Figure 11.1. Design stress-strain relations for concrete in compression specified otherwise (Clause 3.5(1P)). The most important reason is that sustained loading effects, if occurring anyhow, will occur after a considerable time. So when loaded the concrete will be much older (maybe even years) than 28 days. The sustained loading effect is therefore with high probability compensated by the gain in strength between 28 days and actual loading of the structural member (see further the background report of chapter 3.1).
In lightweight concrete, however, the increase in strength after 28 days is smaller than in normal weight concrete . Furthermore it is reported that the sustained loading effect is more pronounced than in normal aggregate concrete. Weigler  reported that the strength of lightweight concrete under sustained loading was only about 70-75% of the short term strength. Similar results were obtained by Smeplass . The results are explained by creep of the matrix, overloading the aggregates. Consequently this phenomenon occurs when the strength of the aggregates it utilized to its maximum . Since further research seems to be necessary here, this would support the idea of introducing a sustained loading factor for lightweight concrete anyhow. Therefore in 10.3.1.5 alcc should preferably be defined as 0.85 "unless specified otherwise".
The tensile strength of lightweight concrete can be obtained by multiplying the corresponding strength of normal density concrete of the same strength class with a factor
where p is the upper limit of the oven-dry density.
Here it should be noted that the average tensile strength fctm normal density concrete follows from fctm = 0.30 fctk2/3 for concretes <C50/60 (11.2a)
and fctm = 2.12 ln (1+fcm/10) for concretes > C50/60 (11.2b)
the characteristic (5%) value follows from fctk = 0.7 fctm (11.3)
An estimate of the mean value of the secant modulus Elcm for LWAC can be obtained by multiplying the corresponding value for normal density concrete by the coefficient
In the previous part of the Eurocode, ENV 1992-1-4 "General rules for lightweight concrete with a closed structure, the equation ^e = (p/2200)2 was mentioned. For normal density concrete, according to chapter 3.1, the E-modulus is calculated from
In the the old version ENV-1992-1-4 it is denoted that for lightweight concrete the creep coefficient ^ can be assumed equal to the value of normal density concrete multiplied by a factor (p/2300)2 for p > 1800 kg/m3. For p < 1500 kg/m3 a factor 1.3(p/2300)2 can be used. For intermediate values of p linear interpolation may be applied. Furthermore the creep strain has to be multiplied by a factor ^2 = 1.3 for lightweight concrete classes lower than LC20/25.
There is however serious doubt on the correctness of the statement in ENV 1992-1-4 that the creep of lightweight concrete is smaller than that of normal density concrete, in spite of the fact that, according to  also other codes like the Norwegian Code NS 3473, the Japanese Code JSCE and the German code DIN4219 give formulations with the same tendency. Kordina  states that creep is a matter of the cement paste and not of the aggregate, which would imply that similar compositions of LWAC and NDC should give the same specific creep. Neville  developed a two-phase model where he distinghuishes the cement paste and the aggregates as two parallel load bearing components. The stiffer the aggregate, the more load will be carried by the aggregate skeleton and the more the stresses in the paste will decrease. A decrease of the stresses in the paste will result in smaller creep deformation of the paste and hence of the concrete. Since most of the lightweight aggregates have a lower stiffness, the stresses in the paste will remain higher and so the creep of LWAC . Anyhow, existing information seems to confirm that there is not difference between normal density concrete and lightweight concrete with regard to the specified creep, see f.i. Fig. 11.2 .
A reconsideration of the formulation for creep of LWAC, as given in ENV 1992-1-4 and provisionally adopted in the version of EC-2 of 1/1/2000 seems to be necessary. The best formulation seems to be that creep of LWAC is the same as creep of NDC and can be calculated with the same formula's.
Was this article helpful?