segments has increased worldwide. The solution of FRC segments, without any reinforcement rebar, provides advantages in terms of cost and precast production operations. Nevertheless, in some parts of a segmentally lined tunnel drive, particular loading conditions, typically under prevalent bending actions such as in cross-passage breakout or under particularly shallow alignment, the FRC solution can not satisfy the requirement.

Fig 1. Results of the beam bending tests for SFRC segments

The possibility of adopting glass fiber reinforced polymers (GFRP) reinforcement rebar in concrete segments can be proposed as an alternative to the traditional steel rebar when a high resistance to environmental attack is required since GFRP reinforcement does not suffer corrosion problems and its durability performance is a function of its constituent parts^(1,2). From the mechanical point of view, GFRP rebar is characterised by an elastic behaviour in tension, and, with respect to the steel rebar cages, present higher tensile capacity, lower elastic modulus, and lower weight⁽³⁾. Compression strength is often neglected, due to its low value. GFRP is also electrically and magnetically non-conductive, but sensitive to fatigue and creep rupture⁽⁴⁾. Furthermore, the structural effects of the low elastic modulus and bond behavior has to be considered⁽⁵⁾.

Due to all these aspects, this type of reinforcement is not suitable for all applications, but it appears appropriate for precast segments, both for provisional and permanent applications.

Fig 2. GFRP cage for the hybrid SFRC-GFRP test segments

To evaluate the potential applications, FRC segments of a typical metro tunnel geometry, were cast with and without GFRP rebar and experimentally tested. Both bending and point load tests were carried out and the structural performance results, both in terms of strength and crack width, are compared and discussed.

Four 1.4m wide x 300mm thick steel fibre reinforced test segments were produced for a 6.4m o.d. ring. Bekaert Dramix 4D 80/60BG steel fibre was added to the concrete matrix with a content of 40kg/m³. The average compressive strength, measured on six 150mm cubes, was equal to 62.35 MPa.

The tensile behavior was characterized through bending tests on eight 150mm x 150mm x 600mm notched specimens according to the EN 14651 code. Nominal stress versus the crack mouth opening displacements (CMOD) were plotted (Fig 2).

Fig 3. Bending test set-up and instrumentation

Fig 4. Comparison of bending test load average displacement

Table 1. Maximum crack widths comparison

Fig 5. Bending test crack pattern at the intrados surface
Load [kN]		125	160	180	210	222	250	270
Crack width
	SFRC	<0.05	0.25	0.35	0.60	1.00	n/a**	n/a**
	SFRC-GFRP	<0.05	0.10	0.15	n/a**	0.35	0.45	0.70

n/a*= measure not available since the crack width was not recorded at this load step

n/a**= measure not available since the segment did not reach this load value

The two hybrid SFRC-GFRP segments were further reinforced with a perimetric GFRP cage (Fig 3). The GFRP rebar had a nominal diameter of 18mm and was characterized by Young’s Modulus of about 40 GPa and ultimate tensile strength equal to 1000 MPa.

Both segment types were then subjected to full-scale flexural and point load tests for evaluation of the structural performances, both in terms of structural capacity and crack pattern evolution, under bending and under the TBM thrust conditions.

Bending tests

The bending tests were performed by adopting a 1000kN electromechanical jacket, with a PID control and by imposing a stroke speed of 10 µm/sec (Fig 4). The segments were placed on cylindrical supports with a span of 2,000mm and the load, applied at the midspan, was transversally distributed be adopting a steel beam and measured by three potentiometer wires and two LVDTs.

The behaviour of the SFRC and SFRC-GFRP segments are compared where the average value of the displacement, measured by the three potentiometer wires, is plotted versus the load (Fig 5). The first cracks appeared for a load value of about 125kN and 120kN for the SFRC and SFRC-GFRP segments respectively. In both the cases the first cracks were opened on the lateral surfaces close to the midspan and propagate on the intrados.

After a first comparable almost elastic response, the SFRC-GFRP segment presented a peak load about 63% higher than the SFRC segment, 367kN as compared to 225kN (Fig 6). The obtained results, including the maximum crack widths measured at different load steps (Table 1), show clearly the synergic effects of the two materials in reducing the crack widths.

Fig 6. Point load test and instrumentation

Fig 7. Point load test where load (single pad) vs time; a) SFRC; b) SFRC-GFRP segment

Fig 8. Point load test crack pattern for each segment type

Table 2. Comparison of maximum crack widths
Load		Maximum crack width [mm]
		SFRC	SFRC -GFRP
1st crack [kN]	1250	0.05	<0.05
Service load [kN]	1580	0.10	0.05
Unblocking thrust*[kN]	2670	0.40	0.25
Unload [kN]	0	0.15	0.10

Note * For metro tunnelling, TBM pushing capacity coincides with unblocking thrust

Point load tests

The point load test was performed by applying three-point loads on each segment and adopting the same steel plates as used by a TBM, each jack having a loading capacity of 2,000kN. The load was continuously measured by pressure transducers with six potentiometer transducers (three located at the intrados and three at the extrados) measuring the vertical displacements, and two LVDTs transducers, applied between the load pads, measuring the crack openings (Fig 7).

Two cycles were performed, with the chosen reference load levels equal to 1,580kN and 2,670kN (for each pad) referring to the service load and unblocking thrust of the TBM (Fig 8). The final crack pattern, after the point load test, is similar for both the SFRC and SFRC-GFRP segments (Fig 9). The first cracks appeared for a load level of 1250kN (for each steel pad) between two pads at the top and lateral surfaces in both the cases. Besides the splitting cracks (between the pads), a bursting crack (under the point load), formed in both the cases. The results show that the addition of the perimetric GFRP cage led to halve the crack width under the service load, and to reduce the crack width by about 37.5% under the unblocking thrust force (Table 2). Furthermore, a reduction of the crack width of about 33% was measured after the complete unloading.

As a conclusion to the test cycle, the results of the bending tests show clearly the synergic effects of the two materials (steel fibre and GFRP reinforcement) by increasing the peak load and reducing the crack width. The results of the point load tests confirm the effectiveness of the solution, since the addition of the perimetric GFRP cage led to halve the crack width under the service load, and to reduce it under the unblocking thrust force, and at the complete unloading, respectively.

While the tests provide positive results, there is yet to be an actual application of a hybrid SFRC-GFRP segmental lining.