Printer Friendly

Fatigue testing for thermex reinforcing bars.

Fatigue Testing for Thermex Reinforcing Bars


Conventional grade 60 reinforcing steel that meets the American Association of State Highway and Transportation Officials (AASHTO) M31-89 (ASTM A615-87) specifications is generally manufactured by alloying certain percentages of rare metals, such as manganese and vanadium, with steel ingots to obtain the required strength. These rare metals are available mostly from foreign sources; 64 percent of the known world reserve of vanadium ore is in South Africa.

Because of the unavailability of these rare metals locally and their cost, the Florida Steel Corporation in 1987 decided to produce grade 60 reinforcing bars by a proprietary Thermex process which reduces the requirements of these alloying elements. This process, developed in eastern Europe, is similar to the "Tempcore" process that has been in use in western Europe since 1974.

Thermex processing combines conventional hot-rolling with a water quenching and a self-tempering heat treatment. The reinforcing bar's exterior is cooled rapidly to a low temperature, facilitating the formation of a harder outer layer that approaches a martensitic condition, while the rebar interior is of lower strength and still quite ductile.

Although the Thermex process produces reinforcing steel that meets the AASHTO M31-89 specifications for grade 60 reinforcing steel, detailed research showed that Thermex steel has different physical properties from conventional reinforcing steel containing alloying elements. (1) Consequently, before it would permit largescale use of reinforcing steel in Federal-aid projects, the Federal Highway Administration (FHWA) investigated the fatigue characteristics of Thermex versus conventional reinforcing steel. The findings of this investigation are described in this article.


Currently, the AASHTO material specification for reinforcing steel (M31-89 grade 60) has no provision for requiring rebars to meet certain fatigue requirements. Moreover, there is no standard test method - or agreement on best method - of performing fatigue test on rebars in the U.S. To develop the AASHTO design specifications, testing was performed on rebars encased in concrete beams loaded in flexture (2). This process simulates the conditions experienced by rebars in service but is impractical for quality control testing of the bars. More recent work has been performed by the firm Wiss, Janney, Elstner, Associates, Inc. (WJE), where rebars were tested in air in a servo-hydraulic load frame. (3) The WJE method correlated fairly well with the concrete encased tests, however, the air test results were very sensitive to the method used for gripping specimens.

Unlike the United States, the British do have a material standard that requires rebars to meet specified fatigue performance criteria. (4) This standard reflects work performed at the Transport and Road Research Laboratory (TRRL) in the United Kingdom, in which specimens were axially tested in air. (5) The British specification requires rebars to survive 5 million load cycles at a prescribed stress range that is dependent on the diameter of the bar being tested. Based on the experience at TRRL and WJE, it was decided to test bars in air for the present study.

Description of Specimens

To resolve the fatigue question, the FHWA requested that the Florida Department of Transportation supply appropriate samples of both Thermex and conventional reinforcing steel to perform fatigue tests at the Turner-Fairbank Highway Research Center (TFHRC) laboratories.

Initially, the TFHRC received 24 rebar samples from the Florida Steel Corporation for testing. These consisted of four Thermex and four conventional bars in three different bar sizes - #5, #6, and #11. The bars were all 460 mm (18 in) long. The Thermex bars were designated by the letter "T" on the bar. These 24 original specimens were used in resolving the "grip problem" discussed below. At a later date, additional samples of #5 and #11 rebars were shipped by the Florida Steel Corporation mill in Knoxville, Tennessee, for further testing.

The manufacturer provided chemical analysis data for both the Thermex and conventional bars. The FHWA obtained a chemical analysis on the same bars by an independent testing laboratory. Table 1 shows the chemical analysis data for both bar types. All residual elements were excluded from the table because they have little effect on tensile properties.

Table : Table 1. - Chemical analysis of Thermex and conventional rebar samples
Process Bar Source C Mn V
 size size (%) (%) (%)
 #5 Manufacturer 0.28 0.70 0.002
Independent 0.29 0.63 <0.005
#11 Manufacturer 0.31 0.82 0.003
Independent 0.41 0.97 <0.005
Conventional #5 Manufacturer 0.43 0.71 0.002
Independent 0.44 0.75 <0.005
#11 Manufacturer 0.34 1.01 0.028
Independent 0.36 0.99 0.033

Comparing the Thermex to the conventional bars showed no significant difference in manganese content; however, the carbon content was about 35 percent lower in the #5 Thermex bars than in the conventional bars. The concentration of vanadium, which is only added to conventional bars larger than the #6 size, was much lower in the Thermex than the conventional #11 bars. The percent of vanadium was negligible in the #5 bars produced by both processes. This last finding substantiates the manufacturer's claim that the addition of vanadium is not required in Thermex-produced bars larger than #6 in order to obtain the tensile strength required by the AASHTO M31-89 grade 60 specification.

Fatigue Testing

All fatigue tests were performed on bars mounted axially in a servo-hydraulic load frame. Cyclic loading was applied at a rate of 20 Hz for the #5 and #6 bars; however, this had to be reduced to 10 Hz for the #11 bars because of the higher loads required. These frequencies resulted in rapid testing of the specimens, with most tests lasting less than 24 hours.

Initially, the rebars were tested at a stress range of 138 MPa (20 ksi), with the minimum to maximum stress ratio of R = 0.2. However, based on discussion with researchers from the United Kingdom's Transport and Research Laboratory, the test stress ranges used in Britain were later adopted, resulting in stress ranges of 200, 185, 150 MPa (29, 26.8, and 21.8 ksi) for the #5, #6, and #11 bars, respectively. (4) Tests were also conducted at a stress ratio of R = 0.2.

Gripping Methods

Previous research has shown that axial tests of bars in air are very sensitive to the way they are gripped in the test machine. (3,5) To eliminate the gripping effects from the test results, several methods of gripping specimens were investigated. Many difficulties arise when trying to test reinforcing bars in a servo-hydraulic load frame. Unlike standard test specimens, the bar diameter cannot be reduced in the center section before testing. The deformations rolled onto the bar surface have an effect on fatigue life, and must remain on the bar during testing. Additionally, the Thermex process produces a hardened surface layer that also must remain. Simply put, the grip problem is that the bar has the same section inside and outside the grip area. Any force exerted by the grips will result in locally higher stress concentrations - which in turn will result in a high probability of failure at the grips.

Special wedge grips filled with babbitt (a lead-based alloy containing 1 to 10 percent tin and 10 to 15 percent antimony) were used for rebar fatigue testing conducted in the 1960's. More recently, in the 1988 National Cooperative Highway Research Project study, Fatigue Behavior of Welded and Mechanical Splices in Reinforcing Steel, WJE developed a special hollowed grip that used a high early strength grout to hold the bar. (3) During fatigue testing, the grout "clamps" the bar deformations in the grip by direct bond between the grout and bar. WJE had a high success rate of fatigue failures away from the grips in this National Cooperative Highway Research Program (NCHRP) pilot test program. The current FHWA test program tried to find a simplified method of bar gripping, thereby speeding up the test procedure. Four types of grip systems were investigated at the TFHRC to try to get failure to occur away from the grip area. The following are brief descriptions of these grip systems:

1. Aluminum or brass shims placed between the bar and the self-aligning, V-type wedge grips. These shims deformed around the bar's raised deformations and prevented the wedge grips from marring bar surface. This method was ineffective in preventing grip failure. Figure 1 shows a top view of a typical grip failure with aluminum shims.

2. Machining raised deformations off the bar ends, then inserting bars in the wedge grips lined with aluminum shims. It was hoped that eliminating the stress concentration effects caused by the raised deformations would prevent grip failure. This method was effective for conventional type reinforcement; however, grip failure was still evident with the Thermex bars.

3. Embedding bar ends in soft (babbitt) metal contained within conical grips. The bar was placed in the body of a standard 12.7-mm (0.5-in) prestressing chuck, and molten metal was poured around the bar. The prestressing chuck was then gripped in the self-aligning grips of the hydraulic load frame. Figure 2 shows a sectional view of this grip system. Both pure lead and babbitt metal were tried in this system. The lead proved to be too soft and fatigued before the bar; the babbitt metal was in-effective in preventing grip failure.

4. Embedding bar ends in high strength metallic grout contained within the same conical grips. This system is essentially the same as the WJE grout system mentioned above. Out of the four systems tested, the grouted conical grips were the most effective in preventing grip failure. Figures 3 and 4 show a typical failure occurring about 1 bar diameter outside of a grouted grip. The size of the prestressing chuck body limited testing with this system to #5 and smaller bars. Larger, specially made grips would be required for larger bars.



Table 2 shows the results of all the fatigue tests performed, indicating the grip type and location of failure. The failures noted as occurring outside the grip area (i.e., at least 1 bar diameter away from the grips) were considered valid. For the conventional bars, the machined-end method was very effective, with valid failures occurring in 75 percent of the tests. For the Thermex bars, the grout and babbitt embedding methods were the only ones providing valid breaks. Of these two methods, the grout appeared to be the most effective, with both tests showing valid failures. The grout method should also work well for the conventional bars, although no tests were performed for verification.

It is apparent that the gripping medium is critical to the validity of tests conducted on reinforcing bars. In service, while some of the axial force is transferred to the rebar by adhesion, most of the axial force is transmitted through direct bearing between the concrete and the raised deformations along the bar. This circumstance results in a shear force at the point where the raised deformation joins the bar's round section. The magnitude of this force is determined by the compressive strength of the grout or other gripping medium. If the shear force introduced by the grip system is too great, large reductions in fatigue strength will result. This effect seems more critical in Thermex than conventional bars, probably because of the hardened surface layer on the Thermex bars.

Figure 5 shows the results of all the tests performed at the TFHRC, with different symbols indicating the types of grips used. The filled symbols indicate "valid breaks" - i.e., the bar broke at least 1 bar diameter outside the grip area. The machined-end and shim-grip systems seem to produce failures earlier than the grout- or babbitt-grip systems. Although data are limited, the grout system seems to produce a higher percentage of valid failures than the babbitt system. For this reason, grouted wedge-type grips should probably be specified if a standard test method for reinforcing steel is developed, even though they are more labor intensive than the other types investigated.

Thermex Versus Conventional Bars

Statistical findings

Figure 6 shows the fatigue test results of the three conventional and three Thermex #5 bars shown in table 2 that are considered valid. These tests all failed at points at least 1 bar diameter outside the grip area. As a reference, the solid and dotted lines in the figure represent, respectively, the lower 95-percent confidence limit and mean regression lines for #5 bars. (4) The Thermex bars all fell above the mean line while the conventional bars fell slightly below the mean line. All of the bars failed well above the lower 95-percent confidence limit. [Tabular Data Omitted]

Given the small sample size and the fact that fatigue tests usually show a large degree of scatter, no statistical conclusions can be drawn from these data. However, the data do tend to support conclusions from testing performed by WJE - i.e., that there does not appear to be any significant difference in fatigue resistance between Thermex and conventional bars. (6)

Comparison with British standard

It should be noted that there is a large discrepancy between these test results and the British standard. For example, for #5 bars, the British standard requires testing at a stress range of 200 MPa (29 ksi). (4) The NCHRP mean line indicates shown in figure 6 indicates the average life of bars to be only about 700,000 cycles at 200 MPa (29 ksi) stress range. It could be argued that the NCHRP mean line was derived primarily from tests on bars embedded in reinforced beams loaded in flexure while the British standard was derived from tests of bars in air. The in-air tests performed by both WJE and the FHWA are relatively close to the NCHRP mean line, indicating that no significant difference exists between the in-air and reinforced-beam tests. Thus, the fatigue strength of U.S. bars seems relatively insensitive to test type. The large difference between the U.S. tests and the British standard remains unexplained.


This preliminary research developed insufficient data to make any conclusions regarding the relative fatigue resistance of bars produced by the Thermex process. It does, however, point up the need to develop a U.S. specification for testing reinforcement bars. What constitutes a valid test needs to be resolved and a level of acceptable performance specified. The discrepancy between U.S. test results and the British standard also should be investigated to see if U.S.-produced bars actually show lower fatigue resistance than those produced in Europe.


(1) John L. Ratliff. "Evaluation of Thermex Reinforcing Steel," unpublished report to Florida Steel Corporation, Applied Research and Development Co., Tampa, FL.

(2) T. Helgason, J.M. Hanson, N.F. Somes, W.G. Corley, and E. Hognestad. "Fatigue Strength of High Yield Reinforcing Bars," NCHRP Report No. 164, Transportation Research Board, Washington, DC, 1976.

(3) C. Paulson, and J.M. Hanson. "Fatigue Behavior of Welded and Mechanical Splices in Reinforcing Steel," Interim Report for NCHRP Project 10-35, Northbrook, Illinois, October, 1988.

(4) British Specification BS 4449 for Weldable Steel Bars for Reinforcement of Concrete, British Standard Institution, London, 1988.

(5) D.S. Moss. Axial Fatigue of High Yield Reinforcing Bars in Air, Supplementary Report 622, Transport and Road Research Laboratory, Crowthorne, Berkshire, United Kingdom, 1980.

(6) Letter from Conrad Paulson; Wiss, Janney, Elstner, Inc. Northbrook, to James Oliver, Florida Steel Corporation, Tampa, regarding fatigue tests of reinforcing bar, February 18, 1991.

Yash Paul Virmani is a research chemist in the Structures Division, Office of Engineering and Highway Operations Research and Development, Federal Highway Administration (FHWA). He serves as the program manager for the Nationally Coordinated Program on Corrosion Protection and the High Priority Program Area Prestressed Concrete Protection. Dr. Virmani is the co-inventor of conductive polymer concrete, a material that is the basis of several cathodic protection systems.

William Wright is a research structural engineer in the Structures Division of the FHWA's Office of Engineering and Highway Operations Research and Development. He is the principal investigator for several research projects involving fatigue and fracture of steel bridges being performed in the Structures Laboratory at the Turner-Fairbank Highway Research Center.

Ronald Nelson is a structural laboratory technician in the Structures Division of the FHWA's Office of Engineering and Highway Operations Research and Development. Mr. Nelson has 25 years experience in design, set-up, and performance of structural and material tests at the Turner-Fairbank Highway Research Center

PHOTO : Figure 1. - Top view of a typical grip failure with aluminum shims.

PHOTO : Figure 2. - Sectional view of grouted conical grip used for testing #5 bars.

PHOTO : Figure 3. - Hydraulic load frame showing grouted conical grips mounted in self-aligning wedge grips.

PHOTO : Figure 4. - Closeup of a typical failure in the grouted conical grip.

PHOTO : Figure 5. - Fatigue results showing the effect of grip type.

PHOTO : Figure 6. - Fatigue results for #5 bars comparing the Thermex and conventional processes.
COPYRIGHT 1991 Superintendent of Documents
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Virmani, Yash Paul; Wright, William (English priest); Nelson, Ronald N.
Publication:Public Roads
Date:Dec 1, 1991
Previous Article:Analysis of advisory speed setting criteria.
Next Article:Laboratory evaluation of Verglimit and PlusRide.

Related Articles
Impact fatigue of a polycarbonate/acrylonitrile-butadiene-styrene blend.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters