The problems with the crack length determination by the unloading compliance method are well known for Charpy size specimens. The final crack lengths calculated for bent specimens do not fulfil ASTM 1820 accuracy requirements. Therefore some investigations have been performed to resolve this problem. In those studies it was considered that measured compliance should be corrected for various factors, but satisfying results were not attained. In the presented work the problem was attacked from the other side, the measured specimen compliance was taken as a correct value and what had to be adjusted was the calculation procedure. On the basis of experimentally obtained compliances of bent specimens and optically measured crack lengths the investigation was carried out. Finally, a calculation procedure enabling accurate crack length calculation up to 5mm of plastic deflection was developed. Applying the new procedure, out of investigated 238 measured crack lengths, more than 80% of the values fulfilled the ASTM 1820 accuracy requirements, while presently used procedure provided only about 30% of valid results. The newly proposed procedure can be also prospectively used in modified form for the specimens of different than Charpy size.

The problems with the crack length determination by the unloading compliance method are well known for Charpy size specimens. The final crack lengths calculated for bent specimens do not fulfil ASTM 1820 accuracy requirements. Therefore some investigations have been performed to resolve this problem. In those studies it was considered that measured compliance should be corrected for various factors, but satisfying results were not attained. In the presented work the problem was attacked from the other side, the measured specimen compliance was taken as a correct value and what had to be adjusted was the calculation procedure. On the basis of experimentally obtained compliances of bent specimens and optically measured crack lengths the investigation was carried out. Finally, a calculation procedure enabling accurate crack length calculation up to 5mm of plastic deflection was developed. Applying the new procedure, out of investigated 238 measured crack lengths, more than 80% of the values fulfilled the ASTM 1820 accuracy requirements, while presently used procedure provided only about 30% of valid results. The newly proposed procedure can be also prospectively used in modified form for the specimens of different than Charpy size.

Neste trabalho os resultados da metodologia de normalização de E. D. Reese e K. H. Schwalbe para a obtenção de curvas J-R, são comparadas com aquelas obtidas pela tradicional Técnica de Variação de Flexibilidade Elástica. Ambos os métodos foram aplicados à um aço de grau nuclear A508-CL3A, quando vários corpos-de-prova miniaturizados 0.394TC[S], foram testados em temperaturas típicas de operação de vasos de pressão nuclear. Corpos-de-prova miniaturizados formam tratados termicamente tentando simular o comportamento exibido pelos aços estruturais quando submetidos à diferentes níveis de exposição de nêutrons, sendo que outros foram ensaiados na condição de como-recebidos. Todos os procedimentos experimentais e analíticos empregados neste estudo foram totalmente computadorizados, incluindo aquisição e refinamento dos dados e o traçado da curva J-R. Esta última foi realizada conforme a mais recente versão das normas ASTM, denominada ASTM-E1820. As variáveis avaliadas influentes do comportamento das curvas de J-R, incluem o incremento de propagação de trinca e o número de pontos de dados de carga versus deslocamento. Também foi analisada a efetividade do método de normalização quando comparada com a técnica de variação de flexibilidade elástica. Os resultados desta pesquisa mostraram que a normalização é uma metodologia confiável e muito simples, tão logo a programação de software seja desenvolvida, para ser aplicada em uma vasta variação de comportamentos mecânicos como amostrados neste estudo. / In this research, results from Reese; Schwalbe (1993) linear normalization methodology for deriving J-R curves are compared to those obtained from the traditional unloading elastic compliance technique. Both methodologies were applied to a nuclear grade steel A508-CL3A. Several miniaturized 0.394TC[S] specimens, were tested at typical NPV operating temperatures. Small scale testpieces were heat treated aiming to simulate the embritlemment exhibited by the structural steel when submitted to different levels of neutron exposure, whereas other specimens were tested in the as-received condition. All experimental and analytical procedures employed in this study have been fully computerized, including data acquisition, data refinement and J-R curve plotting, when the guidelines supplied by most recent version of ASTM standard for J-R curves, namely ASTM-E1820, were followed. The assessed variables affecting J-R curves behavior comprised crack propagation increment, number of load-Ioad line displacement data points, and it was also analyzed the agreement of the normalization method when compared to the compliance technique. Research results have shown that linear normalization is a trustworthy and very simple methodology to be applied in a vast range of mechanical behavior as sampled in this study, as long as computer programming is employed.

Neste trabalho os resultados da metodologia de normalização de E. D. Reese e K. H. Schwalbe para a obtenção de curvas J-R, são comparadas com aquelas obtidas pela tradicional Técnica de Variação de Flexibilidade Elástica. Ambos os métodos foram aplicados à um aço de grau nuclear A508-CL3A, quando vários corpos-de-prova miniaturizados 0.394TC[S], foram testados em temperaturas típicas de operação de vasos de pressão nuclear. Corpos-de-prova miniaturizados formam tratados termicamente tentando simular o comportamento exibido pelos aços estruturais quando submetidos à diferentes níveis de exposição de nêutrons, sendo que outros foram ensaiados na condição de como-recebidos. Todos os procedimentos experimentais e analíticos empregados neste estudo foram totalmente computadorizados, incluindo aquisição e refinamento dos dados e o traçado da curva J-R. Esta última foi realizada conforme a mais recente versão das normas ASTM, denominada ASTM-E1820. As variáveis avaliadas influentes do comportamento das curvas de J-R, incluem o incremento de propagação de trinca e o número de pontos de dados de carga versus deslocamento. Também foi analisada a efetividade do método de normalização quando comparada com a técnica de variação de flexibilidade elástica. Os resultados desta pesquisa mostraram que a normalização é uma metodologia confiável e muito simples, tão logo a programação de software seja desenvolvida, para ser aplicada em uma vasta variação de comportamentos mecânicos como amostrados neste estudo. / In this research, results from Reese; Schwalbe (1993) linear normalization methodology for deriving J-R curves are compared to those obtained from the traditional unloading elastic compliance technique. Both methodologies were applied to a nuclear grade steel A508-CL3A. Several miniaturized 0.394TC[S] specimens, were tested at typical NPV operating temperatures. Small scale testpieces were heat treated aiming to simulate the embritlemment exhibited by the structural steel when submitted to different levels of neutron exposure, whereas other specimens were tested in the as-received condition. All experimental and analytical procedures employed in this study have been fully computerized, including data acquisition, data refinement and J-R curve plotting, when the guidelines supplied by most recent version of ASTM standard for J-R curves, namely ASTM-E1820, were followed. The assessed variables affecting J-R curves behavior comprised crack propagation increment, number of load-Ioad line displacement data points, and it was also analyzed the agreement of the normalization method when compared to the compliance technique. Research results have shown that linear normalization is a trustworthy and very simple methodology to be applied in a vast range of mechanical behavior as sampled in this study, as long as computer programming is employed.

A study is carried out into fracture behavior of steel/steel laminates both experimentally and through finite element analysis (FEM). The laminates produced by hot pressing consisted of low carbon and medium carbon steels with two volume fractions / 0.41 and 0.81. Fracture toughness, JIC has been measured using partial unloading technique assuming a critical value of crack extension. The technique is initially applied to monolithic material and then to the laminates in crack divider orientation. Evaluation of fracture toughness of laminates indicates that there is a substantial improvement of JIC with increase in the volume fraction. The systems under study were also evaluated by FEM modeling with the use MARC package program. To evaluate JIC, the problem has been evaluated in several steps / first two-dimensional plane strain problem is considered. This is followed by three-dimensional case and then by an artificially layered system, all for monolithic materials. Values of JIC derived were close to one another in all cases. Following this verification, the method, as implemented in layered monolithic system, was applied to laminates. This has shown that JIC of laminates can be predicted using FEM analysis, including the delamination. Values of JIC varied in the same manner as the experiment verifying that fracture toughness in the current system increases with increase in volume fraction. It has been concluded that modeling as implemented in this work can be used for useful composite systems incorporating hard/brittle reinforcements both in crack divider and crack arrester orientation.