Concrete Slab on grade – Failure by design
If you are interested in what is happening with concrete volumes you can check out the statistics at the Golden Bay Cement website. The March quarter shows volumes running at 917,314 m3 but this is actually the second quarter running that volumes have declined. The biggest quarter was September 2014 at 990,616 m3.
There is no doubt the beginning of 2015 did feel quieter but now we have all been feeling the pressure as the workload builds. We re certainly seeing that here at Conslab. We monitor the area of place and finish work we do by month. This has risen to an incredible 70,000m2 per month. Our record to date has just been set in June with a massive 76,084 m2.Engineers have the best of intentions when they write their specifications but there are clauses, which when viewed against an end user or owners priority list of what is important in a slab, reduce the chance of the slab construction being successful. Conslab believe that specifications should be reviewed from the this perspective of what makes the slab a success rather than the material perspective from which most specifications originate. In this article Andrew Dallas, Technical Director of Conslab discusses various clauses from his perspective of what makes a successful slab.
GOOD INTENTIONS
Much of what we see with slab on grade specifications is simply copied from previous specifications, with little change between contracts. Generally they are based around “Good Practice” for concrete as a material. However not as much consideration goes into considering if they actually achieve the aims of the client particularly considering the construction risk they introduce.
Eg. Examples of “Good Practice” clauses:
- Low w/c ratios
- No added water after the truck leaves the plant.
- 60mm slump on site prior to adding superplasticiser.
- The concrete may not be pumped.
- Cutting every second bar of the reinforcing.Viewed from a constructors perspective we suggest that these clauses reduce the chances of the client receiving what they might want. But this does require reviewing specifications from the point of view of the clients priority list and not from a materials perspective.
WHAT MAKES A FLOOR SUCCESSFUL?
To assess whether a floor is successful or not then you must have a prioritised list of what will make that floor successful.In my opinion the Aesthetics of the floor is the number one priority. Generally this is not specified. The only specification which we see focus on this is the Ancon Beton specification for the Bunnings retail stores. However our experience is that this is the key consideration for most owners. As long as it looks good this is the first criteria mentioned when assessing the quality of a floor.Flatness is next as this affects the ability of the machinery handling equipment to operate efficiently. This can have a fundamental effect on the economy of that warehouse in the long term. It also provides a flexibility to owners in terms of installing racking at closer centres at some future time.Durability is important. Free joints and tied joints cannot spall and deteriorate as this will affect machinery speed across those joints and could close areas of the warehouse for maintenance at some future stage. The surface must remain sound for the expected life, particularly at corners where hard wheeled forklifts turn regularly.
There can be no obvious failure such as wide random cracking or surface delamination. Fine cracking such as surface crazing has to be accepted but wide cracks are considered an obvious failure. Delamination where the surface starts to fail is a placers worst nightmare as it is difficult and very expensive to fix. If there are wide random cracks or surface delamination it cannot be called a successful floor although it may still be perfectly useable.
The surface cannot dust. No building owner or tenant wants to spend the first six months of the buildings life removing dust off their stock.
That is my list, in order of priority, of what makes a successful slab yet in reading most specifications you might consider that low shrinkage was the fundamental cornerstone of what makes a good slab. This article is not written in defense of soft pump mixes with 13mm aggregate or the uncontrolled addition of water by a pump operator wanting an easy life but it does question why the focus on shrinkage.
Reducing shrinkage by 20% might reduce a free joint width from 25 to 20mm. You still have a free joint and this must be armoured and dowelled to provide load transfer. Probably the one advantage of reducing shrinkage is in reducing curl. Curl at open free joints or roller doors can be very limiting to forklift traffic. Generally to the perimeter of slabs against panels curl does not matter as this will be hidden under racks. However to free joints in aisles and doors it can be very disruptive and very difficult to fix. With an armoured joint you cannot simply grind the surface down.
While you can reduce the shrinkage and thus curl you still will have shrinkage and curl and thus the design must allow for this. Dowels to panels and footings can limit the curl but between two curling slabs the dowels simply bend as the slab opens. Conslab have a patent for Conslab Rhino, an armoured joint which holds the slab down to the foundation and restricts uplift.
So considering the priority list of what makes a slab successful it is worth assessing the various “Good Practice” clauses listed above. All of these clauses have been extracted from recent specifications.
Clauses Focusing on the Concrete Mix
“The concrete shall contain a maximum water cement ratio of 0.45”
Often we will find the specification sets a maximum water cement ratio. The actual figure varies by specification. It is hard to understand quite what drives this clause. If the engineer was designing a wharf deck where chloride attack was an issue or a factory where chemical attack might be an issue then this clause would be sensible, but we are seeing this in retail stores and distribution warehouses. Water cement ratios are intrinsically linked to durability, in particular chloride ingress, porosity and chemical resistance. However these are qualities not generally linked with slab on grade. With slabs on grade the major durability consideration is abrasion resistance. For this you require a minimum cement content of around 300kg (above about 325kg you get little value), you require a repeat power trowelled surface to case harden the surface of the slab and you require good curing. “The abrasion resistance of 22 MPa concrete slab finished by repeat power troweling tested higher than that of a 52 MPa slab finished by single power troweling. Curing was the second most important factor…”. (Chisholm, 1994)
Alternatively the engineers may be specifying this as another way of reducing shrinkage. Shrinkage though is controlled by how much aggregate is in a concrete mix. The greater the aggregate content the lower the shrinkage. Lowering the w/c ratio can in fact increase shrinkage as a low w/c ratio requires greater cement contents. Ready Mixed Concrete producers work to what they consider a lower bound for water in their mixes and once they hit this point to reduce the w/c ratio they add cement. More cement means less aggregate and thus greater shrinkage. It also means greater cost of the mix.
A focus on water cement ratio does not provide the durability generally required of a concrete floor. What you do get with a low water cement ratio is risk.
A low water cement ratio means concrete companies reduce the water to a minimum to keep cement contents economic. In floor slabs this means a lack of bleed and leads to increased risk of delamination as the top dries rather than sets and the placer is forced to get onto the slab to begin finishing. If placed in a windy environment due to being outside or with panels missing off the walls this lack of bleed means that the slab is at significant risk of plastic cracking. There is also just the simple matter of cost as you will normally require a superplasticiser to provide workability adding around $10/m3 to the price.
Most common is the clause: “No water shall be added to the concrete mix after it leaves the plant.” This is clearly to avoid uncontrolled water addition leading to strength loss and greater shrinkage in the concrete, which of course no one would want. However it takes no consideration of why a truck would arrive on site with low slump. Concrete plants very carefully weigh the cement and admixtures. For these items they can be very accurate. Where they are not accurate is in the estimation of the moisture contained within the mass of aggregates and sand they batch. If a truck has not been delayed and arrives on site with a low slump then the cause will almost certainly be not enough water in the mix in the first instance. Adding water to bring the truck load of concrete up to target slump is the sensible thing to do and will not lead to strengths below nor greater shrinkage than in all the other trucks on site. Adding water to bring the concrete to target slump avoids sending away a truck which you can roughly estimate is worth $1000 and has nothing wrong with it other than, ironically, a lack of water. If you use superplasticiser to bring the slump up then you will get uneven set in the floor as the low slump mix sets up early causing finishing issues. What should be avoided on site is the uncontrolled addition of water or addition of water by those whose interest is not in the final quality of the floor slab.
Having a slightly higher slump is also not going to create a large difference in shrinkage. Testing by Allied Concrete on the difference in shrinkage of a mix with 50mm slump and one with 100mm slump showed a difference in ultimate drying shrinkage of 90 microstrain. The 50mm slump was 700 microstrain at 56 days and the 100mm slump 790 microstrain. Analysing the shrinkage of similar mixes using the research within Nevilles’ book “The Properties of Concrete” we’d estimate a 100 microstrain difference. Of course this is not what you are going to see on site. New Zealand, currently at least, does not see 50% relative humidity (as used within the test method) very often and the concrete is likely to have restraint from reinforcing, sub-base friction etc. Thus my experience is that you will see around 50% of the ultimate movement as joint openings on site. Thus that increase in slump from 50mm to 100mm translates to 45 microstrain. Being pedantic you could say this is 10% more than you would otherwise expect but in reality means a difference in the joint opening at a free joint spacing of 25m of 1mm. The shrinkage gain is not worth the increased risk of construction issues trying to work with such mixes.
Along the same lines are low initial slumps brought up to target slump with superplasticiser. “60mm design target slump, 90mm maximum slump with the use of approved plasticiser/water reducing agent.” These mixes are susceptible to the same issues as the low water cement ratios – low or no bleed leading to delamination and plastic cracking issues or an inconsistent set due to varying doses of superplasticiser.
“Concrete shall not be pumped without the engineers approval”. In many instances there is no other practical way to get the concrete on site other than pumping and I would suggest this clause is routinely ignored. The main value in the clause is that it does allow the engineer to cut out the 13mm grout pump mix which would be completely unsuitable. My impression however is that it is often included as another way of specifying for reduced shrinkage concrete. Alternatively the pumping clause says that where pumping is allowed: “..only a standard structural concrete mix design may be used.”. There is no such thing as a standard structural mix. Between concrete companies there are significant differences in mix design philosophy so what might be standard for one is pump mix for the other. Quite how an engineer seeing a mix design every couple of months decides what is a standard structural mix design is beyond me.
It also raises the obvious point. Pump mixes are designed to pump. Structural mixes, outside of areas with rounded aggregates, generally won’t pump otherwise they would be the pump mix. Trying to pump structural mixes will simply lead to the obvious risk of a pump clogging and getting cold joints in the floor. I’d take slightly larger joints with a better finish to the floor every time. It raises the point that it is the aesthetics of the floor which is the major unwritten specification.
Concrete admixtures are to be approved by the engineer, or in a recent specification; “ Superplasticisers and air entrainment are recommended.” Admixtures are used in almost every concrete mix, in particular water reducing admixtures, and no engineer unless they are intimately involved in mixes will have any idea as to the benefits of the various brands and types available on the market.
For concrete slabs I would recommend that superplasticisers (a high range water reducer) and air entrainment are not recommended. While we have touched on superplasticisers air entrainment is another significant risk for slabs on grade. For any slab where the surface has to be power trowelled air entrainment significantly increases the risk of delamination. The air coalesces into lenses under the top surface and this causes the top to break away where there is this discontinuity in the concrete matrix. Air will also reduce bleed, another significant contributor in delamination issues. Air is only required where the concrete surface will be subject to freeze thaw issues. Thus the area outside coolstore doors which is a prime area for surface degradation due to freeze thaw should include air in the mix but otherwise in very few instances.
Floor Construction
Restraint is a major contributor to random cracking in floors. If the floor surface and aesthetics is important to the client then at an early stage in the design of the building the engineer should allow to have the slab floating free within the structure. Sometimes this is difficult with holding up wall panels in fires or retaining wall situations where the floor is acting as a prop. In larger buildings these issues can be avoided through the use of large footings under the free floating slab. In smaller developments this can be impractical. It’s very difficult to see how a free joint or a saw cut can work when a slab is restrained by starters to a footing or a wall panel. It’s only when you get a few metres away from the restraint that the shrinkage control joints can start to take effect. Practically it maybe best to delete the free joint and instead use more sawcuts at closer centres.
Pouring floors hit and miss is seen as reducing the width of shrinkage joints and is fairly commonly specified. It will certainly increase cost, reduce flatness and increase the risk of colour difference. With laser screeds and ride on trowels pours of 2000m2 are easily handled and sizes significantly greater than this are poured. For every additional floor pour this will increase the cost by $500 – $1000 per establishment for what gain? It will take longer in a
time when construction programmes are squeezed. Joints are the main areas where flatness is compromised. The UK Concrete Society Technical report No 34 says that “Joints create unavoidable discontinuities in floors……On new floors, sawn joints do not usually affect surface regularity. Formed joints….may have more effect.
The colour of a floor surface is significantly affected by how the slab has hydrated and been finished. Pouring on different days has an obvious affect on this due to temperature, humidity and wind conditions and the subsequent effects on the finishing process. A single pour is much more likely to have a uniform colour.
“Concrete shall be at least 7 days old before pouring adjacent sections of the floor slab.” During this time the slab is being cured so it has very little ability to dry out and thus will undergo virtually no drying shrinkage. It is hard to imagine what this clause is in a specification for. Possibly it is to protect the surface alongside from early construction traffic and damage. The risk of uncontrolled cracking tends to be from the opposite situation, where the time between pours has been too long – a rule of thumb might be do not have more than a one month gap between slab pours. This situation can cause cracking perpendicular to the joint due to the differential shrinkage between the two pours.
Sawn Joints
At sawn joints we are often asked to cut every second bar of the mesh. Most traditional mesh slabs are designed unreinforced and the mesh is in the slab specifically to hold the sawcuts tight, to allow the concrete to transfer vertical loads through aggregate interlock in the crack at the saw cut. The spacing of free joints is constrained by the area of mesh in the slab as otherwise the mesh might yield at a saw cut and the aggregate interlock mechanism will not work. If you are going to cut every second bar out then you need to put twice as much in to ensure you do not yield the mesh at sawcuts. If you believe you need the full area of mesh to hold a random crack together why would you cut half the mesh out specifically where you are promoting a crack.
Of far greater import in ensuring cracking happens at sawn joints is the timing and the depth of the saw cut. Early age saws are a great way of ensuring you do not get random cracking occurring over the first night. You need to create the line of weakness prior to the slab going through its first major movement. On the first night the slab is similar to a potato crisp – brittle but weak and the drop in temperature will make that slab want to shrink, particularly if it was a nice sunny day on the day of the pour but then a cold night.
Crack inducers work most effectively in creating early cracking. Unfortunately they can often create the crack prior to sawcutting and then you will have sawn joints with cracks meandering along and across the sawn joint. Often the saw cutter will cut the slab even though it has clearly cracked. It might have been better to save the money on the sawcutting and invest it in the crack repair.
Floor Thickness
Floor thickness while critical and specified is almost never checked until a failure has occurred. It is impossible for an engineer undertaking a monitoring visit to know what the thickness is as they can only view it at the external boxing. The only way to know is to have a grid of levels taken to both the sub-base and finished surfaces. There is a tolerance to the subgrade and finished levels which engineers need to understand.
Engineers need to get a feeling for what these construction tolerances are and where in the design or the specification are they allowed for, particularly with proprietary designs aiming for the ever thinner solution.
The levels of the subgrade should be monitored, perhaps by an independent party and then the finished levels of the slab surface measured.
Design Clarity
It gets more complicated in design build contracts where lack of clarity is what generally creates the issues. It appears engineers often feel lost as to where the boundary sits between their overall building design and the design build floor. For example although they have issued loads, sub-grade strengths and be asking for design build they will nominate a slab thickness on the plan. Is this a minimum thickness, is it what has to be provided or was it simply a figure put on the plan to get building consent? The specification will often include significant restrictions around the mixes while taking no responsibility for the floor design or construction.
The lack of detail in the loading specification can create significant differences in the design solution and could easily cause difficulties for the building owner or tenant in future. For example when providing a racking load what foot size should be assumed? If 40’ containers are specified in the yard what is the weight that should be assumed? Is it 32 tonne or can we allow for a lesser weight? At what spacing are the containers? Can they be hard against one another? Will they be placed right against the edge of the slab or can we assume the containers will be kept away from the edge? Should we make allowance for fatigue or are they only an occasional load? The devil is in the detail as they say and can result in what appears to be an apples with apples comparison of designs being anything but. The current issue in New Zealand is how the seismic load for racking is to be handled.
The other major design input provided is the specification of the subgrade strength. Slabs are designed using a modulus of sub-grade reaction, a “k” value. However this is not what is provided. Generally we will see a CBR or an allowable bearing capacity. Interpretation from these figures to a design k value can be very rough. Conslab generally recommend working from a Benkelman beam reading as the loads used in testing give an idea at least of the sub- grade strength over some depth. CBR in particular, depending on how it is derived, can simply give a reading of the sub-base strength which is virtually irrelevant in the slab design.
Clients being their own Worst Enemy
Workmanship makes up a large proportion of the issues in slabs. You can specify what you like but much of the quality will come down to the attitude and knowledge of the guys on the tools. And once poured most floor slabs do not come out. Yet despite the importance of the floor to success of a warehouse or retail building these are generally tendered to the lowest cost operator. They are large lump sum items in a quote and the contractors in a tight tender market need those low bids to win. A system designed to drive quality to the absolute minimum level acceptable. While the engineers often understand this the clients are reluctant to move from a tender system for a concrete slab which they perceive as a commodity item despite its importance to the success of the building.
SUMMARY
Excellent concrete slabs are anything but a commodity item however the lack of consideration of the risks to the floor construction leads engineers to focus on the material characteristics to the detriment of the construction aspects. This focus on the material rather than the customers aims gives rise to clauses that increase the risk of the floor not being successful in the clients eyes.
Compound this with a tendering system that drives quality out of the construction process means that you can easily get issues compounding.
“Good engineering practice” should include a balancing of the required material characteristics with the practical aspects of being able to build a concrete floor. Wet concrete when laid out thinly and exposed to the elements is a fragile beast. Consideration of the risks should be at the forefront of every engineers mind when specifying a concrete floor slab.
Quotes in bold are extracted from recent specifications on New Zealand contracts. They will not be attributed to any contract or engineering consultancy.
ANDREW DALLAS
Conslab Ltd