Working with Pile Foundation Analysis and Design using the Poulos and Davis Method
I spent a few years working through pile foundation problems using the approaches from Poulos and Davis before moving more toward numerical modeling. Their book, "Elastic Solutions for Soil and Rock Mechanics," along with the subsequent papers on pile groups, gives you a fairly complete set of tools for handling single piles and pile groups under both axial and lateral loads. It is still taught in geotechnical programs and used in practice, though with known limitations. The method is built on elastic continuum theory. You model the soil as a homogeneous or layered elastic half-space and the pile as a beam or column with a certain stiffness. The key output you get is how load transfers from the pile to the surrounding soil, how settlements distribute, and how adjacent piles in a group interact with each other. The interaction factors are probably the most widely used part of their work. The main inputs you need are the Young's modulus of the soil, Poisson's ratio, the pile geometry (length, diameter, moment of inertia), the pile modulus of elasticity, and the load case you are analyzing. Everything flows from there through the influence factors and solution charts that Poulos and Davis derived or compiled.
How to Actually Use It in a Design
I will walk through a typical axial pile group settlement calculation because that is where most people run into trouble. The steps are roughly: First, determine the equivalent rigid or flexible pile cap assumption. If the cap is thick and stiff relative to the pile flexibility, treating it as rigid gives a more realistic load distribution. A flexible cap assumption will over-predict settlements in many practical cases. This distinction matters more than people usually give it credit for. Second, calculate the settlement of a single isolated pile under the design load. You can use the standard elastic solution where settlement equals the load divided by the pile stiffness factor, or you can use the more complete expression that includes both tip and shaft contributions. Poulos and Davis provide the full equations with correction factors for embedded depth and diameter ratios.
Third, apply the group interaction factors. This is where the method gets useful and also where errors creep in. You need the interaction factor between each pair of piles, which depends on the L/D ratio, the spacing-to-diameter ratio, the Poisson ratio of the soil, and whether you are looking at axial or lateral interaction. The charts and tables from their papers cover a range of these parameters. If your geometry falls outside the published ranges, you are doing interpolation or extrapolation, and that introduces uncertainty. Fourth, combine the individual pile settlements with the interaction effects. For a rigid cap, the settlement is approximately uniform across all piles, and you solve for the load each pile carries based on its position relative to the centroid and the interaction factors. For a flexible cap, each pile settles according to its own load and the influence from neighboring piles. The computational part is straightforward enough that many people still do it by hand for simple groups or set up a spreadsheet. I have seen full pile group analyses completed in under an hour this way for configurations up to about nine piles. Beyond that, the matrix operations get tedious without automation.
Get the Full Details
A Problem I Ran Into That Was Not Covered in the Textbook
Several years ago I was working on a bridge approach embankment where the pile group was situated above a soft clay layer that was only about three meters thick, sitting on dense sand. The Poulos and Davis method assumes a homogeneous or gradually varying elastic half-space. Our layer was clearly not that. Using the average modulus for the clay gave settlements that were roughly 40 percent too low compared to what we observed later during construction monitoring. The workaround I used was to treat the soft clay layer as a distinct zone with its own settlement contribution calculated separately using a 1D consolidation approach, then add that to the elastic solution for the upper stiffer layer. It is not elegant, and it is not what the original method was designed for, but it brought the predicted and observed settlements into reasonable agreement. If you are dealing with a thin weak layer beneath deeper competent soil, a combined approach like this is worth considering rather than trying to force a single equivalent modulus through the elastic solution.
Where the Method Falls Apart
The biggest limitation is the elastic soil assumption. Real soils are not elastic. They are plastic, strain-hardening or strain-softening, and their modulus changes with confining stress and strain level. The Poulos and Davis method does not account for this. For preliminary design and for cases where the soil stiffness is reasonably well characterized, it gives useful results. For final design on critical structures, you should verify the outcomes against a more refined analysis or monitoring data. Another issue is the treatment of pile installation effects. Driven piles densify the surrounding soil, especially in sands, which increases the actual stiffness compared to what the elastic solution assumes. Bored piles often have a lower interface stiffness due to soil remolding. The method does not include installation effects, and ignoring them can lead to unconservative predictions in dense sands or overly conservative estimates for bored piles in clay. Layered soil profiles are handled somewhat better than fully homogeneous assumptions, but the published solutions are still restricted to a limited number of layer configurations. If your ground profile has multiple thin layers with contrasting moduli, you are likely better off using a finite element program such as PLAXIS or a specialized pile analysis tool like L-Pile or APIR. These tools can model stratigraphy directly and handle nonlinear soil behavior.
Practical Notes for Getting Reasonable Results
The modulus values you choose will dominate the output. A 20 percent error in E leads to roughly a 20 percent error in settlement predictions. Use in-situ test data where possible. SPT correlations, CPT-based modulus estimates, and pressuremeter results all give different values for the same ground. Pick one consistent approach and stick with it throughout the analysis. Poisson's ratio is another input that people tend to take from default tables. For saturated clays undrained conditions are common in short-term analysis, so v = 0.5 may be appropriate. For long-term drained conditions in clays or for sands, values between 0.3 and 0.4 are more typical. Using the wrong value affects the interaction factors more than the individual pile settlement, so getting it right matters when you have a tightly spaced group. The spacing-to-diameter ratio is critical for group interaction. When s/D is less than about three, interaction factors become large and the group efficiency drops significantly. Below s/D of two, the method becomes less reliable because the assumption of non-overlapping stress bulbs breaks down. In those cases, numerical modeling is the safer route.
If you need the original reference material, Poulos and Davis published their foundational work in "Elastic Solutions for Soil and Rock Mechanics" (Wiley, 1974) and followed up with several papers on pile group analysis in geotechnical journals. The interaction factor tables and charts from those sources are still available in most engineering library collections and through academic access platforms. There is no single official download link for the method itself since it is a published analytical approach, not a software package, but secondary sources and textbooks that cite their work often reproduce the key charts and tables. I tend to keep a copy of the original charts in my office files rather than rely on reproductions from later textbooks. The reproduction quality varies, and a slightly misread chart value can shift your group settlement estimate by a meaningful amount on a sensitive project.
Bottom Line
The Poulos and Davis approach to pile foundation analysis gives you a structured way to estimate pile and pile group behavior using elastic theory. It is fast, transparent, and good for preliminary sizing and sanity checks. It is not a substitute for site-specific numerical analysis on complex projects, and it requires careful attention to input parameters, especially soil modulus and layering. If you use it as part of a broader design process rather than the final word, it remains a practical tool.
