Sheng Huoo Ni

Training Artificial Neural Networks with the Aid of Fuzzy Sets

This article presents a technique of training artificial neural networks (ANNs) with the aid of fuzzy sets theory. The proposed ANN model is trained with field observation data for predicting the collapse potential of soils. This ANN model uses seven soil parameters as input variables. The output variable is the collapsibility (whether the soil is collapsible) or the collapse potential (if the soil is judged collapsible). The proposed technique involves a module for preprocessing input soil parameters and a module for postprocessing network output. The preprocessing module screens the input data through a group of predefined fuzzy sets, and the postprocessing module, on the other hand, “defuzzifies” the output from the network into a “nonfuzzy” collapse potential, a single value. The ANN with the proposed preprocessing and post-process techniques is shown to be superior to the conventional ANN model in the present study.

Numerical investigation of the scouring effect on the lateral response of piles in sand

Scouring around a bridge foundation is a problem of much concern to civil engineers. The main purpose of this paper is to gain more understanding of the effect of scouring around the pile on the lateral capacity of the piles embedded in sandy soil. Factors such as soil stiffness, pile head fixity, and pile slenderness ratio (L/B) were studied to show their effects on the variation of the percentage decrease of lateral load capacity (PDC) due to scouring near an isolated pile. The results indicate that the PDC value reaches almost 50% when the scour depth reaches 1.3 to 2.4 times the pile diameter, and that piles with a fixed head are more capable of resisting lateral load loss. Furthermore, the PDC values remain almost constant after the pile slenderness ratio is greater than 10. Hence, the decrease in lateral load due to scouring of the soil will be more serious for short piles.

Estimation of Dynamic Properties of Sand Using Artificial Neural Networks

Dynamic properties of soils are usually determined by time-consuming laboratory tests. This study presents a method for estimating dynamic soil parameters using artificial neural networks. A simple feedforward neural network with back-propagation training algorithm is used. The neural network is trained with actual laboratory data, which consists of six input variables. They are the standard penetration test value, the void ratio, the unit weight, the water content, the effective overburden pressure, and the mean effective confining pressure. The output layer consists of a single neuron, representing shear modulus or damping ratio. Results of the neural network training and testing show that predictions of shear modulus by the neural network approach is reliable although it is less successful in predicting damping ratio.

The horizontal amplitude screening effects of a rectangular strip in-filled trench

The horizontal amplitude screening effects of an in-filled trench are solved using the boundary element method in the frequency domain. Dimensionless parametric studies are performed to assess the screening effects of the trench in terms of the geometry, in-filled material, foundational embedment, and shear wave velocity ratio of the in-filled material to the surrounding soil. The results show that the screening effects of the in-filled trench depend highly on the trench geometry and the velocity ratio of the in-filled material to the surrounding soil. The embedment ratio of the foundation and other factors are relatively unimportant.

Fines content effects on liquefaction potential evaluation for sites liquefied during chi‐chi earthquake, 1999

Abstract The purpose of this paper is to present the effect of fines content on the evaluation of liquefaction potential while using Seeds method (1985) and Tokimatsu and Yoshimis method (1983). The data used in the paper were obtained from holes bored at sites where liquefied during Chi‐Chi Earthquake, 1999. The fines content of soil is the major influence on safety and stability and is the target of the above methods. Student t‐Test results show that the factor of safety calculated by Seeds method and T&Ys method can be regarded as identical when fines content less than 35%. However, the results of T&Ys method will overestimate liquefaction resistance when the fines content is greater than 35%. An appropriate correction factor for fines content greater than 35% is proposed from back‐calculated analysis using data from liquefied sites. The result show the proposed correction factor will improve the efficiency of liquefaction potential evaluation in the T&Y’ s method at high fines contents.

A study on dynamic properties of cement-stabilized soils

In this paper the dynamic property (shear modulus and damping ratio) of cement-stabilized soil is studied with using the resonant column test. The amount of cement admixed, the magnitude of confining pressure, and shearing strain amplitude are the parameters considered. Test results show that the maximum shear modulus of cement-stabilized soil increases with increasing confining pressure, the minimum damping ratio decreases with increasing confining pressure. The shear modulus of cement-stabilized soil decreases with increasing shearing strain while the damping ratio increases with increasing shearing strain. In the paper the relationship of shear modulus versus shearing strain is fitted into the Ramberg-Osgood equations using regression analysis.

Using Continuous Wavelet Transform to Construct the Dispersion Image for Soil Layers

This study used a time-frequency domain analysis for estimating the dispersion curve of a Rayleigh wave by using two receivers. The signals were first transformed using continuous wavelet transform. A similar slant stack procedure was used to analyze the wavelet transform signals and extract a dispersion image. This method is advantageous because it requires no empirical judgment in phase unwrapping and few receivers. To examine the applicability of the method for evaluating the dispersion curve for soil layers with lateral heterogeneity, three synthetic examples and an experience example were investigated. In these examples, numerical simulations of the surface wave seismic test were performed using the finite difference FLAC code. The results revealed that the estimates of the surface wave dispersion curve, obtained using the method, coincide with those of the theoretical values. A high-resolution dispersion image is generated by increasing the spacing of receivers. The method is applicable for extracting a dispersion image for lateral heterogeneous soil layers.

An EMD-based procedure to evaluate the experimental dispersion curve of the SASW method

An empirical mode decomposition (EMD)-based procedure is proposed in this paper to evaluate the experimental dispersion curve of the surface wave (SASW) method. Four field sites were selected to validate the advantages of the suggested EMD-based procedure for determining shear wave velocity profiles, which are also compared with those obtained by using suspension PS logging tests. The results show that the EMD-based procedure can reduce the effects of noise and higher modes on the experimental dispersion curve, and thus can improve the reliability of the evaluated shear wave velocity profiles. In addition, the effective depth of the shear wave velocity profile evaluated by using the SASW method can also be increased when the EMD-based procedure is introduced.

Integrity Evaluation of PCC Piles Using the Surface Reflection Method

The surface reflection method was used in this study to evaluate its applicability and limitations in the evaluation of the integrity of pour concrete-cased (PCC) piles. Parametric studies of the surface reflection technique were carried out numerically and experimentally, and the responses were analyzed both in the time and frequency domains. It was found that the length of the PCC piles can be determined with confidence in both domains. However, detecting the integrity of PCC piles in the time domain seems more appropriate than that in the frequency domain. Also, the relative angle between the impact and receivers should not be larger than 90° to obtain satisfactory accuracy in the determination of the defect depth.

NONDESTRUCTIVE INTEGRITY EVALUATION OF PC PILE USING WIGNER-VILLE DISTRIBUTION METHOD

Nondestructive evaluation (NDE) techniques have been used for years to provide a quality control of the construction for both drilled shafts and driven concrete piles. This trace is typically made up of transient pulses reflected from structural features of the pile or changes in its surrounding environment. It is often analyzed in conjunction with the spectral response, mobility curve, arrival time, etc. The Wigner-Ville Distribution is a new numerical analysis tool for signal process technique in the time-frequency domain and it can offer assistance and enhance signal characteristics for better resolution both easily and quickly. In this study, five single pre-cast concrete piles have been tested and evaluated by both sonic echo method and Wigner-Ville distribution (WVD). Furthermore, two difficult problems in nondestructive evaluation problems are discussed and solved: the first one is with a pile with slight defect, whose necking area percentage is less than 10%, and the other is a pile with multiple defects. The results show that WVD can not only recognize the characteristics easily, but also locate the defects more clearly than the traditional pile integrity testing method.