Shrinkage and creep in post-tensioned slabs
Concrete keeps moving long after the tendons are stressed. How shrinkage and creep are predicted to Eurocode 2 and ACI 209, what they do to prestress, and how restraint from walls and cores turns shortening into cracks.
To run eq. 5.46 or the ACI (Zia et al.) method with your own concrete, humidity and stressing age, use our PT long-term prestress loss calculator.
Shrinkage is the reduction in volume of concrete as it dries and as the cement hydrates. It happens whether or not the slab is loaded. Creep is the gradual increase in strain under sustained stress, and in a post-tensioned slab the stress is always there: the tendons keep the concrete in compression for its whole life. Both effects matter more in PT slabs than in ordinary reinforced slabs, for three reasons:
- they reduce the prestress, because the strand shortens with the concrete around it;
- they shorten the slab, and where that shortening is restrained by walls, cores and columns, the result is cracking and loss of precompression;
- they increase deflection over time; see long-term deflection to Eurocode 2 and ACI 318.
Predicting creep and shrinkage to Eurocode 2
EN 1992-1-1 clause 3.1.4 covers creep and shrinkage, with more detailed expressions in Annex B.
- Creep coefficient. φ(∞, t0) depends on the relative humidity, the notional size h0 = 2Ac/u (Ac the cross-section area, u the perimeter exposed to drying), the age at loading t0, the cement class and the concrete strength. Figure 3.1 gives it graphically for inside (RH 50%) and outside (RH 80%) conditions; Annex B.1 gives the formulas and the development with time. The creep strain is εcc(∞, t0) = φ(∞, t0) · σc/Ec, where Ec is taken as 1.05 Ecm. Where the compressive stress at loading exceeds 0.45 fck(t0), non-linear creep should be considered.
- Shrinkage. The total shrinkage strain is the sum of drying shrinkage and autogenous shrinkage, εcs = εcd + εca. Drying shrinkage develops slowly and depends on humidity, notional size and cement class (Table 3.2 and Annex B.2). Autogenous shrinkage develops early and depends on strength: its final value is εca(∞) = 2.5 (fck − 10) × 10−6.
For a slab, both faces usually dry, so h0 is close to the slab thickness. A thinner slab therefore dries and shrinks faster than a thick transfer slab of the same concrete.
Predicting creep and shrinkage to ACI
ACI 318 does not prescribe a creep or shrinkage model. ACI-based designs normally use the ACI Committee 209 documents:
- ACI 209R-92 gives an ultimate creep coefficient of 2.35 and an ultimate shrinkage strain of 780 × 10−6 for standard conditions, multiplied by correction factors for age at loading, curing, relative humidity, member size and concrete composition, with time functions for how they develop.
- ACI 209.2R-08 compares several models, including ACI 209R-92, Bažant-Baweja B3, CEB MC90-99 and GL2000, and explains how to choose and calibrate them.
The model matters less than the inputs. In the UAE and wider GCC, outdoor humidity varies widely between coastal and inland sites and between seasons, interiors are air-conditioned and often dry, and hot-weather concreting affects early curing. Choose a humidity and curing assumption that reflects the actual exposure, and state it in the calculation report.
Effect on prestress
As the concrete shortens, the strand bonded to it (or anchored at its ends) shortens too and loses stress. EN 1992-1-1 clause 5.10.6 gives a combined expression for the time-dependent loss at the tendon:
Δσp,c+s+r = [ εcsEp + 0.8Δσpr + (Ep/Ecm) φ(t,t0) σc,QP ] / [ 1 + (Ep/Ecm)(Ap/Ac)(1 + (Ac/Ic) zcp2)(1 + 0.8φ(t,t0)) ] (5.46)
The three terms in the numerator are shrinkage, relaxation of the strand and creep under the quasi-permanent stress in the concrete at the tendon level. The denominator accounts for the interaction between them and is usually close to 1.0 for slabs, where the tendon area is small compared with the concrete.
Illustrative example (assumed values, not from any project): C40/50 concrete with Ecm = 35 GPa, strand with Ep = 195 GPa, a final shrinkage strain of 400 × 10−6, a creep coefficient of 2.0, a quasi-permanent concrete stress at the tendon of 2.0 MPa and a relaxation loss of 32 MPa.
- Shrinkage: 400 × 10−6 × 195,000 = 78 MPa
- Relaxation: 0.8 × 32 ≈ 26 MPa
- Creep: (195/35) × 2.0 × 2.0 ≈ 22 MPa
- Numerator about 126 MPa; with a denominator of about 1.02 for a typical slab, the long-term loss is about 123 MPa, or a little under 10% of an initial stress of around 1,300 MPa after friction and anchor set.
The example shows why shrinkage is often the largest single long-term loss in a slab, and why a lump-sum allowance should be checked against the actual concrete, exposure and stressing age. In North American effective-force practice, an effective stress of around 1,200 MPa (175 ksi) is often assumed for unbonded slab tendons, but that assumption still needs to be justified for the profile, length, friction and long-term losses on the project. Short-term losses from friction and anchor set are covered in our guide to friction and wobble losses.
Axial shortening and restraint
A PT slab shortens from elastic shortening at stressing, then from creep under the precompression, shrinkage and, in exposed slabs, temperature change. An illustrative estimate for a 60 m long slab with 1.5 MPa average precompression and a short-term modulus of 30 GPa:
- elastic shortening: 1.5 / 30,000 = 50 × 10−6
- creep with φ = 2.5: 125 × 10−6
- shrinkage: 400 × 10−6
- total about 575 × 10−6, or about 35 mm over 60 m: roughly 17 mm at each end if the slab shortens towards its centre
If the slab is free to move, that shortening is harmless. If stiff elements at the ends of the slab hold it in place, the movement is restrained. Then:
- part of the prestress goes into the walls and columns instead of the slab, so the slab has less precompression than the design assumed;
- restraint tension develops in the slab and cracks form, typically near the restraining walls and across the slab between them;
- the walls and columns take shear and bending from the slab's movement and can crack themselves, especially at the lowest restrained levels and at the ends of long buildings.
Stiff cores, walls and pour strips
The layout of the vertical structure decides how much restraint a PT floor sees. A single core near the centre of the floor plate restrains much less than two stiff cores at opposite ends or long perimeter basement walls. Measures we use, often in combination:
- Pour strips (closure strips). A strip of slab left uncast between two stressed pours, then cast later once a good part of the early shortening has taken place. The width, location, duration and reinforcement through the strip are set by the designer and the specification, and the strip needs temporary propping and its own stressing or lapping details.
- Stressing before restraint. Stressing pours before the slab is locked to stiff elements, and sequencing pours so they shorten towards a free edge or a pour strip.
- Release and sliding details. Temporary or permanent slip joints, dowelled connections or sliding bearings between slab and wall where movement must be allowed.
- Movement joints. On very long structures, especially car parks and podiums exposed to temperature change.
- Restraint reinforcement. Additional bonded reinforcement in the slab near restraining walls, and in walls and columns at the restrained levels, to control crack widths.
In a RAM Concept model, full in-plane restraint at walls can show much less precompression in the slab than will exist if pour strips are used; see our RAM Concept tips. RAM Concept's load history calculation also lets you choose a creep and shrinkage model and apply an external restraint percentage, which increases cracking and long-term deflection where stiff supports restrain shrinkage.
Site practice that limits shrinkage problems
- Early and continuous curing, especially in hot, windy conditions where plastic shrinkage cracking can occur within hours of casting.
- Mixes with a controlled water content and a cement content no higher than needed for the strength and durability requirements.
- Pour sizes and sequences agreed with the designer, and pour strips left open for the period specified.
- Stressing at the specified concrete strength and age, with partial stressing where the design calls for it to control early cracking.
Talk to us about shrinkage and restraint on your project. Our team checks long-term losses, shortening, restraint and pour strip layouts as part of PT slab design and review. Send us the layout for a PT design review or a full post-tension design package, ask about site supervision and stressing support, read about our post-tension design in the UAE, or contact our engineers.
Related PT design guides: temperature and restraint cracking, precompression and balanced load.
Part of our PT technical series. Codes are referred to by clause where we are confident of the reference; always check the edition and national annex adopted for your project.
Questions we are asked about this.
How much prestress is lost to creep and shrinkage?
It depends on the concrete, humidity, slab thickness, stressing age and precompression. In our illustrative example with assumed typical values, the long-term loss from creep, shrinkage and relaxation is a little under 10% of the initial stress. It should be calculated for the project rather than assumed.
What is a pour strip in a PT slab?
A strip of slab left uncast between two stressed pours and cast later, once part of the early shortening has taken place. It reduces the restraint from stiff walls and cores. Its width, location, duration and details are set by the designer and the specification.
Which creep and shrinkage model should be used in the UAE?
Use the model consistent with the governing design code and specification: Eurocode 2 clause 3.1.4 and Annex B for EN 1992-1-1 designs, and an ACI 209 model for ACI 318 designs. The humidity and curing inputs should reflect the actual exposure.
Does shrinkage cause cracking in PT slabs?
Shrinkage itself does not crack a slab that is free to move. Cracking occurs where shortening is restrained by walls, cores, columns or earlier pours, or where early curing is poor and plastic shrinkage cracks form.
Why do walls crack next to PT slabs?
Restrained shortening of the slab pulls on stiff walls and columns, causing shear and bending in them, particularly at the lowest restrained levels and at the ends of long buildings. Pour strips, sequencing and release details reduce this.
Can you check restraint and pour strip layouts?
Yes. We check shortening, restraint, pour strips and stressing sequence as part of PT design and review, and can advise during construction.
Technical information only. This article is general engineering information to help you understand the subject. It is not design advice for a specific project, codes and authority requirements change, and worked examples are illustrative. Designs must be prepared and checked by a qualified structural engineer responsible for the project, to the code edition adopted for the project. Disclaimer · Terms of use.
Send the drawings. Get an answer from our engineers.
WhatsApp or email the architectural or structural set, the location and what you need. The reply comes from the engineers who will do the work, with a proposed scope and programme.
- +971 52 118 2569
- Telephone
- +971 52 118 2569
- Office
- Al Warqa 1, Dubai, UAE
- Hours
- Sunday – Thursday, 09:00 – 18:00 GST


