Abstract
Creep strain, a requirement of the concrete design process, is a complex phenomenon that has proven difficult to model. Although laboratory tests may be undertaken to quantify creep magnitude, these are generally expensive and not a practical option. Hence, empirical code-type prediction models are used to predict creep strain. This paper considers the accuracy of the Hong Kong Code of Practice Model (HKBD, 2020 [1]) and the Japan Society of Civil Engineers Standard Specification for Concrete Structures (JSCE, 2010 [2]). In the case of the JSCE, two models were considered, one applicable to normal strength concretes (referred to as JSCE) and the subsequently introduced model which is applicable to high strength concretes (> 55 MPa), referred to as JSCE HS. The models were evaluated, when compared with the actual creep strains measured, over a period of approximately six months, on a range of concretes under laboratory-controlled conditions, for six mixes (comprising three aggregate types and two water-cement ratios). The HKBD and JSCE models generally over-predicted specific creep at all ages. Furthermore, the JSCE HS generally slightly over-predicted specific creep for the first 28 days and then under-predicted thereafter. The HKBD model was found to be the least accurate with an overall coefficient of variation (ωall) of 56.2 %. The relatively simple JSCE HS, which was applied to all six of the mixes, of which only one mix classified as high strength, was the most accurate model with a ωall of 21.8 %. The JSCE model was less accurate than the JSCE HS, yielding a ωall of 30.2 %. When comparing the accuracy of the three models investigated to the accuracy of 17 other international code-type models that were previously investigated by the author, the JSCE HS was the second most accurate and the HKBD was ranked 17 th in terms of accuracy (of the 20 models considered), respectively.