Expansive Clay & Foundations July 1, 2015
This subject has received attention from many sources for many years, but nevertheless remains a hot topic with regard to design and construction of lightly loaded, shallow foundations in Texas. The prime design problem for these foundations is swelling or expansive clays. A shallow foundation is typically found under a residence or small to mid-size commercial structure. These structures generally apply a load to the surface of the ground of between 150 to 450 pounds per square foot. This is very light loading from a geotechnical and foundation engineering point of view. Such loadings provide almost no restraint to the expansive clay activity, whereas a more heavily loaded foundation would tend to restrain swelling movements. Such movements are typically not uniform and can result in damage to the structure. Certain types of structures with composite foundations using deep piers and soil supported slabs can also be seriously damaged by expansive clays.
Examples of environmental factors include weather such as heavy rains or lack of rain. These can be short term or long term cycles. Moisture attempts to move under the foundation from the outside during wet weather and a wetting front advances slowly underneath the foundation toward the interior. The portion that is wetted expands, whereas the portion that is not yet wet will not expand, causing differential support under the foundation. Similarly in dry weather, assuming the entire underslab soil is at a uniform moisture content, the edges can be dried out causing loss of edge support due to shrinkage.
Other environmental factors could include tree roots extending under the foundations withdrawing moisture and causing the resulting shrinkage phenomena. To some extent, change in temperature and interruption of normal capillary moisture rise and evaporation from the surface also play a part.
A typical example of an artificially induced water condition is a leaking plumbing pipe underneath the foundation. Other examples include standing water such as from excessive watering of a nearby planter, or adverse drainage grades on the surface which do not permit ready shedding of storm water.
When differential volume change occurs under shallow soil supported foundations, the support conditions change from uniform to non-uniform. It is then that the stiffness of the foundation elements become important by bridging the soil waves with adequate structure. The function of any foundation is to remain plane enough under differential support conditions such that the superstructure is not adversely affected.
For some expansive clays, the differential support condition can be quite severe with non-uniform expansions of six to eight inches. For such soils very stiff and strong soil supported slab foundations are necessary. In extreme cases and where the sensitivity of the superstructure and the non-compact shape of the foundation footprint dictate, it may be necessary to utilize a structural floor system with drilled and underreamed piers. This system requires maintaining a complete void space beneath all grade beams and slabs of the foundation and isolating all soil supported elements such as steps, patios, sidewalks or driveway slabs. The entire floor structure is carried on a pier system, which is established at a depth which will remain stable even though the near surface soils move around. Adequate reinforcing of such piers is also necessary. These foundations can be very expensive compared to a soil supported foundation. Three to four times as much cost may be involved.
Methods for predicting and analyzing the differential movements of expansive clay soils include empirical or rule of thumb methods, pressure-volume change relationship predictions, and the more recently introduced soil suction profile variation methods. In the first category of empirical methods might be predictions based on the “PI” such as utilized by the BRAB #33 report. This method utilizes the Atterberg limits, specifically the Plasticity Index, to estimate a support condition relative to a climatic condition. This support condition in turn drives the structural design for stiffness and strength of the shallow slab. The original BRAB #33 procedure permits a weighted average of the PI with the highest weight going to the near surface soils. This is a typical method currently utilized for analyzing subdivision soils for residential slab design.
The second technique uses pressure-volume change relationships derived through laboratory testing of undisturbed samples. This method determines what the potential expansion might be under certain pressure regimes in the ground and assuming unlimited water is made available to the clay soils.
A related technique is the so called “PVR” or potential vertical rise method of the Texas Highway Department. This technique was developed by Chester McDowell and others 30 to 40 years ago and is widely utilized by the Texas geotechnical community today as a method for predicting potential swell. The method relies on Atterberg limits to predict the swell based on extensive correlation with pressure-swell tests. The PVR technique does not always predict accurately (sometimes overpredicting and sometimes underpredicting the actual swell by as much as 100%). It is nevertheless a widely used and convenient method for predicting and categorizing a particular site with regard to expansive clay activity.
The third procedure coming into use is the evaluation of the change of soil suction profiles. This method may be used to either explain what has happened or to predict what might happen in the future under different conditions of drainage, tree root conditions, or typical drought or wet cycle. In this technique, a profile is established of the actual soil suction values under the proposed or existing foundation. A scenario is developed which would change this profile to different suction levels and formulas are applied which will produce predicted vertical movement, either shrink or swell.
This procedure is difficult to implement because of the necessity of obtaining soil suction values and volume change parameters through laboratory tests. Sufficient research has been done to demonstrate what the equilibrium values might be for various localities and conditions. This procedure fairly accurately models the physical phenomena of what actually is occurring under the foundation and accounts for all the variables. The difficulty comes in supplying the actual values of these variables in a given situation. The problem is actually quite complex and involves analysis of partially saturated flow interacting with the volume change properties of the soil and water matrix.
For those into mathematics, soil suction is typically expressed in pF, which is the logarithm of the column of water in centimeters that could be supported by the level of suction in the soil. Extremely high numbers are possible and users should not be concerned that a column of water will not be self supporting at the high suction values indicated. At the very small capillary interfaces in the water-clay-gas three phase system such extremely high suction values are possible. Typical values of pF range from about 2.0 for a very saturated soil condition up to about 4.5 for a very dry condition, which the agriculturists would call the wilting point. At the wilting point tree roots can no longer pull water out of the ground.