An Experimental Study of Autogenous and Drying Shrinkage

Abstract

Shrinkage of concrete is the time-dependent strain in an unloaded and unrestrained specimen at constant temperature. It is usually considered to be the sum of drying shrinkage and autogenous shrinkage. Drying shrinkage is the reduction in volume caused principally by the loss of water during the drying process and this continues perhaps for years after the concrete is cast. Autogenous shrinkage results in the main from various chemical reactions within the cement paste and occurs in the first days and weeks after casting. All else being equal, drying shrinkage increases with an increase in the water to binder ratio and autogenous shrinkage decreases. For higher strength concrete, autogenous shrinkage is significant and must be considered in the design of concrete structures for serviceability.

The standard procedure for measuring shrinkage of concrete involves measuring the total shrinkage strain in a concrete prism between ages of 7 days and 56 days under a specified controlled environment. This fails to account for the autogenous shrinkage that occurs within the first 7 days and frequently leads to early-age cracking when this shrinkage is restrained. The testing method is also inconsistent with the approach specified in codes of practice (such as EN 1992-1-1:2004 and AS 3600-2009) for quantifying autogenous and drying shrinkage separately.

This paper describes an experimental investigation of shrinkage in Australian concrete in which a reliable experimental method for measuring autogenous shrinkage is proposed and used to quantify the autogenous shrinkage in concretes of strengths ranging from 30 MPa to 80 MPa. To date the test data indicates that autogenous shrinkage is underestimated in both the Australian Standard and Eurocode 2. Modifications of the existing expressions for autogenous and drying shrinkage specified in the Eurocode 2 are also proposed.

Appendix – Proposed Modification to Eurocode 2

The existing expression in Eurocode 2 for the autogenous shrinkage strain ε _ca(t) at any time t after casting is the product of a time function β _ca(t) (varying from zero at t = 0 to unity at t = ∞) and the final autogenous shrinkage strain ε _ca(∞) as follows:

$$ \varepsilon_{{\text{ca}}} (t) = \beta_{{\text{ca}}} (t)\,\varepsilon_{{\text{ca}}} (\infty ) $$

(A.1)

where

$$ \beta_{{\text{ca}}} (t) = 1 - \exp ( - 0.2t^{0.5} ) $$

(A.2)

and

$$ \varepsilon_{{\text{ca}}} (\infty ) = 2.5(f_{{\text{ck}}} - 10)10^{ - 6} $$

(A.3)

The above equations lead to an underestimation of the autogenous shrinkage of Australian concretes. To more accurately predict autogenous shrinkage, particularly for Australian concrete, it is here proposed to replace the expression for the final autogenous shrinkage at time infinity in Eq. A.3 by the following:

$$ \varepsilon_{{\text{ca}}} (\infty ) = 4.0(f_{{\text{ck}}} + 10)10^{ - 6} $$

(A.4)

The predictions made using Eqs. A.1, A.2 and A.4 are compared in Figure 4 with the experimental results presented in Table 3, showing much improved agreement.

Fig. 4.

Comparisons between measured autogenous shrinkage and predictions made using the proposed revision to Eurocode 2 (2004).