No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Application of Artificial Neural Network for Diagnosing Pile Integrity Based on Low Strain Dynamic Testing
Abstract
The artificial neural network (ANN) models are presented for diagnosing pile in this paper based on the pile integrity test (PIT) also known as low strain dynamic test. The back-propagation learning algorithm is employed to train the network for extracting knowledge from training examples. There are fifty-three input neurons in the network including the PIT response and pile length, cross-sectional area and wave velocity. In order to obtain the pile condition in quantity, the novel technique is proposed containing two back-propagation ANN models. The first is to identify the defect patters while the second to investigate the exact degree of pile defect by computing the change of equivalent cross-sectional area. Training and testing data were drawn from response records of actual piles. The results from the testing phase indicate that the presented method is successful.
... Others have also applied neural networks to PIT signals. For example, Zhang & Zhang [16] proposed using two ANN models for processing the PIT signals for diagnosing pile integrity. They used the first ANN model for identifying defect patterns and the second ANN model for assessing severity of the defects. ...
Low strain pile integrity testing (LSPIT), due to its simplicity and low cost, is one of the most popular NDE methods used in pile foundation construction. While performing LSPIT in the field is generally quite simple and quick, determining the integrity of the test piles by analyzing and interpreting the test signals (reflectograms) is still a manual process performed by experienced experts only. For foundation construction sites where the number of piles to be tested is large, it may take days before the expert can complete interpreting all of the piles and delivering the integrity assessment report. Techniques that can automate test signal interpretation, thus shortening the LSPIT’s turnaround time, are of great business value and are in great need. Motivated by this need, in this paper, we develop a computer-aided reflectogram interpretation (CARI) methodology that can interpret a large number of LSPIT signals quickly and consistently. The methodology, built on advanced signal processing and machine learning technologies, can be used to assist the experts in performing both qualitative and quantitative interpretation of LSPIT signals. Specifically, the methodology can ease experts’ interpretation burden by screening all test piles quickly and identifying a small number of suspected piles for experts to perform manual, in-depth interpretation. We demonstrate the methodology’s effectiveness using the LSPIT signals collected from a number of real-world pile construction sites. The proposed methodology can potentially enhance LSPIT and make it even more efficient and effective in quality control of deep foundation construction.
... Relating work on inverse analysis and defect identification problems solved by optimization and ANNs can be found in (Stavroulakis et al., 2003(Stavroulakis et al., , 2004Stavroulakis, 2000). Relative work can be, also, found in (Tam et al., 2004;Zhang and Zhang, 2009). These approaches exploit relatively simple ANN topologies, using a few selected inputs. ...
Semi-supervised learning (SSL) is a class of supervised learning tasks and techniques that also exploits the unlabeled data for training. SSL significantly reduces labeling related costs and is able to handle large data sets. The primary objective is the extraction of robust inference rules. Decision support systems (DSSs) who utilize SSL have significant advantages. Only a small amount of labelled data is required for the initialization. Then, new (unlabeled) data can be utilized and improve system's performance. Thus, the DSS is continuously adopted to new conditions, with minimum effort. Techniques which are cost effective and easily adopted to dynamic systems, can be beneficial for many practical applications. Such applications fields are: (a) industrial assembly lines monitoring, (b) sea border surveillance, (c) elders' falls detection, (d) transportation tunnels inspection, (e) concrete foundation piles defect recognition, (f) commercial sector companies financial assessment and (g) image advanced filtering for cultural heritage applications.
... Relating work on inverse analysis and defect identification problems solved by optimization and neural networks can be found in [20][21][22][23]. Post processing of experimental or numerical data, coming from PIT by means of ANNs techniques, have been published in several papers, see among others [24][25][26][27][28]. ...
A genetically optimized neural detector is utilized for the identification of structural defects in concrete piles. The proposed methodology is applied on numerically generated data, involving two major defect types. A coupled finite element and scaled boundary finite element method approach is used to model the pile and its surrounding soil. The oscillation patterns, produced on the surface of the pile, depend strongly on the introduced defect type. The proposed defect detection system provides information about the type and the placement of the defect(s), given the surface's oscillation patterns.
This study investigates variations in the velocity and sensitivity of electromagnetic waves in transmission lines configured in defective model piles for the detection of necking defects containing soil. Experiments are performed with model piles containing defects filled with different materials, such as air, sands, and clay. Five different types of transmission lines are configured in model piles. The electromagnetic waves are generated and detected using a time domain reflectometer. The velocity of electromagnetic waves is highest when the defect is filled with air, and it decreases with an increase in the water content. The velocity is lowest when the defect is filled with clay. The sensitivity of transmission lines for detecting defects decreases with an increase in soil water contents. The transmission line with a single electrical wire and epoxy-coated rebar exhibits the highest sensitivity, followed by that with three and two parallel electrical wires. Transmission lines with a single electrical wire and uncoated rebar and those with two parallel electrical wires wrapped with a sheath exhibit poor sensitivity when the defect is filled with clay. This study demonstrates that electromagnetic waves can be effective tools for detecting necking defects with wet and conductive soils in bored piles.
The reflected wave method is simple and audiovisual, which has come into common use in the testing of pile integrity, but it is difficult to solve the stress wave attenuation by the absorption and scattering of pile and soil. The reflected wave equation including the attenuation factor of pile and soil and the reflecting factors of discontinuity interface, is transformed into the nonlinear least square description, the trust region method is suggested to deal with. The Gauss-Newton direction is used to define the two-dimensional subspace of the trust region to accelerate the convergence. The example for real pile verifies the effectiveness of this method.
Diagnosis of damage of prestressed concrete piles during driving is an important problem in foundation engineering. An effort to build an expert system for the problem is described in this paper. To overcome the bottleneck of knowledge acquisition, an artificial neural network is used as the learning mechanism to transfer engineering experience into usable knowledge. The back-propagation learning algorithm is employed to train the network for extracting knowledge from training examples. The influences of various control parameters (including learning rate and momentum factor) and various network architecture factors (including the number of hidden units and the number of hidden layers) are examined. The results prove that the artificial neural network can work sufficiently as a knowledge-acquisition tool for the diagnosis problem. To apply the knowledge in the trained network, a reasoning strategy that hybridizes forward-and backward-reasoning schemes is proposed to realize the inference mechanism.
We show that standard multilayer feedforward networks with as few as a single hidden layer and arbitrary bounded and nonconstant activation function are universal approximators with respect to Lp(μ) performance criteria, for arbitrary finite input environment measures μ, provided only that sufficiently many hidden units are available. If the activation function is continuous, bounded and nonconstant, then continuous mappings can be learned uniformly over compact input sets. We also give very general conditions ensuring that networks with sufficiently smooth activation functions are capable of arbitrarily accurate approximation to a function and its derivatives.
Previous studies have applied artificial neural networks (ANN) with the back-propagation learning algorithm for diagnosing pre-stressed concrete piles. Recent developments of ANN breed a new form of network architectures for modeling this specific type of classification problems: Probabilistic Neural Networks (PNN). This paper presents this probabilistic neural network architecture for diagnosing the causes of prestressed concrete pile damages. In this paper, the use of neural networks for construction is first presented and the various types of neural networks are introduced. Then, based upon a set of data collected from the previous study on prestressed concrete pile diagnosis, the common features of concrete pile damage and their causes are identified. The PNN model and its architecture are described. Using the set of data, the network is trained and the procedural steps are described. A random seed approach for cross validation is used to assess the reliability of the model. Lastly, the result of the network training is discussed and analyzed, which demonstrates the robustness of the model developed.
The application of neural network method in idendifi-cating the defect of foundation piles
- H C Gan
- J H Lu
- J Tong
- H.C. Gan
A neural network model for diagnosing pile integrity based on reflection wave method
- C H Wang
- W Zhang
- C.H. Wang
The multi-reflected wave analysis on the low strain integrity testing of pile foundation
- C H Zhang
- C.H. Zhang
Diagnosis of prestressed concrete pile defects using probabilistic neural networks. Engineering Structures
- C M Tam
- K L Thomas
- Tong
- C.M. Tam