Due to their ability to provide robustness during handling and installation, as well as long-term durability, steel fibers have become an accepted form of reinforcment for segmental tunnel linings.

Regardless of the properties steel fiber reinforced (SFRC) concrete is chosen for, designers must prepare specifications that direct precast manufacturers in the selection of the the proper fiber type and dosage rate.

During mix development, precast manufacturers use specifications to form a mix that will meet compressive, flexural and bursting strength requirements. Once these parameters are confirmed, the type and quantity of steel fibers is chosen to provide the specified equivalent flexural strength (EFS) value. Confirmation of the EFS value is achieved by conducting bending tests on beams in accordance with standard test methods.

Flexural resistance

Flexural resistance properties are determined using beam tests, and evaluating the load carrying capacity of a beam after the concrete cracks. There are many publications that describe how to determine SFRC flexural properties, but the methodology involves casting small sample beams and testing them in a 3 or 4 point displacement controlled bending apparatus.

Sample sizes vary slightly between test methods, but the generally accepted size of test beams are 150x150mm square cross section with a test span of at least 3 times the depth. The use of smaller beams tends to exaggerate test results, due to size effects and preferential alignment of fibers in the molds and should not be used.

Test methods such as JSCE-SF4, ASTM C-1609, and EN-14489-1 are typically specified based on the country in which the project is located. In general, load deflection curves are produced for the sample beams and values are determined for the post crack flexural strength, usually using data up to a deflection of span/150 or 3mm. The post crack flexural strength is a measurement of the strength provided by the steel fibers.

Design engineers should be aware that each test method provides slightly different performance values. Understaning this is critical when preparing acceptance criteria for steel fibres. After all, it is vital the material performance used in design is in fact specified in the contract documents and used during manufacturing.

Unfortunately, there is currently no clear consensus on how best to specify acceptance criteria.

Figures 1 through 4 show load deflection curves for two different fibers with different acceptance criteria shown. The red and green lines represent an average of five individual tests from manufacturer Red and manufacturer Green. Both Red and Green fibers are advertised as being identical and the same dosage was used in all tests. A value of 3.2Mpa (500psi) is used as the design value for the ultimate flexural capacity of the segments. These figures illustrate acceptance criteria that have been used on actual projects in the past few years.

Figure 1 shows a single point as the acceptance value. Both Red and Green fibers meet this acceptance criteria. The deflection value occurs at 0.5mm, which is very early in the test and is still in what is considered the unstable portion of the load deflection curve.

None of the test methods referenced herein recommend using single point values at deflections less than span/600, which is 0.75mm. Furthermore stress values at span/600 should only be used for serviceability checks and not for ultimate strength determination.

Figure 2 shows toughness performance classes as the acceptance criteria. The design value represents an average value for Level IV performance class. This criteria is somewhat subjective since it is not clear if the entire fiber performance line must exceed the Level IV line or can drop below the line for some percentage. For this criteria Red fibers are acceptable while Green fibers would require a higher dosage to meet the acceptance criteria. Since this criteria uses an average of stress levels to span/ 150 it would be appropriate for ultimate strength determination.

Figure 3 shows two-point acceptance criteria. This criteria is specified using ASTM C1609. It is similar to the toughness performance classes and may use the end points of one level. This criteria is less ambiguous since it is clear acceptable fiber performance must exceed both points. The average value of the two points is used as the design value. Therefore, Red fibers are acceptable and Green fibers would require a higher dosage rate. Since this criteria uses an average of stress levels to span/ 150 it would be appropriate for ultimate strength determination.

Figure 4 shows acceptance criteria using an average value calculated with the area under the load deflection curve. This criteria is specified using JSCE SF-4. An average value is used as the design value. Flexural toughness (Tb) is determined as the area under load deflection curve up to deflection of span/ 150, which is 3mm (dtb) for a 150mm (b) x 150mm (t) x 450mm (L), or 6”x 6”x18”, square beam.

Flexural toughness factor, expressed as an engineering stress, is calculated using:

See Equation 1

[Mpa] (psi)

where (dtb) = 3 mm

Using this criteria, the Red fibers are acceptable and Green fibers would require a higher dosage. This criteria uses an average of stress levels to span/ 150 and would be appropriate for ultimate strength determination.

As can be seen from the above discussion, different conclusions are reached depending on the criteria chosen. Listed in Table 1 are three published SFRC performance level criteria. Although these are designed for steel fibre shotcrete they work equally well for cast concrete.

It is recommended to use one of the methods in Table 1, an equivalent flexural strength value such as an Re3 using an average of the area under the load deflection curve up to 3mm, or an average of the f150,0.75 and f150.3.0 values. In this way, all of the post crack load deflection data, until the end of the test at 3mm deflection, is used to evaluate acceptance of the steel fiber concrete.

Beam tests represent average equivalent or residual flexural strength values. Appropriate statistical methods should be applied to ensure that the specified SFRC mix provides the minimum equivalent or residual flexural strength design values.

The following is an example of specification language in a contract document, using ASTM C1609, to ensure a project is supplied with consistent steel fiber performance:

“Steel fiber reinforced concrete shall provide a post-crack average equivalent flexural strength toughness, of xxx psi, minimum, which represents the average of the f150,0.75 and f150,3.0 values obtained according to ASTM C1609.”

Bursting resistance

Bursting resistance, or splitting tensile strength, of SFRC is less well understood than flexural properties. Bursting in segments occurs from two different types of loads. Bursting stresses are created during installation by the application of ram loads to the edge of the segments and from the in-place forces due to compression in the ring (Figures 5 and 6). The magnitude of the bursting force is related to the configuration of the segments, the TBM and the geology.

It is well known the presence of steel fibers in concrete do not significantly influence its unconfined compressive strength or it’s modulus of rupture in flexure. However, it is not well known that the presence of steel fibers increases the bursting strength of concrete. This is one of the primary benefits in using SFRC in segments.

The forces created by the applied loads are redistributed by the fibers as micro tensile cracks develop. This allows larger loads to be carried by SFRC than can be carried by plain concrete. This is confirmed by large-scale tests and small cylinder splitting tests.

Tests have shown that increases of 10-30% over plain concrete bursting strengths have been achieved with steel fiber dosages of 30-40kg/m3 (50-66lb/yd3). Obviously, the magnitude of the increase is related to the steel fiber type and dosage rate. Another significant observation of full-scale bursting tests is that the failure mode with SFRC is more ductile than with conventionally reinforced segments. Full-scale test have also shown that SFRC segments can provide similar bursting load capacities as those with conventional reinforcement.

Since the concrete used to manufacture segments is usually high-strength, designers may question whether the formulas contained in the structural concrete codes in Europe and US are appropriate for high-strength concrete. A recent paper evaluated the splitting tensile strength formulas for a number of sources for application with concretes up to 120Mpa (17,400psi)[4].

The results of their integral absolute error analysis showed that the CEB-FIB code formula gives a good correlation to observed data. In contrast, the formula in ACI-318-99 did not have a very good correlation, particularly at higher compressive strengths where there is a significant divergence between the two methods. Table 2 shows splitting tensile strength values calculated using the two methods.

The CEB-FIB formula for splitting tensile strength is:

See Equation 2 (metric units)

See Equation 3 (imperial units)

Experience on past projects has shown that when the steel fiber contents are determined for flexural resistance requirements then the splitting tensile resistance requirements are also usually met. Design engineers should specify the required value for splitting tensile needed to carry the applied loads.

Test methods used to determine splitting tensile strength include ASTM C496, BS 1881- Part 117 for cast cylinders or the “Brazilian Method” for cores obtained from finished segments. Since there is no standardized test method for preparing cylinders for splitting tensile tests it may be prudent to place in the specification some language to indicate cylinders for splitting tensile strength should be consolidated using external vibration not by rodding the sample. Rodding will disturb the random orientation of the steel fibers, which may affect results.

All fibers are not equal

Steel fibers have been successfully used for over 20 years to reinforce concrete segments. However, contractors should be wary of selecting a supplier based simply on price, as there are many low-quality look-a-likes currently available on the market. Many suppliers do not manufacture their own fibers. Distributors purchase from manufacturers and place their label on the bag. Often these distributors change suppliers and never inform their customers.

As an example of quality issues that have been identified for steel fiber suppliers, examine the data in the Figures 7 and 8. Figure 7 shows the results of tensile tests conducted on ten individual fibers from ten different suppliers taken over a seven-year span. Most were sampled once but one was sampled four times in one year. The published tensile specification was the same for all of the suppliers: An average of 1300Mpa ±100. The results are startling. Only one test was within the published specification, 11 tests had minimum values significantly outside of specification, and only three were close to the published average value.

Contrast this with the data in Figure 8, which was collected over a one year time span for a single supplier. The results show a consistent pattern of tensile strengths within the published specification.

Fiber quality can be determined by consistency of length, diameter, tensile strength and hook formation. In order to ensure consistent quality it is recommended the following language be placed in the contract documents:

“Test steel fiber in accordance with ASTM A 820 or EN-14489-1. Test 10 fibers per lot.

Determine steel fiber length, diameter and aspect ratio. Individual samples shall be ±10% of specified values and the average of 10 samples shall not deviate more than ±5% of the specified values.

Determine tensile strength of steel fiber. No more than 10% of the individual samples shall be below the specified values.”

In reference to manufacturer, or supplier experience, the following language is also recommended:

“If the supplier of the steel fibers is not the manufacturer, the distributor shall document that they have a minimum 5 years experience in distributing steel fibers from the same source.

If steel fibers are supplied from different manufacturing facilities the supplier shall provide documentation that material being produced meets the same quality at all locations.”

Conclusions

It is imperative that engineers specify acceptance criteria in alignment with assumptions made during design, otherwise the delivered segment performance may not meet expectations.

The recommendations offered herein aim to provide clear and concise guidance for the preparation of specifications. When implemented, steel fibers will be supplied from a consistent quality manufacturer and provide the flexural resistance and bursting strength needed to match the design.


Fig 1 – Single point acceptance criteria Fig 1 Fig 2 – Toughness performance level acceptance criteria Fig 2 Fig 3 – Average two-point acceptance criteria Fig 3 Fig 4 – Fe3.0 acceptance criteria Fig 4 Fig 5 – Ring bursting Fig 5 Fig 6 – Ram bursting Fig 6 Fig 7 – Tensile strengths of various suppliers (specification is 1300Mpa ±100) Fig 7 Fig 8 – Tensile strengths of one supplier (specification is 1260Mpa ±190) Fig 8 Beam test underway Beam test underway Table 1 Table 2 Equation 1 Equation 1 Equation 2 Equation 2 Equation 3 Equation 3