Concrete is one of the most common materials used to construct a variety of civil infrastructures. However, since con-crete is susceptible to fractures, it is essential to confirm the strength development of concrete during the curing process, in order to prevent unexpected collapse. To address this issue, this study proposes an artificial neural network (ANN)-based strength estimation technique using several kinds of strength related factors of concrete materials. In particular, the variations in mechanical properties of concrete were measured through electro-mechanical impedance (EMI) change using an embedded piezoelectric sensor. The ANN was trained to estimate the strength of concrete by using watercement ratio, curing time and temperature, maturity from internal temperature, and 1-CC of the EMI signals. The trained ANN was verified with conventional strength estimation models throughout a series of experimental studies. According to the comparison results, it is noted that the proposed technique could be very effectively applied to estimate the strength of concrete.

The relationship among air permeability, pressure, and pore size from a viscous- to a molecular-flow region is not well understood. In this work, air permeability in a straight circular pipe was studied considering viscous and molecular flows. I learned that the air-permeability coefficient and intrinsic air-permeability coefficient exhibit contrasting pressure dependence: that of the air-permeability coefficient is larger in a larger pore, whereas that of the intrinsic air-permeability coefficient is larger in a smaller pore. I thus proposed a method to obtain the air-permeability coefficient at atmospheric pressure from that measured under vacuum or pressurised condition. From the Reynolds number study, turbulent flow study is unnecessary in air flow in concrete.

We reported “Influence of Relative Mechanical Strengths between New and Old Cement Mortars on the Crack Propagation of Recycled Aggregate Concrete” in Journal of Advanced Concrete Technology, Vol. 15 (2017) No. 3 p. 110-125; doi: 10.3151/jact.15.110. After publication of the paper, we have become aware of the copyright violation from the original research. We therefore wished to retract the paper and apologize for any inconvenience caused by the retraction.

In marine environments, deterioration of concrete infrastructures under airborne chloride attack is a common problem, which raises the need for a reliable prediction model of airborne chloride penetration into concrete structures to evaluate the service life of concrete structures. This study proposes a time-dependent computational model for predicting the amount of airborne chloride ingress into concrete under actual environmental conditions. The proposed model calculates the amount of chloride penetration by considering the amount of advection and diffusion of airborne chloride on the concrete surface. To compute the amount of airborne chloride penetration, the proportion of dry and wet sections on the concrete surface is assumed, and to ensure accurate prediction of chloride penetration into concrete structures under actual environment conditions, the washout effect of rainfall is taken into account in the calculation. The proposed model was verified through comparison of the experimental and on-site measurement results.

Concrete is tested using slump testing, but slump testing is a method for evaluating concrete consistency, and it is thus difficult to directly evaluate concrete separation resistance by this method. Generally, it is common to perform slump plate tamping after slump testing, and observe concrete deformation to evaluate the concrete separation resistance, but this method is not quantitative. This research proposes a simple method for evaluating concrete separation resistance, based on compaction completion energy, by confirming the presence of circles on the top of a concrete sample after performing hammer tamping used in air volume measurement until the slump flow of a sample reaches 47cm after slump testing. This paper is an extended version of the author’s previous publication (Maruya et al. 2013).