BITS Faculty Publications
Permanent URI for this communityhttp://localhost:4000/handle/123456789/1867
Browse
10 results
Search Results
Item Neural Networks for Damage Detection in Steel Railway Bridges(IABSE, 1995) Barai, Sudhir KumarThe paper presents Artificial Neural Networks developed for typical steel railway bridges for the purpose of damage detection. Multilayer perceptrons have been used for generating the architecture for the bridges of different configurations. The back propagation algorithm has been adopted for training the network with simulated damage states. The training pairs have been generated using a standard finite element program. The weights of the trained networks have been stored and can be used as a knowledge source independently. It is demonstrated that the trained networks have practical relevance.Item Ensemble modeling or selecting the best model: Many could be better than one(1999) Barai, Sudhir KumarIn the course of data modeling, many models could be created. Much work has been done on formulating guidelines for model selection. However, by and large, these guidelines are conservative or too speci c. Instead of using general guidelines, models could be selected for a particular task based on statistical tests. When selecting one model, others are discarded. Instead of losing potential sources of information, models could be combined to yield better performance. We review the basics of model selection and combination and discuss their di erences. Two examples of opportunistic and principled combinations are presented. The rst demonstrates that mediocre quality models could be combined to yield signi cantly better performance. The latter is the main contribution of the paper; it describes and illustrates a novel heuristic approach called the SG (k-NN) ensemble for the generation of good quality and diverse models that can even improve excellent quality models.Item Ensemble modelling or selecting the best model: Many could be better than one(CUP, 1999-11-01) Barai, Sudhir KumarIn the course of data modelling, many models could be created. Much work has been done on formulating guidelines for model selection. However, by and large, these guidelines are conservative or too specific. Instead of using general guidelines, models could be selected for a particular task based on statistical tests. When selecting one model, others are discarded. Instead of losing potential sources of information, models could be combined to yield better performance. We review the basics of model selection and combination and discuss their differences. Two examples of opportunistic and principled combinations are presented. The first demonstrates that mediocre quality models could be combined to yield significantly better performance. The latter is the main contribution of the paper; it describes and illustrates a novel heuristic approach called the SG(k-NN) ensemble for the generation of good-quality and diverse models that can even improve excellent quality models.Item Performance of the generalized delta rule in structural damage detection(Elsiever, 1995-04) Barai, Sudhir KumarThe paper examines the suitability of the generalized data rule in training artificial neural networks (ANN) for damage identification in structures. Several multilayer perceptron architectures are investigated for a typical bridge truss structure with simulated damage states generated randomly. The training samples have been generated in terms of measurable structural parameters (displacements and strains) at suitable selected locations in the structure. Issues related to the performance of the network with reference to hidden layers and hidden neurons are examined. Some heuristics are proposed for the design of neural networks for damage identification in structures. These are further supported by an investigation conducted on five other bridge truss configurations.Item Sensitivity‐Based Weighted‐Average in Structural Damage Assessment(ASCE, 1994) Barai, Sudhir KumarTo quantify the global damageability of a structure, the weighted‐average function defined in terms of damage condition and the importance or weightages of structural elements has generally been used in literature. Judgment of structural importance of a structural element is a difficult task. The opinion provided by experts could have different perspective or bias, which may reflect in the overall assessment process. Structural analysis being precise these days, structural importance of a member could be realistically estimated using standard structural procedures. In this paper it is proposed that the weightage computation could be linked to damage sensitivity of the element response. It is argued that the weighted‐average computation based on damage sensitivity is a more realistic index for integrity assessment. In this perspective, the concept of damage sensitivity and its computational aspects based on the finite‐element method are also presented. Finally, examples of weighted‐average computation comprising fuzzy sets and importance factor obtained from normalized damage sensitivity are illustrated.Item Time-delay neural networks in damage detection of railway bridges(Elsiever, 1997-01) Barai, Sudhir KumarThe recent developments in multilayer perceptron using the backpropagation algorithm, has opened up new possibilities in structural identification. Limitation of traditional neural networks (TNN) in dealing with patterns that may vary in time domain has given birth to time-delay neural networks (TDNN). In the present paper the TNN and the TDNN have been implemented in detecting the damage in bridge structure using vibration signature analysis. A comparative study has been carried out for the various cases of complete as well as incomplete measurement data. It has been observed that TDNNs have performed better than TNNs in this application.Item Structural Sensitivity as a Measure of Redundancy(ASCE, 1997) Barai, Sudhir KumarThe conventional definition of redundancy is applicable to skeletal structural systems only, whereas the concept of redundancy has never been discussed in the context of a continuum. Generally, structures in civil engineering constitute a combination of both skeletal and continuum segments. Hence, this paper presents a generalized definition of redundancy that has been defined in terms of structural response sensitivity, which is applicable to both continuum and discrete structures. In contrast to the conventional definition of redundancy, which is assumed to be fixed for a given structure and is believed to be independent of loading and material properties, the new definition would depend on strength and response of the structure at a given stage of its service life. The redundancy measure proposed in this paper is linked to the structural response sensitivities. Thus, the structure can have different degrees of redundancy during its lifetime, depending on the response sensitivity under consideration. It is believed that this new redundancy measure would be more relevant in structural evaluation, damage assessment, and reliability analysis of structures at large.Item Vibration Signature Analysis Using Artificial Neural Networks(ASCE, 1995-10) Barai, Sudhir KumarDamage detection by measuring and analyzing vibration signals in a machine component is an established procedure in mechanical and aerospace engineering. This paper presents vibration signature analysis of steel bridge structures in a nonconventional way using artificial neural networks (ANN). Multilayer perceptrons have been adopted using the back-propagation algorithm for network training. The training patterns in terms of vibration signature are generated analytically for a moving load traveling on a trussed bridge structure at a constant speed to simulate the inspection vehicle. Using the finite-element technique, the moving forces are converted into stationary time-dependent force functions in order to generate vibration signals in the structure and the same is used to train the network. The performance of the trained networks is examined for their capability to detect damage from unknown signatures taken independently at one, three, and five nodes. It has been observed that the prediction using the trained network with single-node signature measurement at a suitability chosen location is even better than that of three-node and five-node measurement data.Item Evaluating machine learning models for engineering problems(Elsiever, 1999-07) Barai, Sudhir KumarThe use of machine learning (ML), and in particular, artificial neural networks (ANN), in engineering applications has increased dramatically over the last years. However, by and large, the development of such applications or their report lack proper evaluation. Deficient evaluation practice was observed in the general neural networks community and again in engineering applications through a survey we conducted of articles published in AI in Engineering and elsewhere. This status hinders understanding and prevents progress. This article goal is to remedy this situation. First, several evaluation methods are discussed with their relative qualities. Second, these qualities are illustrated by using the methods to evaluate ANN performance in two engineering problems. Third, a systematic evaluation procedure for ML is discussed. This procedure will lead to better evaluation of studies, and consequently to improved research and practice in the area of ML in engineering applications.Item Multilayer perceptron in damage detection of bridge structures(Elsiever, 1995-02-17) Barai, Sudhir KumarRecent developments in artificial neural networks (ANN) have opened up new possibilities in the domain of structural engineering. For inverse problems like structural identification of large civil engineering structures such as bridges and buildings where the in situ measured data are expected to be imprecise and often incomplete, the ANN holds greater promise. The detection of structural damage and identification of damaged element in a large complex structure is a challenging task indeed. This paper presents an application of multilayer perceptron in the damage detection of steel bridge structures. The issues relating to the design of network and learning paradigm are addressed and network architectures have been developed with reference to trussed bridge structures. The training patterns are generated for multiple damaged zones in a structure and performance of the networks with one and two hidden layers are examined. It has been observed that the performance of the network with two hidden layers was better than that of a single-layer architecture in general. The engineering importance of the whole exercise is demonstrated from the fact that measured input at only a few locations in the structure is needed in the identification process using the ANN.