Compressive behaviour of engineered cementitious composites (ECC) under high strain rates

In the past few decades, an interest in exploring a new class of cementitious composite called engineered cementitious composites (ECC) has been gaining ground. A considerable number of studies have indeed shown that ECC offers superior attributes compared to normal concrete in many aspects, including tensile strength and strain capacity, toughness, and durability. These findings surely suggested that there is a great potential for its application as a modern construction material, especially for structures requiring considerable amount of toughness and ductility, i.e. structures subjected to extreme loading conditions such as earthquake, impact and blast. In the present study, ECC dynamic response under compressive loading is of particular interest since the initial stress imposed on the target is in the form of compressive stress wave when impact or blast loads attack a target structure. Hence, if the material has insufficient compression capacity, the target structure may be severely crushed at the initial stage—even before the incident waves are reflected as tensile stress waves. In addition, the compressive behaviour of concrete-like materials can actually be associated to the tensile behaviour since the compression failure is governed by the formation of tensile cracking based on Griffith theory of brittle fracture. The Split Hopkinson Pressure Bar (SHPB) or Kolsky Bar is an apparatus commonly used for determining mechanical properties of materials at high strain rate. Cylindrical samples have been traditionally and routinely used as SHPB test samples due to their ease of machining on a lathe for solid metal samples. For cementitious composites, however, cuboidal specimens may be preferable than cylindrical samples due to these reasons: (1) smooth, flat, and parallel contact surfaces are formed immediately once the specimen is hardened, thus surface grinding process is no longer necessary and the specimen preparation process can be expedited; (2) the flat face of cuboidal specimens facilitates better surface mapping for high-speed imaging, hence allows the use of optical measurement technique. Since limited information is available to assess the suitability of using cuboidal specimens for SHPB tests, the comparison of compressive behaviour between cylindrical and cuboidal specimens is presented and the factors that may cause discrepancies on the results are explained in this study. The outcomes of the research indicate that the cuboid and the cylinder specimens will show comparable dynamic stress-strain curves in the early stage of loading (i.e. low strain regime), but will disclose noticeable difference at the later stage of loading (i.e. large strain regime, after the test specimens reaching the peak capacity). The phenomenon is attributed to the change in boundary friction condition at the bar-specimen interfaces. At the beginning of the test, the boundary friction can be kept to a minimum owing to sufficient amount of lubricant applied at the interfaces. Nonetheless, at the latter stage when the specimen has undergone significant deformation, the boundary friction will gradually increase due to diminishing thickness of the lubricant layer—the lubricant might be squeezed out or might be absorbed into flaws or cracks formed in the specimen at large deformation. Further examination of the failure crack patterns with the aid of a high-speed camera suggests that the cylinder specimens offer more reasonable and reliable results than the cuboid specimens; while the cylinder specimens are characterized by a splitting tensile failure mode, the cuboidal specimens are characterized by a combination of splitting tensile with corner or diagonal failure mode. The comparison of dynamic compressive behaviour between ECC and mortar is also investigated in this study. It is evident that ECC still outperforms mortar even at high strain-rate conditions. The post-peak branches of ECC stress-strain curves display gentler slope than that of mortar, indicating ECC greater energy absorption capacity. The recovered specimens at the end of the tests also indicate less catastrophic damage appearing in ECC specimens; ECC specimens break up into few large fragments, but mortar specimens shattered into many small fragments. Finally, the rate-sensitivity of ECC mechanical properties—in terms of the compressive strength and the critical strain—is discussed and the empirical DIF formula useful as inputs for impact or blast simulation is derived.MASTER OF ENGINEERING (CEE