eg1776
Geotechnical
- Mar 14, 2008
- 8
Has anyone checked ILLICON 2016 against exact solutions? I am writing a user's manual for my finite element consol program that I wrote several years ago, and was verifying my current version with Hamilton Gray's 1945 solution for 3 cases. However, I was also trying to verify that I understood how to use ILLICON 2016 and developed three data files to check its use before I recommended the program to Parsons. ILLICON gives very wrong results when checked against classical Terzaghi and two layer (Gray, 1945) solutions.
The first figure attached here shows how ILLICON fails the simple check against the Terzaghi classical solution:
I have attached another figure that compares the degree of consolidation with time from Gray 1945, from my finite element program, and from ILLICON. the 3 cases use the following properties:
Hamilton Gray 1945 Check for a two layer system of 10 feet thick each.
Case 1 A has both top and bottom draining.
Case 1B has top draining and bottom impervious,
and Case 1C has the top impervious and the bottom draining.
The ILLICON results are in solid line colors of blue CASE 1a and Red CASE 1c. CASE 1b looks to plot over CASE 1c with an orange curve even though I used a different BC condition.
Third, I compared ILLICON and the D-Settle code (developed by the Dutch) for a live field test conducted in the Texcoco clay of Mexico City and the results from my Finite Element code TEXCOCON and measured field settlements with time. Both ILLICON and D-Settle give spurious solutions and apparently would predict infinite settlement at infinite times. See the attached comparison plot.
My conclusion is that ILLICON should not be used as it fails the most basic verification problems. I talked to Dr. Mesri in three different emails and conversations about this 6 months ago to alert him of this problem. As far as I know no fixes have been made to the code to correct these fatal errors. Furthermore, both D-Settle and ILLICON give wrong predictions of settlement at Mexico City's Texcoco test site. By contrast the FE code I wrote many years ago matched the measured data on the first interpretation of the soil layering and properties. No iteration or adjustment of properties or boundary conditions was conducted.
The first figure attached here shows how ILLICON fails the simple check against the Terzaghi classical solution:
I have attached another figure that compares the degree of consolidation with time from Gray 1945, from my finite element program, and from ILLICON. the 3 cases use the following properties:
Hamilton Gray 1945 Check for a two layer system of 10 feet thick each.
Case 1 A has both top and bottom draining.
Case 1B has top draining and bottom impervious,
and Case 1C has the top impervious and the bottom draining.
The ILLICON results are in solid line colors of blue CASE 1a and Red CASE 1c. CASE 1b looks to plot over CASE 1c with an orange curve even though I used a different BC condition.
Third, I compared ILLICON and the D-Settle code (developed by the Dutch) for a live field test conducted in the Texcoco clay of Mexico City and the results from my Finite Element code TEXCOCON and measured field settlements with time. Both ILLICON and D-Settle give spurious solutions and apparently would predict infinite settlement at infinite times. See the attached comparison plot.
My conclusion is that ILLICON should not be used as it fails the most basic verification problems. I talked to Dr. Mesri in three different emails and conversations about this 6 months ago to alert him of this problem. As far as I know no fixes have been made to the code to correct these fatal errors. Furthermore, both D-Settle and ILLICON give wrong predictions of settlement at Mexico City's Texcoco test site. By contrast the FE code I wrote many years ago matched the measured data on the first interpretation of the soil layering and properties. No iteration or adjustment of properties or boundary conditions was conducted.