M. Jaksa

Recent advances and future challenges for artificial neural systems in geotechnical engineering applications

Artificial neural networks (ANNs) are a form of artificial intelligence that has proved to provide a high level of competency in solving many complex engineering problems that are beyond the computational capability of classicalmathematics and traditional procedures. In particular, ANNs have been applied successfully to almost all aspects of geotechnical engineering problems. Despite the increasing number and diversity of ANN applications in geotechnical engineering, the contents of reported applications indicate that the progress in ANN development and procedures is marginal and not moving forward since the mid-1990s. This paper presents a brief overview of ANN applications in geotechnical engineering, briefly provides an overview of the operation of ANN modeling, investigates the current research directions of ANNs in geotechnical engineering, and discusses some ANN modeling issues that need further attention in the future, including model robustness; transparency and knowledge extraction; extrapolation; uncertainty.

A transfer coefficient method for rock slope toppling.

A common block toppling mode is adopted for analysing toppling stability of rock slopes. A transfer coefficient method is developed based on Goodman and Bray’s kinematic model, by introducing two parameters: the transfer coefficient and equivalent toppling weight, which, given the geometrical and geotechnical parameters of the slope, can be determined as functions of the relative position of the block to the slope crest. Accounting for the case that the block base is not normal to the dominant discontinuities, the arithmetic is extended. The method describes a concise relationship between the normal force of the blocks and their weight represented by their equivalent toppling weight. The analysis process is formulated and easily implemented. A spreadsheet procedure is presented and a typical case is illustrated. It is demonstrated that the transfer coefficient method provides a considerably straightforward means for analysing block toppling of rock slopes and is computationally very efficient.

Measuring the Risk of Geotechnical Site Investigations

The site investigation phase of any geotechnical design plays a vital role, where inadequate characterization of the subsurface conditions may contribute to either a significantly over designed solution that is not cost-effective, or an under design, which may lead to potential failures. Although it is intuitive to expect that the financial risk of a design will reduce as the site investigation scope increases (i.e. additional sampling), it is not known to what degree the risk is reduced, nor whether other uncertainties have an impact on this relationship. As such, this paper discusses research to measure the impact of varying the scope of a site investigation, on the financial risk of a foundation design project. The financial risk is defined as the total cost, which includes costs associated with undertaking the site investigation, constructing the foundation and superstructure, and any works required to rehabilitate a foundation failure. The analysis is numerically based, where a foundation design simulation model is incorporated into a Monte Carlo framework, in order to generate GSP 170 Probabilistic Applications in Geotechnical Engineering Copyright ASCE 2007 Geo-Denver 2007: New Peaks in Geotechnics 2 expected costs, and a measure of the financial risk. Results indicate that the risk of a foundation design is considerably reduced as the scope of a site investigation increases. However, results also indicate that there is an optimal site investigation expenditure, which leads to the least financial risk, and where additional sampling becomes redundant.

Assessing soil correlation distances and fractal behavior

An essential parameter used in reliability analyses of geotechnical engineering systems, particularly foundations, and probabilistic modeling of soil profiles is the correlation distance, which is also known as the scale of fluctuation and the range of influence. This parameter is often difficult to measure, as it requires extensive field testing, which is usually beyond the scope of most geotechnical investigations, and is hence associated only with uncommon research studies. This paper presents the results of a number of investigations in clay and sand, which were designed to quantify the spatial variability and, hence, the correlation distance of these soils. The paper describes and makes use of random field theory, geostatistics and fractal theory to facilitate the assessment of spatial variability and the correlation distances in both the vertical and horizontal directions.

Effect of sample location on the reliability based design of pad foundations

Site investigations that aim to sufficiently characterize a soil profile for foundation design, typically consist of a combination of in situ and laboratory tests. The number of tests and/or soil samples is generally determined by the budget and time considerations placed upon the investigation. Therefore, it is necessary to plan the locations of such tests to provide the most suitable information for use in design. This is considered the sampling strategy. However, the spatial variability of soil properties increases the complexity of this exercise. Results presented in this paper identify the errors associated with using soil properties from a single sample location on a pad foundation designed for settlement. Sample locations are distributed around the site to identify the most appropriate sample location and the relative benefits of taking soil samples closer to the center of the proposed footing. The variability of the underlying soil profile is also shown to a have a significant effect on the errors due to sampling location. Such effects have been shown in terms of the statistical properties of the soil profile. The performance of several common settlement relationships to design a foundation based on the results of a single sample location have also been examined.

Intelligent Computing for Predicting Axial Capacity of Drilled Shafts

In the last few decades, numerous methods have been developed for predicting the axial capacity of drilled shafts. Among the available methods, the cone penetration test (CPT) based models have been shown to give better predictions in many situations. This can be attributed to the fact that CPT-based methods have been developed in accordance with the results of the CPT tests, which have been found to yield more reliable soil properties, hence, more accurate axial capacity predictions of drilled shafts. In this paper, one of the most commonly used artificial intelligence techniques, i.e. artificial neural networks (ANNs), was utilized in an attempt to obtain more accurate axial capacity predictions for drilled shafts. The ANN model was developed using data collected from the literature that comprise CPT results and drilled shaft load tests of 94 case records. The predictions from the ANN model were compared with those obtained from three commonly used available CPT-based methods. The results indicate that the ANN-based model provides more accurate axial capacity predictions of drilled shafts and outperforms the available conventional methods.

Assessing the influence of root reinforcement on slope stability by finite elements

This paper aims to investigate the effect of root reinforcement on slope stability using finite element methods. It is well recognised that plant roots can improve the shear strength of soils by their high tensile strength and closely spaced root matrix system. The increase in soil shear strength due to root reinforcement is considered as an increase in apparent soil cohesion, called root cohesion, cr. In this paper, a freely available (http://www.inside.mines.edu/~vgriffit/slope64) finite element code called slope64 described by Griffiths and Lane (Géotechnique 49(3):387–403, 1999) is used to model the effect of root reinforcement on slope stability. The root cohesion is added directly to the soil cohesion for the soil elements that are reinforced by plant roots. The results from the finite element analyses demonstrate that the factor of safety of a slope increases when the effect of root reinforcement is taken into consideration. A series of stability charts are developed which can be used for assessing the influence of root reinforcement on slope stability.

Use of proctor compaction testing for deep fill construction using impact rollers

1107 USE OF PROCTOR COMPACTION TESTING FOR DEEP FILL CONSTRUCTION USING IMPACT ROLLERS BRENDAN SCOTT and MARK JAKSA and YIEN LIK KUO School of Civil, Environmental & Mining Engineering, The University of Adelaide, North Terrace, Adelaide SA 5005, Australia. E-mail: Brendan.Scott@adelaide.edu.au School of Civil, Environmental & Mining Engineering, The University of Adelaide, North Terrace, Adelaide SA 5005, Australia. E-mail: Mark.Jaksa@adelaide.edu.au; Yien.Kuo@adelaide.edu.au.

AN IMPROVED STATISTICALLY BASED TECHNIQUE FOR EVALUATING THE CPT FRICTION RATIO

This paper examines a statistical technique known as the cross-correlation function (CCF) for determining the shift distance associated with the cone penetration test (CPT). When evaluating the friction ratio, FR ( = fr/qc), for soil classification purposes, it is essential that the measured values of ), for soil classification purposes, it is essential that the measured values of qc and fs are shifted relative to one another because of the physical offset between the cone and the friction sleeve. Generally, the shift distance is estimated by means of empirical and subjective methods, a value of 75 to 100 mm is adopted, or it is ignored all together. Using a series of case studies, this paper demonstrates that the CCF is a useful and objective technique for estimating the shift distance. In addition, a phenomenon associated with sleeve friction measurements related to elastic rebound of clay soils is discussed.

Passive noise datasets at regolith sites

The data presented in this article contain datasets of passive noise measurements at regolith sites in Adelaide, South Australia. The data were acquired using three component (3C) LE-3Dlite Lennartz seismometers with an eigenfrequency of 1 Hz. The data were acquired at eight sites across Adelaide׳s regolith in a hexagonal array layout. Four tests, each with a duration of 30 min, were conducted at different times. The ambient noise data can be used for both horizontal to vertical spectral ratio (HVSR) analysis and array analyses, which are essential to obtain the site fundamental frequency and the ellipticity of the fundamental mode Rayleigh waves at the measured site. The array analyses are useful to obtain the dispersion curves, which are needed to estimate the shear wave velocity profile.