Joints in water-retaining structures 3

[renko]

Hello all,

I've been designing water tanks of various shapes and sizes for a couple of years now. The design is to BS8007 with max crack width of 0.2mm.

My question is to do with the provision of movement joints to control early thermal and shrinkage crackin, as I seem to get conflicting advice from several people (that's probably going to happen here too).

Anyway most tanks I've designed do not exceed say 30 - 35m in length, so I generally design for Option 1 - full restraint which theoretically will control craking to a max design width and spacing. And 90% of the time it works, the other 10% I blame the Contractor (just kidding). This method has been approved by a number of my colleagues, but others recommend using joints anyway.

Does anyone know if there is any definite recommendation when the size of a water tank would necessitate contraction joints to control early cracking.

Any advice or comments are appreciated.

 

[JKW05]

ACI 350 Table 7.12.2.1 specifies varying shrinkage and temperature reinforcement ratios related to the length between joints.

 

[Ussuri]

Have you had a look at Ciria Report 91, although I think there is a newer version now.

 

[renko]

Thanks Gents,

JKW05 I don't really have access to ACI Tables or codes.

Ussuri,
Ciria Report 91 appears to back up my decision to design for full restraint using reinforcement to control cracking without movement joints.

The tank is a Chlorine Contact Tank approx 30m x 15m on plan with 5m high walls, with a hollowcore roof. The walls and base are 400mm thick and the crack control steel is T12 @ 150 (horizontal steel). I'm also recommending the use of PFA or GGBS concrete for lower heat of hydration and increased durability.

I'll try to remember to post in this thread again in January when the walls should be up to one or two weeks cast, just for an update.

Thanks again.

 

[Ussuri]

Be sure to advise the contractor of the longer curring times when using the cement replacement, and hence longer to wait to strip the forms. I have in the past, on D&B projects, been told to revise a design from a GGBFS mix to a pure OPC mix. The saving in steel tonnage was out weighed by the impact on the programme of the longer curing times.

p.s. the 'T' rebar designation is actually now defunct, it was replaced by H,A,B or C in BS8666:2005.

 

[renko]

I've advised the conctractor regarding the curing times. they are still deciding whether to use full OPC or not.

Yeah some of our rebar suppliers are still using the older BS4466 designation but we're catching up.

 

[BarryEng]

I have designed reservoirs (excavated in sand & lined with 150 mm concrete) up to 200 m x 200 m. Wall slopes from 1.25 H to 1 V, up to 2 H to 1 V, depending on the soil conditions & type of soil. I usually have a control joint at the base slab/sloping wall concrete & a control joint at the intersection of the sloping walls (at the corners).

All other joints are then construction joints with reinforcement continuous (100 %) thru the joint. In some cases I have used a 10 m wide pour & up to 170 m long. Construction branch then went back & cut 'dummy' joints at a 10 m spacing. I found this out later. The dummy joints were 20 mm x 20 mm & filled with a polyurethane sealant (NEVER USE POLYSULPHIDE SEALANTS - they will 'rot' under the action of suphate reducing bacteria in a potable water storage).

The problem with dummy joints 20 mm x 20 mm, is that they really do not do anything - they are supposed to develop a crack under the sealant but if the joint depth in less than 25 to 35 % of the concrete slab thickness, you cannot guarantee a crack forming at that location. To me, dummy joints are a waste of time.

I design to the Australian water retaining code, & use the min reinf specified, N12 - 0.48%, N16 - 0.64%, N20 - 0.80%, etc. So for a 150 mm slab, the min is N12-150 EW in one layer - usually 50 mm top cover.

When BS 5337 came out (& later BS 8007) I had the local cements checked out for the 3 day tensile strength. Our local cements will give a 3 day tensile strength of over double the BS value. I assume that the local cements are finer ground & basically are a high early strength cement. This is reflected in the higher min reinf requirements of the Oz code. The 3 day strength is used to calculate the min reinf = fct/fy. This is linear with respect to fct, so a higher early strength cement requires a higher min reinf.

Your design (30m x 15m on plan with 5m high walls, with a hollowcore roof & 400 mm walls) would require a much larger reinf % as a min if you were constructing the tank in Oz. Have you thought of using the hollow plank roof as a tension tie & design the side walls as fixed on three sides & supported on the fourth?

For the design of this type of rectangular plate, have a look at thread181-27567 with the following reference: -

http://www.usbr.gov/pmts/hydraulics_lab/pubs/EM/EM27.pdf

Back to your query - I try NOT to use joints unless I am forced into it. Likely reasons are: -
* Larges changes in direction of a thin slab (as in my example above).
* Articulation due to a change of section.
* Changes of restraint as in an inlet pipe, outlet pipe, overflow pipe, scour pipe etc.
* Structural change such as a channel between a clarifier & a filter.
* Vertical shaft next to a tank.
* Possibility of differential settlement (for any reason).

The joints that I use are construction or control joints. I don't believe that contraction joints are ever needed (at least not in my design environment).

Joints are required for: -
* Thermal effects - initial, within 3 days of a pour.
* Thermal effects - long term (daily, weekly, yearly etc).
* Shrinkage (of the cement paste not thermal).
* Swelling (due to absorption of water).

Calculation of each of these effects can then be summed up depending on combinations (the Oz code gives guidance as well as in several publications).

 

[BarryEng]

In my thread above, I said: -
I don't believe that contraction joints are ever needed (at least not in my design environment).

Sorry about that - it should read "I don't believe that EXPANSION joints are ever needed (at least not in my design environment)."

It's been a looong day.

 

[renko]

Thanks BarryEng - very helpful reply. I'm going to try to get my hands on the Aussie code to compare.

I am using the hollowcore roof to support the top of the wall, but the wall length is too long to span two ways.

The BS8007 code divides the section up into surface zones, 200mm each in this case. Then the design will theoretically control the crack width and spacing with respect to the bond between the reinforcement and the concrete.

The problem is that the concrete is often far from homogenous and the theory doesn't match reality.

Anyway I've gone for full restraint with only construction joints. I'll post a month after the tank is built to update the condition.

Thanks again.

 

[BarryEng]

The Oz code is AS 3735 & is in two parts, the code & the commentary. The surface zone is 250 mm for slabs but only 100 mm in contact with the ground.

The BS code provision of Ast = fct/fy was developed by Barry Hughes & there is a full explanation of this development in his book "Limit State Theory for RC Design".

In your thread "I am using the hollowcore roof to support the top of the wall, but the wall length is too long to span two ways." I think that if you have a look at that reference I posted above, there can be a substantial reduction in the bending moment in the wall using the propped cantilever concept (I assume that was your intention from your thread).

I agree that the wall will have a very limited two way action even for the short wall, but the tables in that pub will allow you to very quickly determine the value of the BM in the beam direction (highest at the support in the horizontal direction) & the cantilever direction (highest at the bottom of the wall in the centre of the wall). The tables have heaps of coefficients & once you are familiar with them, you can (in a minute) find the maximum value & then determine the wall thickness.

You can then calculate the max moment & then determine which reinf layer will be on the inside or outside face. I use the tables by going backwards - find the coefficient that corresponds to the min reinf (allowed by the code for your selected wall thickness) & you will find (in most cases) most of the wall will be min reinf. You can then use the same bar spacing for min & max reinf, & then you only need to determine the bar size for the max BM.

I find that I can design (for example) filters much quicker using these tables than by setting up a 'spacegas' or 'strand' model.

If you are going to use the combinations part of the Oz code, be careful with the temperature stresses (short term loading). The method (tables in the commentary) is correct but I do not agree that the thermal gradient is correct. It does not take into account the 'surface effect' (that effectively reduces the thermal gradient). For this effect, have a look at the EURO code (13445) - I consider that 13445 is closer to reality compared to the Oz version.

I agree with your method of full restraint & only const joints. As I said before, if the calculated reinf is based on the qty of reinf to fully distribute the cracks, why do you also have control joints? I also do not understand the use of partial joints in the BS - I can only assume that the concrete, environmental conditions, construction practices & soil types vary from what I am used to.