Assessment of cast steel anchorage fracture toughness of a cable-stayed bridge by small punch test

Depto. de Ingeniería Mecatrónica. Instituto Tecnológico de Celaya, Celaya, Guanajuato, México. E-mail: alejandro.alcaraz@itcelaya.edu.mx Depto. Ciencia e Ingeniería del Terreno. E. T. S. de Ingenieros de Caminos, Canales y Puertos. Universidad de Cantabria, Cantabria, España Depto. de Ingeniería Mecánica. Instituto Tecnológico de Celaya, Celaya, Guanajuato, México Instituto Mexicano del Transporte. Querétaro, México © Universidad De La Salle Bajío (México) Assessment of cast steel anchorage fracture toughness of a cable-stayed bridge by small punch test Evaluación de la tenacidad a la fractura de anclajes de aceros colados de un puente atirantado por ensayo miniatura de punzonamiento


Resumen
La evaluación de la vida residual de componentes estructurales en servicio requiere conocer el valor de la tenacidad a la fractura, pero los métodos convencionales para medir la tenacidad demandan remover grandes cantidades de material del componente, lo cual es generalmente impráctico. Sin embargo, el Ensayo Miniatura de Punzonamiento (EMP) (que utiliza especímenes miniatura no estandarizados) ha sido empleado como una alternativa práctica y conveniente para evaluar las características de fractura del material de componentes en servicio. El propósito de esta investigación fue encontrar una relación entre la deformación a la fractura equivalente de EMP εaf y la tenacidad a la fractura JIC de aceros colados de baja aleación procedentes de anclajes de un puente atirantado localizado en el Golfo de México. La tenacidad a la fractura JIC se calculó a partir de los datos experimentales reportados de KIC en un trabajo previo y la deformación a la fractura equivalente se obtuvo mediante el EMP empleando especímenes de 10x10 mm 2 por 0.5 mm de espesor. A partir de los resultados de εaf y JIC obtenidos, y correspondientes datos de aceros de baja aleación de la literatura, una correlación lineal fue propuesta para estimar la tenacidad de fractura a partir de la deformación a la fractura equivalente del EMP para los aceros colados pertenecientes a este caso de estudio.

Introduction
Fracture toughness is one of the most important material properties for assessing the structural integrity (Webster, 2000) or investigate the causes of component failures (Urriolagoitia, 2012;Delgado, 1998). In the case of components in service, a sample of material must be removed to evaluate fracture toughness which could endanger the integrity of that component; additional concerns arise when the critical testing region is so small that specimens cannot be obtained with the minimum size requirements of the standard test methods. Consequently, test methodologies have been developed focusing on sub-sized specimens that allow significant material data to be derived from a small quantity of sample material.
In the last decades, the Small Punch Test (SPT) has proved to be a promising testing technique in assessing mechanical properties by using reduced size specimens. The SPT was initially used in mechanical property characterization of irradiated material within the nuclear industry (Manahan, 1981;Mao 1991) and later was introduced as a quasinondestructive method for evaluating local mechanical properties in service structural elements with large dimensions (Fleury, 1998;Viswanathan, 1994;Lacalle, 2008;Madia, 2013;Guan, 2011;Dogan, 2012;Cárdenas 2012  During the test, the force-displacement curves are registered. From these curves, tensile mechanical properties have been successfully estimated (Fleury, 1998;Rodríguez, 2009;Ruan, 2002) along with fracture toughness properties , Mao 1991. Several researchers Guan, 2011;Misawa, 1989;Wang, 2008) have reported that there exists a linear relationship between the fracture toughness parameter JIC and the SPT data, which can be expressed by Eq. 1: Where k and J0 are material constants, whereas εaf is the equivalent fracture strain given by Eq. (2): In the last equation, t0 and tf are the initial and final thickness of the fracture specimen, respectively; see Although good correlations between fracture toughness parameters and SPT data have been found, values of the constants k and J0 constants are distinctive for each material (García, 2015). In the case of structural steels, most of SPT studies have focused on wrought steels, while only a few investigations have concentrated on cast steels, regardless of the brittle nature of these steels.
Therefore, more extensive investigation under the SPT technique is needed for cast steels.
The cast steel investigated in this work corresponds to the cable anchorages of a cablestayed bridge that contains 112 cable anchorages; see Figure 3. One of the cable anchorages failed during normal operation in 2000 after five years of service. Since then, several scientific and engineering studies have been realized for assessment of its structural integrity. López et al. (2009) presented an analysis using ultrasonic and liquid penetrant techniques, and the results revealed the presence of several micro-structural defects, such as pores, cracks, inclusion and large grain size, Based on an ultrasonic study (López, 2009), 16 anchorages in critical condition (large number of microstructural defects) were removed from the bridge along with four anchorages considered in "good condition" to study several properties. Two of them were used in the present research to evaluate whether the SPT could be a technique to determine the fracture toughness of the remaining in-service cast steel anchorages on the bridge.

Previous work: Mechanical characterization by standard methods (Alcaraz, 2012)
The materials under study are cast steels from two different cable anchorages (see Figure 3), where the chemical composition, microstructural details, and tensile and fracture toughness properties were determined according to current standards. The chemical composition was obtained by applying the optical spark emission technique under the ASTM E1019 and ASTM E415 standards.
The microstructure characterization was performed by optical microscopy after specimen polishing and chemical attack with Nital-3%. A total of 9 rectangular specimens from anchorage 1 and 34 rectangular specimens from anchorage 2 were prepared for tensile testing according to the ASTM E8 standard. In addition, plane-strain fracture toughness KIC was determined following the guidelines of the ASTM E399 standard; SE(B) three-point bend specimens were fabricated with a maximum thickness allowed by the limited dimensions of the respective anchorages. Thus, from anchorage 1, two 52-mm-thick specimens were obtained, whereas one specimen of 60 mm thickness was fabricated from anchorage 2.

Small punch technique characterization
As mentioned, a widely used correlation for assessing the fracture toughness of metallic materials by means of the small punch technique is given by Eq.
(1). This expression, which relates the fracture parameter JIC with the equivalent fracture strain εqf, has proved to give good approximations. In fact, fitting coefficients k and J0 for several materials have been reported by some authors Wang, 2008;Guan, 2011). In the present work, the fitting parameters mentioned above have been used for the fracture toughness estimation of the cast steels under study, and they are shown in Table 1.   = • ̅ − [Wang, 2008] = .

Model Reference
• ̅ + . [Guan, 2011] Moreover, additional k and J0 values were also obtained by using (εqf, JIC) values from anchorage 2 and corresponding low-alloy steel values from literature (Guan, 2011), see Table 2, where the main alloying element was Chromium (Cr). It is worth mentioning that (εqf, JIC) values from weld steels were not considered from such reference because the steel under study is a cast steel without welding. The εqf value from anchorage 2 was calculated by using Eq.
(2), while JIC was obtained by means of Equation (3), where E is Young's modulus and represents Poisson's ratio of the material.
The small punch test specimens considered were 10×10 mm 2 squares of 0.5 mm thickness. Eigth test pieces were prepared from anchorage 1 and eigth samples from anchorage 2. The testing

Mechanical characterization by standard methods
Chemical analysis results from previous work of the two cable anchorage steels are shown in Table   3. Based on their alloying elements, the materials were classified as low-alloy cast steels, where anchorage 1 is named 1Cr-0.5Ni, and anchorage 2 as 0.8Cr-0.6Ni. Figure 4 presents metallographic examinations from the two anchorages. In both cast steels, a characteristic ferrite-pearlite microstructure is observed; however, some differences can be distinguished concerning the phase distribution, possibly due to a deficient heat treatment (López, 2009). Also, some sulfides and pores are also observed ( Figure 5), as well as macroscopic discontinuities ( Figure 6).    The tensile mechanical properties obtained from previous work are shown in Table 4. It is observed that the steel from anchorage 1 has a yield and ultimate strength higher than those of anchorage 2; however, the ductility is higher for the cast steel from anchorage 2, as evidenced by the elongation percent. Figure 7 shows stress-strain curves for the two cast steels for comparison purposes. The stress-strain curve from anchorage 1 is typical of high-resistance steel, whereas the corresponding curve for anchorage 2 is representative of high-ductility steels with extensive plastic strain prior to fracture.  The fracture toughness testing results according to ASTM E399 standard are shown in Table 5. By comparing fracture toughness KIC values for both anchorages, anchorage 2 presents a higher value than anchorage 1. The difference can be explained by the high ductility exhibited by anchorage 2 with respect to anchorage 1; see Table 4.

Small Punch Test Characterization
Load-displacement curves obtained from small punch testing for both anchorages are shown in Figure 8. These curves display a characteristic ductile behavior for both steels (Lacalle, 2012 capable distinction between the two types of cast steels; nevertheless, the maximum load values slightly fluctuate. This last behavior is linked to local strain concentrations arising before the maximum load is reached, leading to material instabilities (Lacalle, 2012).
The failure morphology displayed by the small punch specimens for both anchorages showed a hemispherical surface, and fracture occurred along the circumference where the strain is highest; see Figure 9. This failure mode is typical of ductile steels.  To analyze the SPT fracture toughness results, the equivalent fracture strain εqf is used. This parameter is calculated from a modified version of Eq.
(2) , which is given as Where d* is the displacement at fracture, which is obtained from the load-displacement curves for the fractured specimens. Thus, d* and εqf values for the specimens from anchorages 1 and 2 are shown in Table 6. By observing the displacement values d* from both anchorages, it is then clear that the cast steel from anchorage 2 shows higher ductility, as previously concluded from Figure 8 and Table 4.   (Wang, 2008;Guan, 2011)

Conclusions
This work involved a comparative analysis of fracture toughness results from standard specimen tests and from small punch rests for low-alloy Cr-Ni cast steels coming from the anchorages of a small punch specimens showed ductile fracture, whereas standard samples showed quasi-cleavage fracture mode.