Developing and Checking Soil Structure Interaction Models

Published on January 2017 | Categories: Documents | Downloads: 26 | Comments: 0 | Views: 223
of 6
Download PDF   Embed   Report

Comments

Content

Does the Output Mean Anything? Developing and Checking Soil Structure Interaction Models Kevin J. DiRocco, P.E.1 and Michael P. Walker, P.E.2
Project Manager, GEI Consultants, Inc., 1021 Main Street, Winchester, MA 01890; PH (781) 721-4097; FAX (781) 721-4073; email: [email protected] 2 Senior Project Manger, GEI Consultants, Inc., 1021 Main Street, Winchester, MA 01890; PH (781) 721-4057; FAX (781) 721-4073; email: [email protected] Abstract The development of more powerful computers, sophisticated computer models, and analytical programs with easy to use graphic interfaces provides the designer with powerful tools for modeling and evaluating soil structure interaction. However, the complexity of available computer models, and the enormous amounts of data generated by the modeling programs, can make the results of these programs difficult to verify. This paper discusses the selection of soil properties and boundary conditions for staged excavation models. Methods to check the output using such things as simple hand calculations and duplicate models are discussed. A case history is used to illustrate the methods. Introduction In recent years, the development of more powerful computers and more sophisticated computer models and software have dramatically increased our ability to model and to better predict the behavior of deeper and more complex excavation support systems. As the need for more sophisticated models increases, affordable and user-friendly software is becoming available to more design engineers. Much of the commercially available software uses sophisticated graphical interfaces that allow for simple, rapid model construction and for rapid interpretation of the results. However, the ease and speed of software use should not be confused with output validity or accuracy. Before using software the designer must answer the following questions: • What type of excavation support system is appropriate? • What type of computer model is appropriate for the selected excavation support system and site conditions? • Do you understand how the software performs and have you checked the software using test problems? • What are the software input parameters (soil properties, boundary conditions, etc.)? • Are case studies or local experience available that can be used to select the most appropriate model and input parameters? • What type of investigation is needed to obtain the program input parameters? Are special field tests required? Is laboratory testing required? Once the designer has answered the above questions, developed the computer model, and generated output, the designer must perform a detailed review to verify the validity of the results. Reviewing the results can be the most important step in the design process. Reviewing the results is not simply the mechanical process of checking that the input parameters have been correctly entered into the software. The review process should answer the following questions: • Have appropriate input parameters been selected? • How does variation of the input parameters impact the output? • Do the results make sense (reasonable brace loads, wall embedment, movements, etc.)? This paper discusses the process for selecting and developing the model, and verifying the results, and discusses the need for providing construction monitoring to link the predicated performance to the actual performance and to provide additional data for future models. This process will be illustrated using a case history.
1

1

Software Quality Assurance We recommend that either on an individual or corporate basis, quality assurance be performed on any analysis software package prior to its use for design. In the last ten years we have found computational errors in several widely used, commercially available geotechnical software packages. To perform quality assurance on a software package, the engineer must first understand the analytical model used to correctly use the model features in the software. The software output can then be checked by performing analyses on a series of simple quality assurance problems designed to test specific features of the program both individually and in combination. This approach allows the engineer to perform a quality assurance check of the program while developing an understanding of how the software package works. Also, as a secondary check and part of the model validation process, we often compare output from new software packages to software that has already been checked. This will be discussed further in a case history. Problem Definition and Model Selection The first step in the process of selecting and using the appropriate model is to define the problem, which means evaluating the site constraints, selecting the wall type and bracing type for the excavation support system, and developing the construction sequence. Then the appropriate analysis method and/or computer model(s) can be selected. When selecting the analysis method and computer models, consideration must be given to the complexity of the excavation support system and construction sequence, the sensitivity of adjacent structures and utilities to ground movements, the importance of adjacent structures and utilities, and local experience. Simple Empirical Analyses Given the availability of sophisticated computer analysis methods, simple empirical analyses, such as apparent pressure envelopes, are not commonly used as the primary analysis method for critical deep excavations in urban environments. As noted in Pearlman, et.al. (2004), empirical analyses are generally used as upper-bound analyses for flexible wall system design and may not be appropriate for stiffer wall systems, such as concrete diaphragm walls. Use of empirical analysis techniques as the primary analysis methods for excavation support design is generally limited to shallow excavations in areas where excavation support movements are not likely to impact nearby structures and utilities. However, as explained later, empirical analyses are useful tools for evaluating design loads from more sophisticated computer analysis methods and for estimating ground movements when beam on elastic foundation methods are used. Beam on Elastic Foundation Method The beam on elastic foundation method is a numerical method that models the soil and wall elements as a series of inter-related springs and models the bracing elements as either user-defined loads or springs. Using commercially available software, this method can be quickly and easily used for modeling complex, staged excavation/ bracing sequences and can be easily modified to quickly evaluate impacts of changes in excavation/ bracing sequence on the excavation support system design. The beam on elastic foundation method is useful for predicting wall and bracing loads and wall movements. The primary limitation of the beam on elastic foundation method for most applications is that the method does not directly model the behavior of the soil mass behind the wall and does not directly estimate ground movements. However, for most applications, the movement of the soil mass behind the wall can be adequately predicted using empirical relationships. Finite Element Method Finite element models are numerical models that can be developed to model either two- or three-dimensional stressstrain behavior of the excavation support system and surrounding soil mass. Historically, finite element models were difficult to use due to complex soil property and geometry input requirements, but with the development of more user-friendly software, finite element models are becoming more commonly used. Finite element models have the advantage of modeling the behavior of the soil mass surrounding the excavation support system and using the stressstrain relationships of the soil mass to predict ground movements and structure response. However, as noted in Pearlman, et. al. (2004), the finite element method is more sensitive to selection of soil parameters than the beam on elastic foundation method, and the finite element method is most accurate when soil parameters can be calibrated using local case history data.

2

Input Parameters Input parameters may be evaluated at any point in the design process, but generally evaluation occurs in conjunction with problem definition and model selection. Input parameter may be determined from a combination of literature search, previous project experience, subsurface investigations, field testing, and laboratory testing. The effort required to determine input parameter depends on the selected model and analysis method used, the depth and complexity of the excavation staging, the proximity of sensitive structures and utilities, site geology, and local experience. The importance of thorough evaluation of input parameters can be overlooked when using more sophisticated models and when trying to minimize project costs. The design engineer must be familiar with the required parameters for the selected analysis model and must understand the impacts of incorrectly selected or overly conservative parameters. If the designer is uncertain of the impacts of varying a given input parameter, user-friendly software can be used during data collection to rapidly perform parametric studies to evaluate which parameters must be accurately measured or estimated. For beam on elastic foundation and finite element models, the importance of experience, case histories, and local experience cannot be overstated. Review of Results (QA/QC) Reviewing the results is one of the most important steps in the excavation support design process. Reviewing the results is not only the mechanical process of confirming that input parameters have been correctly entered, but is the process of verifying that the appropriate parameters have been selected for the chosen model and that the results “make sense.” The importance of reviewing the results can be overlooked, especially when using powerful, userfriendly software programs. Ease of input and generating output can provide an inexperienced designer with a false sense of security when the software program provides an answer without identifying an error, lending truth to the expression “garbage in is garbage out.” The first step in reviewing the results is to confirm that the selected parameters have been correctly entered into the program. This process can be complicated when using graphically driven programs, which do not provide a detailed summary of the input parameters. For graphically driven programs, we create an input parameter checklist (soil properties, boundary conditions, etc.) that the reviewer uses to either go through the programs input pull down menus to verify that specific parameters have been entered correctly or to review graphical representations of the input in the programs preprocessor to verify that the model has been constructed correctly. The second step in the review of the results is to confirm that the results “make sense.” As with other steps in the design process, the level of review must be commensurate with the complexity of the design. For less complicated designs, review may simply consist of a “gut check” based on local experience or simple hand calculations. However, for more complicated projects, the review process may begin with a “gut check” but will likely include parametric studies, comparison of the design to local case histories, and supplemental analyses. Supplemental analyses may be performed using any of the methods or combinations of analysis methods discussed previously. The use of supplemental analyses will be discussed in the case history. Construction Monitoring Program Regardless of the complexity of the design, a monitoring program should be implemented to ensure the safety of workers and the public and to ensure the protection of sensitive utilities and structures. The results of the monitoring program are also useful for confirming that an appropriate analytical model was selected for your specific project and for providing case history data for future projects. For simple projects, monitoring may consist of daily inspections of the excavation support elements for signs of distress, visual observation of the ground surrounding the excavation for signs of excessive movements, and vertical survey of nearby structures or utilities. For more complicated excavation support systems, a detailed instrumentation program may be required to monitor ground movements, excavation support system movements, brace loads, and structure and utility movements. For complex excavations, the monitoring program should be tied to a mitigation plan to arrest excavation support movements or reduce brace loads if they are greater than anticipated.

3

Case History Background The United States Capitol Visitor’s Center (USCVC) is a three level subterranean addition to the U.S. Capitol Building in Washington D.C. The USCVC has a footprint area of approximately 193,000 square feet, more than doubling the footprint of the U.S. Capitol Building. When completed, the USCVC will house exhibits, food services, theaters, and auditoriums designed to enhance the experience of visitors; contain mechanical facilities, storage, and service tunnels for deliveries; and provide much needed security for the U.S. Capitol Building. Construction of the USCVC required excavation to average depths of about 50 feet with maximum local excavation depths of up to 75 feet. See Figure 1.

Figure 1. Excavation Support Wall at the Capitol Problem Definition and Model Selection The team of Nicholson Construction Company of Cuddy, Pennsylvania and GEI Consultants, Inc of Winchester, Massachusetts designed and constructed the USCVC excavation support system for both temporary and permanent conditions. Concrete diaphragm walls supported by tiebacks were selected for the excavation support system because the diaphragm walls provide excellent ground movement control, provide temporary and permanent groundwater cutoff, and act as the permanent walls for the structure. Due to the complexity of the project and the need to estimate ground movements, GEI elected to design the excavation support system using both the beam on elastic foundation method and the finite element method. GEI selected the commercially available beam on elastic foundation program WALLAP (1997) and finite element program PLAXIS (1998) because of their ability to model complex staged excavations and our experience with these programs on several successful projects. Because of variations in soil stratigraphy and new structure configuration along the wall alignment and the presence of new and existing structures and utilities behind the diaphragm walls, evaluation of numerous design sections was required. We modeled all of the design sections using WALLAP because the program allowed us to quickly build and evaluate complex staged models for each design section. Using WALLAP, we were able to quickly evaluate both wall and tieback design loads. PLAXIS could have been used to evaluate the design loads, but WALLAP models can be constructed and evaluated more quickly and easily than PLAXIS models. The WALLAP program provides wall movement estimates, but does not provide soil movement estimates. Soil movements can be estimated from empirical relationships based on the wall movements obtained from the WALLAP analysis, but better estimates of soil movements can be obtained using the finite element method, As a result, we used PLAXIS to perform a two-dimensional finite element analysis of the most critical design section at the U.S. Capitol Building to evaluate soil movements. PLAXIS soil movement estimates include both horizontal and vertical deformations, and our movement estimates from PLAXIS were used to evaluate the impacts of potential soilmovement-related distortions on the U.S. Capitol Building. Also, WALLAP wall movement estimates are generally conservative because of lower-bound soil modulus values selected to provide better estimates of wall and tieback design loads, and the PLAXIS output was used to confirm that WALLAP provided reasonable estimates of wall movement with the input soil modulus values.

4

Input Parameters We reviewed available information provided in the contract documents (including geotechnical boring logs and laboratory testing data) and performed a literature search of case history data. Our local knowledge and experience allowed us to use data from previous work, including full scale load testing and empirically derived soil modulus values, to confirm that parameters provided in the contract documents were suitable and reasonably conservative for the selected analytical methods. After review of the available information, we selected soil modulus values from case histories (O’Rourke, 1975) that were higher than values estimated from laboratory testing because laboratory test modulus values for stiff clays are typically underestimated. Because of concerns about potential wall movements impacting the U.S. Capitol Building, we selected modulus values at the low end of the expected range from the case histories to provide an upper bound estimate of soil and wall movements. Review of Results (QA/QC) Because of the complexity of the USCVC excavation support design and the importance of protecting the U.S. Capitol Building, emphasis was placed on quality assurance and review of the results. The following steps were taken for quality assurance: • A numerical check and a “gut check” were performed on the results of all individual WALLAP and PLAXIS analyses. Because PLAXIS is a graphically driven program, a detailed summary of the input parameters is not provided. To perform the numerical check of the PLAXIS input, we developed a quality assurance checklist containing the required input parameters and boundary conditions, and the numerical check was performed by someone familiar with the program reviewing the input in the program’s preprocessor. • Select WALLAP analyses were checked using apparent pressure envelopes. As discussed previously, apparent pressure envelopes are not recommended as the basis of design for diaphragm walls, but are tools useful for gauging if the analysis results are reasonable. • The results of the PLAXIS analysis and WALLAP analysis performed on the same design section were compared to confirm that there was reasonable agreement between the two models. • Ground movements estimated from the PLAXIS model were evaluated using empirical relationships between wall movements and expected ground movements (Clough, 1990). • An independent reviewer was engaged to evaluate the selection of design parameters and the output from critical sections at the U.S. Capitol Building. Monitoring Program The monitoring program for the excavation support system consisted of: • Survey monuments on critical structures and utilities. • Inclinometers installed in the diaphragm walls. • Optical survey points on the diaphragm walls. Results of the monitoring program indicated that movements in critical areas were well within the expected range of movements. See Figure 2 for comparison of the estimated wall movements from the WALLAP and PLAXIS analyses with measured movement from inclinometers. Conclusions The development of more powerful computers and more sophisticated computer programs has dramatically increased our ability to design excavation support systems quickly and efficiently, but the importance of thorough data collection, understanding computer models, and evaluating and understanding computer output cannot be overemphasized. Through the use of case history data, multiple analytical methods, and thorough review of analysis results, we were able to design an excavation support system that minimized wall movements and protected the U.S. Capitol Building from damage. Preliminary design for the project, performed during the bid stage,, occurred in about 10 days, and the final comprehensive design was performed in less than a month, highlighting the advantages of using modern computational models.

5

Figure 2. Modeled and Measured Wall Displacement Data References Clough, W.G. and O'Rourke, T.D., 1990. "Construction induced movements of in-situ walls." Design and Performance of Earth Retaining Structures, ASCE GSP No.25, 439 - 470. O’Rourke, T. D., 1981. "Ground movements caused by braced excavations," J. Geotech. Engrg Div., ASCE, 107 (GT9), 1159-1178. O’Rourke, T. D. 1975. "A study of two braced excavations in sands and interbedded stiff clays," Ph.D. dissertation, Univ. of Illinois, 255 pp. Pearlman, S. L., Walker, M. P., and Boscardin, M. D. (2004). "Deep Underground Basements for Major Urban Building Construction.” Proc., GeoSupport 2004, Drilled Shafts, Micropiling, Deep Mixing, Remedial Methods, and Specialty Foundation Systems, J. P. Turner and Paul W. Mayne, eds., Geo-Institute, ASCE, Geotechnical Special Publication No. 124, 545-560. PLAXIS, 1998. Finite Element Code for Soil and Rock Analyses. Brokgreve and Vermeer, et al., (ed.), Balkema. Rotterdam, Brookfield, Version 7, A.A. Terzaghi, K., Peck, R. B., and Mesri, G., 1996. Soil Mechanics in Engineering Practice, Third Edition, John Wiley & Sons, New York, NY, 349-360. WALLAP, 1997. Anchored and cantilevered retaining wall analysis program, D.L. Borin, MA, Ph.D., CEng., MICE. Geosolve, Users Manual, Version 4. Weidlinger Associates, Inc. 2001. "Geotechnical Engineering Study, United States Capitol Visitor Center", Prepared for RTKL Associates Inc. and The Architect of the Capitol.

6

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close