High strain-rate compressive strain of welded 300W asteel joints

Abstract

The split Hopkinson pressure bar (SHPB) test is the most commonly used method for determining material properties at high rates of strain. The theory governing the specifics of Hopkinson bar testing has been around for decades; however, it has only been for the last decade or so that significant data processing advancements have been made. It is the intent of this thesis to offer the insight of application of SHPB to determine the compressive dynamic behaviour for welded low carbon steel (mild steel). It also focuses on the tensile behaviour for unheat-treated and heat-treated welded carbon steel. The split Hopkinson Pressure bar apparatus consists of two long slender bars that sandwich a short cylindrical specimen between them. By striking the end of a bar, a compressive stress wave is generated that immediately begins to traverse towards the specimen. Upon arrival at the specimen, the wave partially reflects back towards the impact end. The remainder of the wave transmits through the specimen and into the second bar, causing irreversible plastic deformation in the specimen. It is shown that the reflected and transmitted waves are proportional to the specimen's strain rate and stress, respectively. Specimen strain can be determined by integrating the strain rate. By monitoring the strains in the two bars and the specimen's material, stress-strain properties can be calculated. Several factors influence the accuracy of the results, including the size and type of the data logger, impedance mismatch of the bars with the specimens, the utilization of the appropriate strain gauges and the strain amplifier properties, among others. A particular area of advancement is a new technique to determine the wave's velocity in the specimen with respect to change in medium and mechanical properties, and hence increasing the range of application of SHPB. It is shown that by choosing specimen dimensions based on their impedance, the transmitted stress signal-to-noise ratio can be improved. An in depth discussion of realistic expectations of strain gages is presented, along with closed form solutions validating any claims. The thesis concludes with an analysis of experimental and predicted results. Several recommendations and conclusions are made with regard to the results obtained and areas of improvement are suggested in order to achieve accurate and more meaningful results.