Sara Khoshnevisan

Robust design in geotechnical engineering – an update

This paper presents an update for the robust geotechnical design (RGD) methodology, which seeks an optimal design with respect to design robustness and cost efficiency, while satisfying the safety requirements. In general, the design robustness is achieved if the system response is insensitive to the variation in the uncertain input parameters (called “noise factors”). In other words, a design is considered robust if the system response exhibits little variation, even though there is high variation in the input parameters. Robust design achieves this desirable outcome by carefully adjusting ‘design parameters’ (i.e., the parameters that can be controlled by the designer, such as the geometry and dimensions) without reducing the uncertainty in the noise factors. In this paper, the existing RGD methodology is updated with a gradient-based robustness measure and a simplified procedure for seeking the knee point. The RGD methodology and its simplified version (with new updates) are illustrated with three design examples. The results presented in this paper show that the RGD methodology and its simplified version are effective design tools that considers safety, cost and design robustness simultaneously. The advantages of the simplified RGD approach are discussed.

Efficient robust geotechnical design of drilled shafts in clay using a spreadsheet

Abstract This paper presents an efficient robust geotechnical design (RGD) approach that considers performance requirements, design robustness, and cost efficiency simultaneously. In this paper, design robustness is measured via the variation in the performance function of concern that can be evaluated using reliability analysis. Furthermore, the performance requirements of the system are also evaluated using reliability analysis. Thus, the evaluation of design robustness and the evaluation of performance requirements share common computational steps, referred to herein as computational coupling. This coupling for computational efficiency is a significant feature of the proposed RGD approach. Within the framework of the proposed RGD approach, design robustness, cost efficiency, and performance requirements can be considered simultaneously by means of multiobjective optimization. Furthermore, a practical and efficient procedure is developed for such optimization using a feature resident in a popular spread...

Response surface-based robust geotechnical design of supported excavation – spreadsheet-based solution

ABSTRACT The robust geotechnical design (RGD) approach which involves optimization to obtain a design that is safe, cost-efficient, and robust in the face of uncertainties, can be computationally challenging for complex geotechnical structures. In this study, the RGD approach has become practical by introducing a response surface as a surrogate to finite element- or finite difference-based computer code that is used for analyzing the system, and developing a fast algorithm for the optimization process. For demonstration purposes, a real-world supported excavation project is designed using this modified RGD approach and it is compared with the one designed by a local expert.

Simplified procedure for reliability-based robust geotechnical design of drilled shafts in clay using spreadsheet

ABSTRACT This paper provides a simplified procedure for reliability-based robust geotechnical design (RGD) using spreadsheet. In the RGD methodology, design robustness is achieved by adjusting “design parameters” without reducing the uncertainties in noise factors. This design approach generally involves a multi-objective optimisation, which is computationally challenging. To improve the efficiency of the RGD methodology, the design robustness is evaluated in terms of sensitivity index and the safety requirement is evaluated using mean value first order second moment (MFOSM). To ease the concern that the reliability index obtained with MFOSM may not be sufficiently accurate, a mapping function that relates MFOSM to a more accurate method such as first order reliability method is introduced. To further improve the efficiency of the proposed simplified RGD method, a new simplified procedure along with a more accurate robustness measure is developed that eliminates the need for multi-objective optimisation. With these modifications, the proposed simplified RGD method can efficiently be implemented in a single Excel spreadsheet. The proposed simplified method, which goes beyond any existing reliability-based RGD methods in terms of ease of use and computational efficiency, is illustrated in this paper with an example of robust design of drilled shaft in clay.

Robust Geotechnical Design of Shield-Driven Tunnels Using Fuzzy Sets

This paper presents a new geotechnical design concept - robust geotechnical design (RGD) - which seeks to minimize the variation of the tunnel circumferential performance to the uncertainty of geotechnical parameters (which is a measure of design robustness) in addition to meeting the requirements of safety and cost efficiency. To this end, uncertain geotechnical parameters are represented as fuzzy sets (or more specifically, fuzzy numbers) to copy with the scarcity of data, which provides an adequate measure of the uncertainties when the knowledge is incomplete or data is limited. Given fuzzy input parameters, the circumferential performance of a shield-driven tunnel is uncertain and logically expressed as a fuzzy factor of safety, Based upon which the failure probabilities and robustness of the tunnel circumferential performance can be computed in the RGD framework. Finally, to aid in the selection of the optimal design that maximizes robustness and minimizes cost simultaneously while the safety is brought to the target level, non-dominated optimization is performed and a Pareto front is obtained. the knee point located on the Pareto front is further identified and recommended as the optimal design, showing the best compromise between robustness and cost. The proposed RGD methodology of shield-driven tunnels is deterministic in nature. Through an illustrative design example presented, the effectiveness and significance of the proposed RGD methodology in the design of shield-driven tunnels for the circumferential performance is demonstrated.

Robust design optimization applied to braced excavations

Design of an excavation support system must satisfy the minimum factors of safety for stability requirements and the wall deformation and/or ground settlement requirements. In this paper, the authors present an application of robust geotechnical design (RGD) method on the design of braced excavations. The essence of RGD is to derive an optimal design through a careful adjustment of the design parameters so that the response of the braced excavation system is insensitive to the variation of “noise factors” such as uncertain soil parameters, model errors, and construction variation. The robust design of a diaphragm-wall-supported excavation requires an optimal selection of the design parameters such as the length (L) and thickness (t) of the wall, the vertical spacing of the struts (S), and the stiffness (EA) of the strut. Within the RGD framework, the effect of uncertainties in the noise factors on the variation of the system response is evaluated. Furthermore, the design robustness is sought along with the cost efficiency and safety. Thus, the RGD methodology involves a multi-objective optimization. As cost and robustness are conflicting objectives, such optimization usually leads to a Pareto front, which offers a tradeoff that can aid in making an informed decision.

R-LRFD:RobustLoad and Resistance Factor Design

In the modern design codes adopting load and resistance factor design (LRFD), the resistance factors for various models for the design of geotechnical systems were calibrated with the load factors that had been determined by structural engineers. Although the consistent performance or risk level was intended for LRFD, the uncertainty remains in the geotechnical design obtained with LRFD, as the variability of input parameters at a given site could be quite different from those variation levels assumed in the calibration of resistance factors. Thus, the question of how likely the completed geotechnical system meets the performance requirement does not vanish even with LRFD. A logical question then is “how can we perform LRFD so that the geotechnical system will remain feasible (i.e., being able to satisfy the performance requirement) in the face of uncertainties.” In this paper, we proposed an approach that integrates the robust design concept with LRFD, termed robust LRFD (R-LRFD) approach, to help solve this problem. We demonstrated this R-LRFD approach with a drilled shaft design example documented in ETC10. Comparison was made between the LRFD solution with that obtained using the proposed R-LRFD approach.

Assessing Liquefaction-Induced Lateral Spreads Using CPT Cases from the Christchurch, New Zealand, Earthquakes

During past strong earthquakes, liquefaction-induced ground movements such as settlements and lateral spreads have caused considerable damages to buildings, bridges, and lifelines. Assessment of such ground movements is of great interest to engineers. This paper focuses on the subject of liquefaction-induced lateral spreads. An evaluation of the existing methods for predicting lateral spreads with the newly compiled dataset from the more recent earthquakes is desirable. In this paper, we compiled case histories of lateral spreads from two major events of the 2010-2011 Canterbury/Christchurch Earthquake Sequence. These cases were assessed from raw data and compiled into a database of lateral spread cases with seismic parameters (moment magnitude and PGA) and soil data (primarily consisting of CPTs). The newly compiled cases were used to evaluate the performance of the existing empirical methods. The results show large scatter in the plot of the predicted versus observed lateral displacements, although the general trend is quite satisfactory.