The concrete slab-on-ground is among the most performance-critical structural elements in industrial and commercial buildings, yet it has long been designed using methods whose fundamental limitations are explicitly acknowledged by the profession's most common reference. ACI 360R-10, Guide to Design of Slabs-on-Ground, notes that the three historically common design methods, namely American Cement Association, Wire Reinforcement Institute, and U.S. Army Corps of Engineers," are based on Westergaard's work and the assumption that slabs are always fully supported by the subgrade," and that "they give erroneous results for slab thickness where the slab is not in contact with the subgrade." ACI 360R-10 further notes that upward curl is the expected condition for indoor slabs, not an exceptional one. These three historically common design methods have no mechanism to account for curl, nor for load transfer efficiency (LTE), or the capacity of a slab-on-ground.
To address these gaps, a modern design framework was developed for concrete slabs-on-ground, described in Mechanistic-Empirical Design of Concrete Slabs-on-Ground. The framework builds on ISLAB2000 finite element analysis, which retains Westergaard's elastic foundation principles while explicitly modeling slab curling via an equivalent built-in temperature differential (EBITD), and including complications like load transfer at joints. A surrogate neural network trained on over 1.3 million ISLAB2000 analyses compresses FEA-level stress prediction to milliseconds, making iterative design optimization practical for everyday engineering workflows. The framework also replaces the beam-derived allowable stress used in traditional methods with a calibrated slab capacity model, correcting a long-standing and consequential inconsistency that the capacity of a concrete slab-on-ground is greater than that of a beam test. Implemented in a design tool, this framework makes it possible, for the first time in a practical design environment, to compare seemingly very different slab-on-ground types on a consistent mechanistic-empirical basis — under identical loading and support conditions, with curling, loss of support, and load transfer all explicitly accounted for. The following design examples do exactly that.
Figure 1. Examples of 100 ft.+ (30m+) joint spacing using shrinkage compensating concrete in a Home Depot distribution center.CTS Cement - Komponent Division
The following design examples walk through how different slab types account for curling, loss of support and load transfer.
Common Design Assumptions
A 172,800 sq.ft. (16,050 square meter) facility with 60 ft. x 60 ft. (18 m x 18 m) column spacing and a 3:1 width-to-length ratio is used as the common design scenario across seven common slab-on-ground types.
The loading includes back-to-back racks and a reach lift truck with the following details:
- Racks: 6 racks high with 2,000 lb. (910 kg) per pallet and 6 in. (15 cm) square base plates. An additional pallet sits on the ground, making the system 7 pallets tall, for a building clear height of 36 to 40 ft. (11 to 12 m). Seismic conditions are not considered; shelf utilization is 80 percent, resulting in post loads of 9,600 lb (4,350 kg). In this example, the allowable stress ratio is 0.8 for a stationary load, and a static k-value is used.
- Lift truck: A 3-wheeled reacher with small, hard wheels carrying 3,000 lb (1,360 kg). The allowable stress ratio is 0.45 for infinite load repetitions per ACI 360R-10 Table 5.3, and a dynamic k-value is used.
The common concrete slab-on-ground design variables used are:
- Support: A subgrade resilient modulus of 5,800 psi (40 MPa), correlating to a dynamic k-value of 150 psi/in. (4.15 kg/cm3) and a static k-value of 75 psi/in. (2.05 kg/cm3). Because a 15-mil vapor barrier is used, a friction coefficient of 0.50 is used per ACI 360R-10 section 10.3.2.
- Concrete: A 90-to-28-day strength ratio of 1.1 and coefficient of thermal expansion of 4.5 x 10-6/° F (8.1 x 10-6/° C).
- Thermal: A concrete setting temperature of 100° F (38° C) and a mean annual temperature of the facility of 70° F (21° C).
Design of Slab-on-Ground Alternatives
Seven common slab-on-ground types are considered. Types 1, 2, and 3 are all traditionally jointed per Fig. 6.6 of ACI 360R-1; type 4 extends joints to column lines; and types 5, 6, and 7 extend joints to 120 ft. (36), which is two column spacings in this example. Figure 2 compares the length of joints needed for each type and serves as a visual indicator of the trend towards extended joint spacing to mitigate common joint performance concern.
Figure 2. Extended joint spacing dramatically reduces total joint length.The Slab Designer LLC
Additional details on each of the seven slab types considered:
- Unreinforced – while dowels are included in construction joints, joints are expected to open approximately 0.13 in. (0.34 cm), resulting in less than 4 percent LTE in sawcut joints. The concrete mix is assumed to be of typical low-end specifications, with relatively high shrinkage and slab curl.
- Strategically Reinforced – plate dowels are used in all joints, with the LTE being increased to 90 percent by design in both construction and sawcut joints.
- Enhanced Aggregate Interlock – while dowels are included in construction joints, sawcut joints have 0.1 percent steel through them, providing 37 percent LTE.
- Extended Joint with Macrosynthetic Fibers + Shrinkage Reducer – joints are extended to the 60 ft. (18 m) column spacing. This combination reduces concrete shrinkage and slab curl and stresses while also increasing slab capacity.
- Crack Width Control with Steel Fibers – joints, spaced at 120 ft. (36 m), are armored because an expectation of up to 1 in. (2.6 cm) of joint opening, and slab curl is expected to be between that of typical concrete mixtures and those with SRAs.
- Shrinkage-Compensated Concrete (ShCC) – also at a joint spacing of 120 ft.(36 m), but slab curl is reduced while concrete strength and abrasion resistance are increased. Compared to Type 5, joint opening is reduced about 40 percent.
- ShCC + Macrosynthetic Fibers – as with the ShCC slab-on-ground but with an increase in slab capacity due to the use of macrosynthetic fibers.
The varying design parameters for each of the seven alternates are summarized in Table 1 (below).
Table 1. Varying Design Parameters Across the Seven Slab-on-Ground Types EvaluatedThe Slab Designer LLC
Table 2 (below) presents the minimum required concrete slab thickness for each type alongside key performance notes that reflect the combined effect of structural loads, curling, reinforcement/jointing/doweling strategy, and concrete strength.
Table 2. Minimum Concrete Slab Thickness and Serviceability NotesThe Slab Designer LLC
Key Trends & Takeaways
These seven design examples, evaluated under identical loading and support conditions, reveal several consistent trends that hold implications for engineers, specifiers, owners, contractors, and prospective tenants alike.
For Owners & Tenants
Slab thickness is a poor proxy for slab capacity. Unreinforced, at 10 in. (25 cm), is the thickest slab in this comparison and has the worst serviceability profile. Type 7, at 5 in. (13 cm), is the thinnest and performs best. Notably, four of the seven types (4 through 7) converge on essentially the same minimum thickness despite employing very different strategies; among those four, the differentiator is serviceability and joint maintenance burden, not slab thickness. Thus, specifying or evaluating a slab based on thickness alone, without knowledge of joint design, load transfer, and curl, provides no reliable information about what loads the slab can carry or how it will perform over time.
Contractors should not accept responsibility for joint spalling on under-designed floors without positive load transfer in joints subjected to lift truck traffic.
These findings carry practical implications beyond the design office. Prospective tenants evaluating industrial facilities should request documentation of the slab's design loads and joint details because thickness alone is not sufficient to assess fitness for purpose. ACI 360R-10 provides clear guidance on allowable differential deflection across joints: 0.010 in. (0.25 mm) for small hard wheels and 0.020 in. (0.51 mm) for large rubber wheels. A facility whose joints cannot meet these thresholds under the tenant's lift truck traffic will require ongoing maintenance regardless of slab thickness.
A Note To Contractors
This advancement in design provides contractors with a valuable tool to protect themselves by ensuring that the designs they construct are adequate for the specified use. Contractors should not accept responsibility for joint spalling (see Figure 3) on under-designed floors without positive load transfer in joints subjected to lift truck traffic.
Figure 3. Joint spalling due to the impact of small, hard wheels when there is high differential deflection of the joint (e.g., no dowels).Metzger McGuire, 2025
For Engineers & Specifiers
LTE is a high-leverage design decision. The difference between unreinforced at 10 in. (25 cm) and strategically reinforced at 7 in. (18 cm) is entirely attributable to LTE because they have the same joint spacing, concrete, and curl, but strategically reinforced is a 30 percent thinner slab. Dowels are among the most cost-effective line items in a slab specification, especially when the dowel design is optimized using the enhanced integrated dowel model (EIDM).
Curl governs more than most designers assume. Types 1 through 3 all share the same conservative EBITD of -25° F (-14° C). Reducing curl, through SRA, ShCC, or macrofibers, is the primary mechanism driving thickness reductions in Types 4, 6, and 7. A slab that curls less carries more load; a slab that curls more requires a thicker slab to compensate.
Every joint is a potential maintenance liability under lift truck traffic so reducing the amount of joints reduces that exposure.
Fewer joints typically means lower long-term maintenance costs; this is why the slab-on-ground market has moved towards reducing joints. Types 4 through 7 have 78 to 91 percent fewer joints than traditionally jointed concrete slabs-on-ground. Every joint is a potential maintenance liability under lift truck traffic so reducing the amount of joints reduces that exposure. For facilities with small, hard-wheeled equipment such as reachers and order pickers, this matters most, as differential deflection limits are tightest and joint edge damage accumulates fastest.
More recently developed slab types support tighter flatness and levelness. Lower curl directly improves the ability to achieve and maintain Ff and Fl tolerances. Despite a longer joint spacing, Types 4 through 7 achieve higher tolerances than Types 1 through 3, which is especially relevant in narrow-aisle and very-narrow-aisle racking environments where defined-traffic Fmin values may govern (Figure 4).
Figure 4. Extended, column line joint spacing for a very-narrow-aisle facility.Lampasona Concrete
Limitations & Opportunities for Future Developments
This comparison reflects one specific combination of loading, support, concrete, and environmental inputs. Seismic conditions were explicitly excluded in these design examples; in seismic zones, amplified rack post loads can govern slab thickness and alter the relative performance of the different slab types. Make sure to consider local building codes.
The comparison presented here illustrates the direction and magnitude of trends across common slab-on-ground types, but they are not universally transferable.
Another related practical constraint worth noting is anchor bolt embedment: slab thickness must accommodate the required bolt length plus clearance, commonly bolt length plus 1 in. (2.5 cm), meaning that rack design and bolt specification can set a minimum thickness floor independent of structural load calculations, and should be coordinated early in the design process.
The comparison presented here illustrates the direction and magnitude of trends across common slab-on-ground types, but they are not universally transferable. Each project warrants its own analysis with site-specific inputs for subgrade, curl, concrete strength, and enhancements, etc.
Good curl data for local conditions is not always available. The EBITD values used here represent informed estimates based on the assumed concrete mix, curing, and environment; field measurements can differ. Systematic collection of curl number data from existing facilities in similar environments, as described in Collecting Curl Data to Improve Concrete Slab Designs, will allow engineers to calibrate EBITD assumptions to local reality and reduce conservatism in future designs.
Finally, this analysis did not optimize dowel size and spacing independently; it assigned LTE targets.
