A study of coarse grain heat affected zone of accelerated cooled structural steels

Abstract

Modern structural steels have significantly different (generally leaner) compositions than equivalent conventional steels developed over 20 years ago. To compensate for the lower carbon and alloy content more sophisticated thermomechanical treatments have been introduced to give very fine ferrite grain size and hence good strength, ductility and fracture toughness. Despite the fact that these changes were largely introduced to improve weldability, the modern structural steels have given problems of heat affected zone (HAZ) hydrogen induced coldcracking leading to considerable debate over the causes of this apparently increased susceptibility. The present work aimed to provide the basis for a better understanding of the metallurgical behaviour of the HAZ of these new modern steels. This was accomplished by employing three low carbon(C<0.15%) low sulphur (S<0.005%) content steels produced by the OLAC process which combines controlled rolling with accelerated cooling. A normalized low carbon (0.12%C) higher sulphur (0.031%S) steel was also included in the present project for comparison purposes. The starting point was to determine the transformation temperatures of the various microconstituents in the HAZ’s by Tn Situ’ thermal analysis. A new approach was developed in which thermal data obtained from real weld thermal cycle measurements were employed in the thermal analysis. The steels studied were analyzed in the light of the results of transformation temperatures, microstructural examinations and hardness measurements in the coarse grained region of the HAZ. This approach offers much greater accuracy and consistency than previous methods of determining thermal cycles and HAZ transformation temperatures, giving a better opportunity for studying the anomalous HAZ behaviour of modern steels. Between the OLAC steels, two steels presented very low hardness(<300HV5) within the entire range of heat inputs used (0.8-3.5kJ/mm),whereas the other gave hardnesses higher than those of the normalized steel. Higher carbon (0.13%C) and a poorly balanced alloy content was considered to be an explanation for the HAZ behaviour of this steel. An investigation on the accuracy of current carbon equivalent in predicting hardenability and hardness formulae was carried out but none of the existing formulae were completely satisfactory in indicating the trend in hardenability found in all the steels. An alternative formula for predicting the variation of hardness based on one put forward by Yurioka was developed, which proved to be suitable for three of the steels studied. However, the restricted nature of all empirical carbon equivalent formulae is demonstrated and the need to limit any such formula to a restricted range of steels is re-emphasized.