Concrete and reinforcing testing in the United States

Concrete is the material whose quality is tested most routinely on a building site, and for good reason. Its final properties depend on decisions made at the batch plant, in the truck, during placement and during curing, and the most important property, strength, cannot be measured until weeks after the concrete has been placed and covered with more work. The testing system therefore combines checks on the fresh concrete at the point of placement with strength tests on samples, statistical acceptance rules and, when results are low, investigation of the concrete in place. This lesson covers that system and the inspection of reinforcing and post-tensioning on the Larkspur Commons podium and transfer slab.

In the United States, structural concrete requirements come mainly from ACI 318, the building code requirements for structural concrete, referenced by the International Building Code, together with ASTM test methods and specification documents such as ACI 301. The specification on Larkspur Commons follows these documents, and Treasure Valley Materials Testing, the owner's testing agency, performs the tests in a laboratory accredited for concrete testing, as part of the special inspection program described in lesson 9.

How concrete is specified

Concrete can be specified by performance, where the owner states the required properties and the producer designs the mix, or by prescription, where the owner states the mix proportions. Most building concrete today is specified by performance: the specification sets the specified compressive strength, the exposure conditions, the maximum aggregate size, the air content where freezing and thawing apply, and any special requirements, and the ready-mix supplier submits a mix design with evidence that it can meet them.

ACI 318 organizes durability requirements through exposure categories and classes, such as category F for freezing and thawing, S for sulfate, W for contact with water and C for corrosion protection of reinforcement, each with limits on water-cementitious materials ratio, minimum strength and air content. On Larkspur Commons, the parking garage slab falls into the higher freezing and corrosion classes because vehicles bring de-icing salts in from the street, while the interior podium slabs and the transfer slab are not exposed to freezing.

The specified compressive strengths on the project are 4,000 psi at 28 days for the podium slabs and walls, 5,000 psi for the transfer slab and 4,500 psi with entrained air for the parking garage slab. The submittal review in lesson 4 confirmed that each mix design met its exposure requirements and that the supplier's strength records supported the target strength.

Tests on fresh concrete

When each truck arrives, the delivery ticket is checked against the mix ordered: the mix identification, the volume, the time of batching and any water added. The testing agency's technician then samples the concrete as specified and runs the fresh concrete tests. These results are available within minutes, so the concrete can be accepted or rejected before it is placed.

Adding water at the site is the most common way fresh concrete is spoiled. Extra water raises the slump and makes placing easier, but it raises the water-cementing materials ratio and lowers strength and durability. Specifications usually allow a limited, measured amount of water to be added at the site within the mix design limits, with the amount recorded on the ticket, and forbid any addition beyond that. Dana's rule on Larkspur Commons is that only the testing technician or Mei may authorize water, and only within the mix design, and that any load in which a crew member adds water without authorization is rejected.

Strength tests

Strength is measured on cylinders cast from the sampled concrete, cured under controlled conditions and crushed at a set age, usually 28 days for acceptance, with earlier breaks at 7 days and sometimes at the age needed for an operation such as stressing tendons or stripping forms. A strength test is the average strength of the cylinders made from one sample and tested at the same age. The number of cylinders in a test and their size are set by the governing standard and the specification.

Cylinders must be handled with care, because a badly made or badly cured cylinder gives a low result that says nothing about the concrete in the structure. The first day matters most: cylinders left on a hot or freezing slab, or knocked around in the back of a truck, can lose much of their apparent strength. The specification requires initial curing within a temperature range for the first day, then transport to the laboratory for standard curing. Separate field-cured cylinders, kept beside the structure under the same conditions, are used to decide when the in-place concrete is strong enough for stressing or stripping, but not for acceptance.

For acceptance, the Larkspur Commons specification follows ACI 318: the strength level of each class of concrete is satisfactory when the average of every three consecutive strength tests equals or exceeds the specified strength and no individual strength test falls below the specified strength by more than 500 psi when the specified strength is 5,000 psi or less, or by more than 10 percent of the specified strength when it is greater. ACI 318 requires samples for strength tests at least once a day for each class of concrete, at least once for each 150 cubic yards and at least once for each 5,000 square feet of slab or wall surface area placed, with further rules for small projects. Readers should confirm the clauses in the edition their project references.

The two rules work together. The individual-test rule catches a bad load or a localized problem; the three-test average catches a mix that is running slightly weak over time. A mix that occasionally produces a test just under the specified strength can still be acceptable, because concrete strength varies and the rules allow for that variation.

# Strength acceptance check, podium slabs, specified strength 4,000 psi (illustrative results)
fc = 4000
tests = [4650, 4420, 3940, 4260, 4830, 3790, 4010, 3720, 4340, 4500]   # psi, each the average of one sample's cylinders
low_limit = fc - 500 if fc <= 5000 else 0.9 * fc
for i, t in enumerate(tests, 1):
    flag = "" if t >= low_limit else "  INDIVIDUAL TEST TOO LOW"
    line = f"test {i:2d}: {t:5d} psi{flag}"
    if i >= 3:
        avg = sum(tests[i - 3:i]) / 3
        line += f"   avg of last 3 = {avg:6.0f} {'ok' if avg >= fc else 'BELOW SPECIFIED STRENGTH'}"
    print(line)

test  1:  4650 psi
test  2:  4420 psi
test  3:  3940 psi   avg of last 3 =   4337 ok
test  4:  4260 psi   avg of last 3 =   4207 ok
test  5:  4830 psi   avg of last 3 =   4343 ok
test  6:  3790 psi   avg of last 3 =   4293 ok
test  7:  4010 psi   avg of last 3 =   4210 ok
test  8:  3720 psi   avg of last 3 =   3840 BELOW SPECIFIED STRENGTH
test  9:  4340 psi   avg of last 3 =   4023 ok
test 10:  4500 psi   avg of last 3 =   4187 ok

No single test fell below the individual limit, but the average of tests 6 to 8 dipped below the specified strength. That is a warning, not a structural failure. Mei and the supplier reviewed the tickets and found that tests 6 to 8 came from pours on hot afternoons in which the slump had been raised with site water close to the maximum the mix design allowed. The supplier adjusted the mix with a water-reducing admixture so the crews could get the workability they needed without water, and the next results returned above the specified strength. The engineer of record reviewed the results and accepted the concrete, because the individual tests met the minimum and the in-place strength needed by the design was not in doubt.

Required average strength

Because strength varies, a producer cannot aim at exactly the specified strength; about half of the tests would then fall below it. The mix must be designed for a higher average, the required average strength, which depends on how variable the producer's concrete is. A producer with a low standard deviation, achieved through consistent materials and batching, can use a lower margin and therefore a more economical mix.

ACI 318 and ACI 301 give formulas for the required average strength based on the producer's standard deviation, and CSA A23.1 sets out a comparable statistical approach in its requirements for mix designs. The script below uses the ACI form, in which the required average strength for specified strengths up to 5,000 psi (35 MPa) is the larger of the specified strength plus 1.34 standard deviations and the specified strength plus 2.33 standard deviations minus 500 psi (3.5 MPa). It is shown here to explain the idea; designers should use the formulas in the standard and edition their project references.

# Required average strength from the producer's standard deviation (ACI form, illustrative)
fc = 4000   # psi
for s in (350, 500, 650):
    a = fc + 1.34 * s
    b = fc + 2.33 * s - 500
    fcr = max(a, b)
    print(f"s = {s} psi: fc + 1.34s = {a:5.0f}, fc + 2.33s - 500 = {b:5.0f} -> required average {fcr:5.0f} psi")

s = 350 psi: fc + 1.34s =  4469, fc + 2.33s - 500 =  4316 -> required average  4469 psi
s = 500 psi: fc + 1.34s =  4670, fc + 2.33s - 500 =  4665 -> required average  4670 psi
s = 650 psi: fc + 1.34s =  4871, fc + 2.33s - 500 =  5014 -> required average  5014 psi

A producer whose standard deviation falls from the higher to the lower value can lower the target average by several megapascals or several hundred psi, which means less cement and lower cost for the same specified strength. This is one reason quality pays for producers as well as contractors: consistency is cheaper than excess strength.

Low strength results

When an individual test falls below the limit, or the running average falls short, the first step is to check whether the test itself was valid. The technician's records, the cylinder curing records and the laboratory reports are reviewed. If a test was made or cured badly, the result may be discounted with the engineer's agreement. If the test seems valid, the engineer of record evaluates whether the low strength matters for the element in question, often with a structural check using the actual strength.

If the strength of the concrete in place is still in doubt, cores are drilled from the structure in the area represented by the low test and tested. Both CSA A23.1 and ACI 318 give criteria for judging core results, generally treating the concrete in the area as adequate when the average of three cores reaches about 85 percent of the specified strength and no single core falls below about 75 percent, with specific conditions about moisture conditioning and core size. Readers should check the exact criteria in the referenced standard. If cores do not resolve the question, the engineer may require a load test or strengthening. Under most specifications, the contractor pays for coring and testing when the concrete is found deficient, and the owner pays when it is found acceptable.

Reinforcing inspection

Reinforcing steel must be the right size and grade, in the right place, with the right cover and the right laps, and held firmly so that it stays in place during the pour. Mill certificates confirm the grade and are reviewed when the steel is delivered; bundles are tagged and stored off the ground. The field inspection before each pour checks the placement against the placing drawings and the structural drawings.

Cover is the most common problem, and the most consequential for durability. Too little cover lets moisture and chlorides reach the steel and start corrosion; too much reduces the effective depth and the element's strength. After the concrete is placed, cover can be checked with an electromagnetic cover meter, which estimates the depth of bars below the surface. The cover problem found at a podium wall, mentioned in lessons 1 and 5, was confirmed this way.

# Cover meter survey of podium wall W-7 after stripping (illustrative)
# Specified cover 1 1/2 in; placing tolerance -3/8 in for this element
spec, tol = 1.5, 0.375
readings = [1.7, 1.6, 1.5, 1.4, 1.1, 1.05, 1.2, 1.55, 1.65, 1.75, 1.3, 1.1, 1.0, 1.45, 1.55, 1.7]   # in, grid of readings
low = [r for r in readings if r < spec - tol]
print(f"readings: {len(readings)}, minimum {min(readings):.2f} in, mean {sum(readings) / len(readings):.2f} in")
print(f"below {spec - tol:.3f} in: {len(low)} readings -> {', '.join(f'{r:.2f}' for r in low)}")
print(f"share below limit: {len(low) / len(readings):.0%}")

readings: 16, minimum 1.00 in, mean 1.41 in
below 1.125 in: 4 readings -> 1.10, 1.05, 1.10, 1.00
share below limit: 25%

A quarter of the readings were below the allowed minimum, concentrated in one area of the wall where the outer mat had been pushed toward the form by a bundle of conduit tied to it. The engineer accepted a repair using a specified protective coating on the exterior face, which faces the parking garage ramp. The initial phase that should have caught the pushed mat before the pour had been done as a quick look rather than a detailed check, which is why lesson 5 makes so much of doing it properly.

Post-tensioning inspection

Post-tensioned slabs add their own checks. Before the pour, tendon profiles are checked against the shop drawings at marked points, usually by measuring chair heights; anchorages are checked for correct placement and support; and sheathing is checked for damage, which would allow the tendon to bond to the concrete or to corrode. After the concrete reaches the stressing strength, each tendon is stressed with a calibrated jack, and the measured elongation is compared with the elongation calculated by the PT supplier's engineer.

Elongation is the main field check that the tendon has the intended force, because friction along the tendon reduces the force at the far end and a blocked or damaged tendon elongates less than expected. Specifications commonly allow measured elongations to differ from the calculated values by a small percentage, often about 7 percent for individual tendons, with any tendon outside the tolerance reported to the engineer. Readers should use the tolerance in their project's specification.

# Stressing record check, transfer slab pour 1 (illustrative; tolerance per specification)
tol = 0.07
tendons = {  # tendon: (calculated elongation in, measured in)
    "T01": (7.00, 7.15), "T02": (7.00, 6.90), "T03": (7.12, 6.50), "T04": (7.12, 7.31),
    "T05": (6.94, 7.05), "T06": (6.94, 7.52), "T07": (7.20, 7.13), "T08": (7.20, 7.25),
}
for t, (calc, meas) in tendons.items():
    dev = (meas - calc) / calc
    flag = "ok" if abs(dev) <= tol else "REPORT TO ENGINEER"
    print(f"{t}: calculated {calc:.2f} in, measured {meas:.2f} in, deviation {dev:+.1%}  {flag}")

T01: calculated 7.00 in, measured 7.15 in, deviation +2.1%  ok
T02: calculated 7.00 in, measured 6.90 in, deviation -1.4%  ok
T03: calculated 7.12 in, measured 6.50 in, deviation -8.7%  REPORT TO ENGINEER
T04: calculated 7.12 in, measured 7.31 in, deviation +2.7%  ok
T05: calculated 6.94 in, measured 7.05 in, deviation +1.6%  ok
T06: calculated 6.94 in, measured 7.52 in, deviation +8.4%  REPORT TO ENGINEER
T07: calculated 7.20 in, measured 7.13 in, deviation -1.0%  ok
T08: calculated 7.20 in, measured 7.25 in, deviation +0.7%  ok

Two tendons fell outside the tolerance. T03 elongated less than expected, which can mean extra friction or a blockage; T06 elongated more, which can mean a measurement error, a slipped wedge or a broken strand. Both were reported to the structural engineer and the PT supplier before any tails were cut, which is exactly why the transfer slab ITP has a hold point there. On investigation, T03's low elongation was traced to a profile that had been placed with an extra kink near a sleeve, and the engineer accepted it after checking the slab with the reduced force; T06 had been measured from the wrong reference mark and was re-measured within tolerance. Neither question could have been answered after the tails were cut and the pockets grouted.