Journal of Materials in Civil Engineering

A sulfur‐recovery process from phosphogypsum produces a by‐product slag that can potentially be used as aggregate in portland cement concrete. The phase composition and microstructure of samples of slag from a bench‐scale (Florida Institute for Phosphate Research [FIPR]) and a pilot‐scale (campaign 5) run of the process were characterized. Both slags have a frothy appearance. The FIPR slag consists of elongate, and often radiating crystals of gehlenite $\left({\mathrm{Ca}}_2\mathrm{Al}\right(\mathrm{Si}\text{,}\mathrm{Al}\left)O_7\right)$, enclosing larnite $(\beta ‐{\mathrm{Ca}}_2{\mathrm{SiO}}_4)$. Some amounts of iron sulfide and magnetite are also present. Campaign 5 slag does not contain gehlenite since quartz is present. Its microstructure is also different, consisting of an intimate, heterogeneous mixture of calcium silicates and opaque phases. Residual gypsum is present in both slags but in greater amounts in campaign 5; the latter also contains some ettringite and native sulfur. The FIPR bench‐scale run produced a completely burned residue while the pilot‐scale run residue was underburned. However, mineralogically, both are suitable as aggregate in concrete as long as the amount of sulfate phases is not too high (<4.5%).

ASTM A615 grade 40 ordinary deformed‐steel reinforcing bars with a specified minimum yield strength of 276 MPa (40 ksi) and ASTM A722 high‐strength prestressing thread bars with a specified ultimate strength of 1,083 MPa (157 ksi) were experimentally evaluated for their low‐cycle fatigue behavior under axial‐strain‐controlled reversed cyclic tests with strain amplitudes ranging from yield to 6%. All tests were performed on virgin (unmachined) specimens to closely simulate seismic behavior in structural concrete members. A methodology is suggested to identify incipient (first‐cracking) failure of test specimens. The experimental data were evaluated with existing fatigue models, which related stress‐strain quantities to the failure life. Additional energy‐based fatigue models are proposed that relate various stress and/or strain quantities to the dissipated energy. The study demonstrates that the modulus of toughness and low‐cycle fatigue life for both the low‐ and high‐strength materials are similar. Based on fatigue considerations, it is concluded that existing design codes are overly restrictive in not permitting the use of high‐strength thread bars in seismic‐resisting elements.

Experimental results of previous research showed that the bond strength between structural steel shapes and concrete in composite sections at the age of one year is approximately 30% of that at the age of three weeks. A literature survey reveals that no research has been conducted to investigate bond strength in composite construction at ages over one year. This paper reports the results of tests conducted on 15 battened composite specimens to investigate bond strength between steel and concrete at the age of five years. Test results of this investigation show that the bond strength at the age of five years is more than twice that at the age of 360 days. The reduction in bond strength in specimens of five years of age is significantly less than that in specimens 360 days old, when compared with results of 21‐day‐old ones.

In this paper, two analytical approaches based on the effective modulus method (EMM) and the age‐adjusted effective modulus method (AEMM), are proposed for the prediction of deflections due to creep and shrinkage in reinforced‐concrete beams with a steel‐fiber concrete matrix under sustained loads. In the EMM‐based approach, the deflection due to creep is evaluated from an elastic analysis using a modulus of elasticity that accounts for the reduced creep coefficient in steel‐fiber concrete, while the deflection due to shrinkage is analyzed using the equivalent tensile‐force method, which accounts for the reduced shrinkage strains in the steel‐fiber concrete. The AEMM‐based approach considers the aging of steel‐fiber concrete and involves a simultaneous linear‐elastic analysis for deflections due to creep and shrinkage using an age‐adjusted effective modulus of elasticity. Tests on 12 SFRC beams as well as on the material properties were carried out. The proposed methods, using creep coefficients and shrinkage strains obtained from compressive creep test and free‐shrinkage test on steel‐fiber concrete, respectively, are found to predict the creep and shrinkage deflections with very good accuracy.

The vertical lift fixed‐wheel gate is a common type of control gate found at hydroelectric generating stations. The material used for the wheels of such gates ranges from cast iron for the older vintage stations to alloy and low‐carbon steels in more modern stations. In recent years, several cast‐iron wheels on such gates have failed in service. An experimental program, sponsored by Manitoba Hydro, was undertaken at the University of Manitoba to study the performance characteristics of cast‐iron wheels and to determine their failure capacity under static loading conditions. Eleven 685‐mm‐diameter cast‐iron wheels obtained from a spillway gate at the McArthur Falls Generating Station in Manitoba were tested to failure under radial and a combination of radial and lateral loads. The material properties were obtained from standard coupon testing. In addition to the experimental work, the elastic behavior of cast‐iron gate wheels was investigated using the ANSYS finite‐element (FE) program. The results from the experimental program and the finite‐element analysis were compared to the results predicted by current design criteria and a number of recommendations for improving these criteria were made.

In this paper the flexural behavior of a strain‐hardening engineered cementitious composite (ECC) is studied and compared with that observed in a regular fiber‐reinforced cementitious composite (FRC). Unlike concrete or regular FRC, ECC materials are characterized by their ability to sustain higher levels of loading after first cracking while undergoing additional straining. This strain‐hardening behavior gives ECCs a significant advantage under flexural loading. In quasi‐brittle material such as regular FRC, the ratio of flexural strength (the modulus of rupture) to tensile strength is known to vary between 1 and 3, depending on the details of the reinforcement and the geometry of the specimen. In this paper, the strain‐hardening behavior observed in an ECC led to a high flexural‐strength‐to‐tensile‐strength ratio. In a third‐point bending test, the flexural strength of an ECC was measured to be five times its tensile (first‐cracking) strength. This result was also predicted by a simple theoretical model. The model can be used for the purpose of optimizing the flexural strength of ECCs.

The first‐crack strength and flexural toughness of steel‐fiber‐reinforced concrete specimens with different dimensions were determined using the procedures outlined in ASTM C1018 and JSCE‐SF4 and the ASTM toughness indices and JSCE toughness parameters were calculated. It was found that variations in specimen size influenced not only the stress and deflection at first crack, and the ultimate strength, but also all the toughness parameters. The ASTM toughness indices and the JSCE toughness factor were both dependent on specimen dimensions unless the specimens were geometrically similar. It was found that toughness parameters decreased in an increase in the span‐to‐depth ratio of the specimens. All toughness parameters were significantly affected by the width of the specimen, even when both depth and span were unchanged; the toughness increased with an increase in width.

A theory founded on a suitable failure criterion for concrete and on a constitutive law in uniaxial tension derived from the failure criterion is developed for axially loaded confined concrete columns. The influence of the tension‐softening behavior of the external concrete shell on the compressive strength of columns is considered, and curves are derived describing the failure axial stress as a function of the tensile strain in the cover. A critical value of the transverse reinforcement percentage is found, for which axial stress and load always increase in the total range of tensile strain. Simple expressions for axial stress and strain and for the corresponding compressive strengths are provided for three particular failure limit states. The experimental comparisons show that for static or quasi‐static and for dynamic loading, failure occurs when the tensile peak or ultimate strain values are reached in the cover, while yielding of confining steel is even reached but only in the case of low transverse reinforcement percentages.

A new method of studying the dynamic development of microcracks in hardened cement paste (HCP) plates subjected to dynamic load sources was developed by Avnon and Yankelevsky in 1991. This method is used to study the dynamic response of high‐strength‐concrete (HSC) two‐dimensional samples. The dynamic loading is produced by special cylindrical explosive charges containing about 10 mg PETN each. When initiated, these charges produce a dynamic pressure pulse for about 3 ms. The charges are inserted inside boreholes perpendicular to the plate faces in predefined locations. The dynamic loading causes cracking to the plates, enabling the measurement of the crack‐propagation velocity and to investigate the crack‐interaction phenomenon. Crack‐propagation velocities were found to be in the range of 60–400 m/s, much lower than the theoretical values. The new method enables simple measurements of crack‐propagation velocities in HSC samples of different shapes and dimensions under dynamic tensile stresses, including small samples, in which crushing of concrete is observed in common methods during similar tests.

Recent research into the strength of dimension lumber has led to the development of an alternative failure criterion based on the notion of a critical energy density for the material. The strength of wood is known to be related to the duration of an applied load; however, duration‐of‐load effects cannot be explained within the context of a theoretical failure criterion based solely on stress. Conversely, a theoretical relationship between ultimate stress and load duration can be demonstrated when an energy‐based failure criterion is combined with a viscoelastic stress‐strain model. In the present paper, the theoretical equations expressing the relationship between ultimate stress and load duration are derived for constant intensity loads. Two related energy‐based failure models are proposed and investigated. Using available experimental data as a reference, a numerical experiment is conducted to compare the two energy models with a widely used damage accumulation model.

Wood fiber reinforced cement composites provide the highest performance‐to‐cost ratio among fibrous cement composites being considered for the replacement of asbestos cement. Fairly strong and stiff wood fibers are particularly suited for the reinforcement of thin‐sheet cement products. There is, however, concern regarding the long‐term performance of wood fiber reinforced cement composites exposed to repeated freezing‐thawing conditions. An experimental study was undertaken in order to investigate the performance of wood fiber reinforced cement composites containing 2% mass fraction of kraft and mechanical pulps. Comprehensive replicated test data were generated for various test cycles, and were statistically analyzed using the analysis of variance and multiple comparison techniques to derive reliable conclusions. The results of this investigation showed, at a 95% level of confidence, that non‐air‐entrained wood fiber reinforced cement composites performed desirably under repeated freeze‐thaw cycles, while plain cementitious matrices were susceptible to the freezing and thawing environment.