In February 2009, Concrete Construction magazine and Scurto Cement Construction Ltd., Gilberts, Ill., joined forces to conduct a field study focused on curling and shrinkage in warehouse floor slabs. Additionally, CC brought together a task group to work out the details of the study and develop the concrete mixes that would be used, while Scurto offered to make the floor available for measuring curling for a two-year period. The 60,000-square-foot floor was placed over a vapor barrier during a three-day period starting on Feb. 10, 2009. Scurto's workforce placed 20,000 square feet each day.
There were several reasons for conducting a study about curling in a setting (“real-crete”) rather than in a controlled laboratory setting (“lab-crete”). In a laboratory, beams are cast without restraint to the top or the bottom of the beam, measuring shrinkage more accurately than curl. In a real-crete situation, concrete is mixed by a ready-mix producer with the many variables common to a batch plant, hauled as long as 40 minutes in a concrete truck, and placed under conditions common to jobsites—in this case on a vapor retarding membrane. In a real setting, with less than perfect control of variables, how would the concrete mixes used perform?
The Bartlett, Ill.-located warehouse is a development project for sale or rent. However, due to the economy, the unit has remained empty for the past year. So the floor surface was profiled and the relative humidity (RH) of the concrete recorded several times in a fully enclosed environment where few people visited. The ambient temperature during the summer months averaged 70° F with RH in the 75% range and during the winter month's room temperatures ranged between 40° F to 50° F and RH readings averaged 50%. Given these ideal conditions for concrete, floor panels profiled three months after placement showed some curling activity at panel edges, but after six months, it was discovered the previous curling at panel edges had relaxed, causing panels to be flatter than they were three months prior. Now, one year after placement, panels are again curling at their edges—probably as a result of recent additional heating of the building and a resulting drop in RH to accommodate a business moving into a portion of the space.
At the time of placement, samples of fresh concrete were taken at regular intervals to check air content, record placing slumps, and cast cylinders. The cylinders were broke at 7 and 28 days to determine compressive strength. Figure 1 shows the test results.
On the morning after each placement, a few hours after control joints were cut, a 3-D laser scanner and a D-Meter were used to profile the previous day's placement to establish baseline information about the installed flatness of the floor. Profiling measurements were taken again after three months, six months, and one year. Each reading was compared to the initial profile to account for any surface flaws due to placing and finishing. Each time the floor was profiled, construction-joint width measurements also were collected along the east-west axis of the floor.
Monitoring moisture movement
Howard Kanare, a member of the project's task group and a senior principal scientist with CTLGroup, Skokie, Ill., has suspected for a long time that the hardness of a floor finish is related to how fast a slab is able to dry out. To test this idea, he installed several finishes in the enclosed, temperature-controlled sprinkler room of the building. The finishes include a bullfloat finish, handtrowel finish, hard-trowel finish, and a bead-blasted area. Then he embedded RH probes in each area at ½-, 1-, 2-, and 4-inch depths to record the RH at different slab elevations.
RH probes also were placed at 2.4-inch depths in each mix design area in accordance with ASTM F 2170 requirements for 6-inch-thick floors. Figure 2 shows readings taken on Jan. 22.
No one besides the project task group hopes that a floor will curl, certainly not the owner of the warehouse. But the primary goal for this field study is to measure curling differences between concrete mix designs. However, at the end of one year, there have been only small amounts of curling present in the panels due to ideal ambient conditions. Measurements over time all point to this conclusion—profiling the floor surface, measuring joint width, slab RH and temperature measurements, and inside warehouse ambient RH and temperature records. However, several things have been learned up to this point.
Measuring curl. Measuring curl is easier and more accurate to measure than shrinkage in a floor. Shrinkage manifests at joints and cracks, which also curl, so top widths are different than bottom widths. Measuring joints and cracks captures both shrinkage and curling.
Curling changes the surface elevation of concrete panels so a 3-D laser scanner can be used to graphically look at the entire floor and a D-Meter can measure elevations to the nearest 0.001 inch of selected panels. Floor flatness (FF) measurements are stopped 1 foot from panel edges, but for the purpose of the field study, the center panel of a bay (a panel surrounded by sawed control joints) was measured diagonally with the D-Meter all the way to the sawed joints (the place where panel elevations are highest).
Sherical curling. Unlike the D-Meter, 3-D laser scans also plot the elevation at the center of a panel, which typically falls below the bench elevation as perimeter elevations of a panel rise (the center of a panel supports the additional weight of the lifted edges). Allen Face, another member of the task group and president of the Allen Face Co., Wilmington, N.C., believes that panel curling takes in a spherical shape and that this radius can be calculated. From the information generated by the study, he was able to confirm this theory.