|Note: a more complete description of the development of this set of on-line routines can be found by clicking here.
Dennis and Olson Method
The Dennis and Olson method is an advance of the API methods for estimating driven pile capacity. A description of the method (with worked examples) is shown here.
This is an on-line demonstration of the method; it is not intended for use in actual design, but for educational purposes only. There are many simplifications employed to make this easy to use. Some of them are as follows:
- The soil is assumed to be homogeneous.
- The pile is assumed to have a uniform cross section all along its length, both for cross sectional area and perimeter. There is a separate input for pile toe area (as opposed to cross-sectional area,) which allows the analysis of piles such as closed ended pipe piles.
- The water table is located somewhere between the pile head and the pile toe. If it is input outside of these limits, the program will adjust the location of the water table.
- The soil input is greatly simplified, with limited options.
- At this time calculations are only available in U.S. units.
- A pile database is included for convenience.
It is important to note that pile plugging of open ended pipe piles is considered, using a very rudimentary criterion. If the L/D ratio of the pile is greater than 20 in a true open-ended pipe or concrete cylinder pile driven into a cohesive soil, soil plugging is assumed to take place, and the toe area is changed to the toe area of a closed-ended pile. A more detailed explanation of this can be found in the Soils and Foundations Reference Manual, which can be downloaded.
TAMAXIAL Load Test Simulator
This program attempts to simulate the results of a load test of the pile analysed by the Dennis and Olson Method. Using the same information, it generates a pile head load-deflection curve. It does this by using a finite difference program and generated t-z curves along the pile. The program is based on the PX4C3 routine developed by Harry Coyle (Texas A&M University) and Lymon Reese (University of Texas) in the 1960's. The program uses the given load transfer vs pile movement curves and computes the load settlement characteristics of an axially loaded pile. Load transfer vs pile movement curve is designated as t-z curve. The program interpolates point bearing values corresponding to a given toe movement from a previously input point bearing vs toe movement curve. The program then solves the pile head deflection using a finite difference technique.
Since we do not have actual load-deflection results or experimentally determined t-z curves, we used an eclectic collection of t-z curves from research data. These can be found in the Theoretical Manual for Pile Foundations. From this document, we used the SSF4 and SF5 methods for pile shafts and toes (respectively) in sand and CSF1 and the method of Ashenbrener and Olson (1984) methods for pile shafts and toes in clay.
Once these load-displacement curves are generated, methods such as Davisson's methods can be used to determine the pile capacity, which in turn can be compared with the results of the Dennis and Olson method.
TAMBENT Pile Group Simulator
This set of routines computes the horizontal and vertical deflection and rotation of a group of piles joined together by a pile cap. It also computes the lateral deflections and moments on piles in the group. It is based on the BENT1 program written by Dr. Frazier Parker, Jr. at the Waterways Experiment Station in Vicksburg, MS under the guidance of Lymon Reese and Hudson Matlock. The report by F. Parker and N. Radhakrishnan (1975), "Background Theory and Documentation of Five University of Texas Soil-Structure Interaction Computer Programs," Miscellaneous Paper K-75-2, discusses the theory and application of the BENT1 program and its subroutines in detail.
One of the subroutines for BENT1 was COM62, which is the direct ancestor of the COM624 program that made lateral pile loading analysis practical using the p-y curve method. COM62 uses a finite difference routine to analyse the effects of lateral loading of each pile in the group. The results are then applied to the entire structure. As this is a non-linear phenomenon, the analysis is cyclical, thus the results for the entire analysis are iterative, as can be seen in the results. We have adapted BENT1, COM62 and the subroutine MAKE that generates the p-y curves to php. The load-settlement curves for the piles are generated using the TAMAXIAL program described above.
In addition to the data entered for the single piles, it is necessary to enter the information for each pile location. TAMBENT is a two-dimensional analyser, but multiple piles at a given location are permitted. The chart below shows the geometry used for the tabular entry at the bottom of the page. Please note the units carefully.
BENT1/COM62 was a pioneering effort; however, the following should be noted about the input and results of the analysis:
- Piles are always assumed to be driven to grade, to have uniform cross section, material, and diameter/size throughout their length.
- The soil is homogeneous except for the presence of the water table. If the water table falls between the head and toe of the pile, the program assumes there are two soil layers to account for the variation in effective unit weight and effective stress. Only one soil profile is used per analysis.
- The "load centre" of the horizontal load, vertical load and moment on the pile cap is the geometric centre of the system and the pile head locations should be taken from this point.
- A single pile lateral analysis is not possible because of the way the equations are solved. For this problem, a better solution (in every respect) is the CLM2 spreadsheet, which can be downloaded from here. (The COM624P program, a direct descendant of this routine, can also be found in the same place.
- The p-y curves in this program are always generated, based on the theory shown in the paper above. Work since then has vastly advanced our understanding of the actual profile of p-y curves from what is shown in this program. The program always generates six p-y curves for each pile, one at the head, one at the toe, and the other four evenly spaced in between. p-y curves for segments between these points are found through linear interpolation. This of course is strictly for teaching purposes; for anything beyond very basic instruction and certainly for analysis of real piles, one should use a program such as COM624P or LPILE.
- No lateral or axial group effects are included in this analysis. No seismic or cyclical loading can be analysed by this routine.