It could cost you an extra bag of cement to hit the required strength for a new job when mix proportions are based on lab trial mixes. Our analysis of lab vs. field strengths for 84 projects from two state departments of transportation shows what most producers already know: Average strengths of production concrete often fall below required average compressive strengths based on laboratory trial batches. This strength shortfall isn't accounted for by the ACI 318 overdesign factor and can easily reach 700 psi for a 4000-psi design strength.The data we've collected gives some guidance for estimating the strength shortfall. Then it's up to you to decide how much more cement you'll need, and bid accordingly.When concrete producers have test records to establish their standard deviation for compressive strength, the required average compressive strength can be calculated as follows: f'cr = f'c + 1.34s f'cr = f'c + 2.33s - 500 where: f'cr = required average compressive strength, psif'c = specified compressive strength, psis = standard deviation, psiThe larger of the two calculated values is used as the required average compressive strength. Then the producer selects mixture proportions that will produce that average strength, based on a suitable experience record or trial lab mixes.The details and rationale for this procedure are described in Building Code Requirements for Structural Concrete, ACI 318-95. The procedure is based on statistical methods intended to provide a probability no greater than 1 in 100 that:

• Averages of any three consecutive test results will be below the specified strength, f'c
• Any individual test result will be more than 500 psi below f'c
The overdesign value (f'cr minus f'c) is intended to yield a mix that provides these probabilities.A different approach is used to establish concrete proportions from trial mixtures when data aren't available to establish a standard deviation. The required average compressive strength is chosen from a table. Then proportions needed to produce that strength are determined in the laboratory.The overdesign values established from field records or taken from the table may seem conservative, but they don't account for differences between strengths of field-produced and lab-produced concrete. They only account for the variability of field-produced concrete as indicated by standard deviation. Failure to account for lower average strengths of field-produced concrete increases the probability of test results falling below f'c.How much of a shift can be tolerated before unacceptably low strength-test results occur? For specified strengths between 3000 psi and 6000 psi, this depends on the producer's standard deviation.But what is a reasonable estimate for the difference between field-produced and lab-produced concrete? And does the strength difference vary with the required strength of the lab concrete? To seek answers for these questions, we analyzed strength-test results supplied by the Colorado and Wyoming departments of transportation and Blue Circle Williams.

CDOT requires contractors to submit a lab mix with a strength that equals or exceeds the specified compressive strength by 25%. The mix proportions of the approved lab mix are then used in the field. We analyzed field-test data from 1992 through 1997 for 66 CDOT projects in which ready-mixed concrete was used.WDOT doesn't have a strength requirement per se. Its specification gives a minimum cementitious materials content and a maximum water-cementitious materials ratio for each class of concrete. Using samples of proposed materials submitted by the contractor, WDOT personnel proportion the mix and determine strength in the laboratory. These mix proportions are used in the field, and laboratory strength becomes the target average strength for field-produced concrete. We obtained field-test data for 18 WDOT projects in which ready-mixed concrete was used. Strength requirements for the Blue Circle Williams projects were determined in accordance with ACI 318-95.For all of the projects, cylinders were made and initially cured in the field, but were then taken to the lab within 48 hours for further curing and testing.For 70 of the 84 projects analyzed, average strength of the field-produced concrete was less than the required average strength based on lab-produced trial batches. Did strength differences vary with lab-batch required strength? There seems to be a trend toward greater differences between lab and field strengths at higher required average strengths, starting at 5000 psi. To further study the relationship between strength difference and required strength, we also grouped the data in 500-psi required-strength increments for the 70 projects in which required strength exceeded field strength. The data suggest two interesting observations:

• At required strengths higher than 5500 psi, field strength is much more likely to fall short of lab strength.
• Also, at required strengths higher than 5500 psi, the discrepancy between lab and field strengths is likely to increase substantially.
Based on our data from the 87 projects, strength differences between field-produced and lab-produced concretes with specified strengths above 4000 psi often exceed 700 psi. How should the producer adjust mixes to account for these lab- field strength differences? One option is a procedure similar to the one in ACI 211.4R: When selecting high-strength concrete proportions based on laboratory trial batches, determine the required average strength f'cr from the equation: f'cr = (f'c + 1400)0.90 For specified strengths between 4000 and 5000 psi, use 1200 instead of 1400 psi in the numerator, and use a lab-to-field factor in the denominator between 0.80 and 0.90 as indicated by your experience. For 6000-psi concrete use 1400 psi in the numerator.

keywords: cylinder test, compressive strength, overdesign