Neural networks for prediction of deflection in composite bridges

Neural networks have been developed for prediction of deflections, at service load, in steel-concrete composite bridges incorporating flexibility of shear connectors, shear lag effect and cracking in concrete slabs. Three neural networks have been presented to cover simply supported bridges, two span continuous bridges and three span continuous bridges. The use of the neural networks requires a computational effort almost equal to that required for the simple beam analysis (neglecting flexibility of shear connectors, shear lag effect and cracking of concrete). The training and testing data for neural networks are generated using finite element software ABAQUS. The neural networks have been validated for number of bridges and the errors are found to be small. Closed form solutions are also proposed based on the developed neural networks. The networks/ closed form solutions can be used for rapid prediction of deflection for everyday design.     

Estimating wall deflections in deep excavations using Bayesian neural networks

In this study, a neural network algorithm has been used to model the soil-structure interaction behavior of deep excavations in clays. The hybrid evolutionary Bayesian back-propagation (EBBP) neural network was used in this study and utilizes the genetic algorithms and gradient descent method to determine the optimal parameters within a Bayesian framework to regularize the complexity of learning and to statistically reflect the uncertainty in data. The EBBP analysis was carried out on an extensive database of braced excavation performance from finite element analyses. Additional parametric studies indicate that the model gives logical and consistent trends. Back-analyses of some instrumented case histories from the literature also indicate that the trained neural network model gives reasonable predictions in comparison to the actual measured values. The trained model can serve as a simple and reliable prediction tool to enable estimates of maximum wall deflection for preliminary design of braced excavations in clay. The model is able to take into consideration various factors such as the wall stiffness, support stiffness, the in-situ stress state, non-homogeneous soil conditions, and the variation of soil properties with depth. An added advantage of this approach is that it provides meaningful error bars for the model predictions.     

Nonlinear static analysis of continuous multi-span suspension bridges

This paper performs the static analysis of multi-span suspension bridges using the deflection theory. Applying the deflection theory conventionally used in three-span suspension bridges to multi-span suspension bridges, the horizontal forces of the main cables due to live loads are obtained by converting the suspension system to the equivalent beam system and using the compatibility equation of cables for four spans. Iterative computations are used due to nonlinearity of the differential equations. The results, such as deflections and moments of girders and horizontal forces of the main cable, are compared with the finite element analysis for verification. The resultant values from two methods are almost the same. Finally, using the linearized deflection theory, parametric studies are performed by influence line analyses for parameters such as the side-to-center span ratio, the tower stiffness ratio and the sag to span ratio. From the parametric studies, alternatives to reduce displacements of girders in the center span and girder moments of all spans are investigated.     

Advanced analysis of multi-span suspension bridges

In this paper, an efficient analysis method considering both geometric and material nonlinearities is proposed for predicting the ultimate strength and behavior of multi-span suspension bridges. The geometric nonlinearities of the cable members due to sag effects are captured using the catenary element, while the geometric nonlinearities of the beam-column members due to second-order effects are captured using the stability functions. The material nonlinearities of the cable and beam-column members are considered using elastic–plastic hinge and refined plastic hinge models, respectively. A simple initial shape analysis method is presented to determine the deformed shape and initial cable tension of the bridge under dead loads. Numerical examples are presented to verify the accuracy and efficiency of the proposed method. In addition, a case study on a four-span suspension bridge is carried out to show the capability of the proposed method in estimating the strength and behavior of very large scale structures.     

On-line building energy prediction using adaptive artificial neural networks

While most of the existing artificial neural networks (ANN) models for building energy prediction are static in nature, this paper evaluates the performance of adaptive ANN models that are capable of adapting themselves to unexpected pattern changes in the incoming data, and therefore can be used for the real-time on-line building energy prediction. Two adaptive ANN models are proposed and tested: accumulative training and sliding window training. The computational experiments presented in the paper use both simulated (synthetic) data and measured data. In the case of synthetic data, the accumulative training technique appears to have an almost equal performance with the sliding window training approach, in terms of training time and accuracy. In the case of real measurements, the sliding window technique gives better results, compared with the accumulative training, if the coefficient of variance is used as an indicator.     

基于小波变换的神经网络空调负荷预测研究

基于小波变换的思想建立了递归BP网络模型来预测空调负荷,改进了网络权值、闽值的修改算法,引入了折扣系数法以提高近期预测精度,结合一实例进行了空调逐时冷负荷预测,结果表明该方法预测精度高,适用于空调负荷预测。     

Assessment of the load-carrying capacity of multi-span masonry arch bridges using fibre beam elements

An approach that makes use of non-linear beam elements with fibre cross-section has been used for modelling the ultimate behaviour of multi-span masonry arch bridges. The proposed approach proves able to take into account the interaction among the spans and the non-linear material behaviour with low computational effort. In order to validate the use of the model for the assessment of masonry arch bridges, the load-carrying capacity for typical multi-span railway bridges has been evaluated and the results compared to experimental results and to currently used limit analysis methods. It is shown that, by parity of constitutive assumption, the method provides the same results as limit analysis, both in terms of maximum load prediction and hinges position at collapse; however, taking the effective ductility capacity into account, a strong reduction in load-carrying capacity with respect to classical limit analysis was found, depending on rise-to-span arch ratio, piers slenderness and backfill height. The approach is then applied to a seven span viaduct of the Italian railway network, for which the effective mechanical properties of masonry were identified through an experimental campaign on brickwork samples according to the effective material properties surveyed on site.     

Improving strength and ductility of continuous composite plate girder bridges

The ultimate load carrying capacity of continuous composite plate girder bridges is usually limited by the local buckling failure of steel girders at interior supports. This paper presents a simple reinforcement method which changes the failure mechanism of the continuous girder from local buckling to formation of plastic hinges at the interior supports and mid-span. Such a change in failure mechanism greatly improves the strength and ductility of the superstructure. In this method the compressive portion of the web near the interior support is braced against local buckling by bolting pairs of stiff bracing elements on opposite sides of the web. The bracing elements prevent local buckling failure of the support section and create a section which can rotate inelastically at plastic moment allowing the second hinge to form at mid-span. The bracing elements may be plates or longitudinal stiffeners which should be designed to remain elastic while the section undergoes plastic deformation. The behavior of plate girders which are reinforced by such bracing elements is studied using nonlinear finite element analyses.     

Ultimate strength of continuous composite box-girder bridges with precast decks

In order to apply precast decks to continuous composite bridges, several experiments and analytical studies were performed. From many previous studies, design criteria for crack controls in transverse joints of prefabricated slabs were confirmed. These considerations were needed for serviceability. The bridges which satisfy service limit states, also, should be evaluated for ultimate strengths to define limit states. In this paper, experimental and analytical studies of two-span continuous composite bridges with open box girder section were conducted. Cracking, yielding and ultimate loads were evaluated and compared with the test results for design of continuous composite bridges with precast decks. To evaluate yielding loads of continuous bridges, an uncracked section method considering moment redistribution which is defined in EUROCODE 4, was considered. In calculation of ultimate strengths, full or partial shear connection and sectional classes which were defined in EUROCODE or AASHTO LRFD specifications were considered. Also, through numerical analysis considering material nonlinearities, moment-curvature relationship and moment redistributions were estimated.     

Concrete compressive strength prediction using ultrasonic pulse velocity through artificial neural networks

Numerous attempts to use ultrasonic pulse velocity (UPV) as a measure of compressive strength of concrete has been made due to obvious advantages of non-destructive testing methods. The present study is conducted for prediction of compressive strength of concrete based on weight and UPV for two different concrete mixtures (namely M20 and M30) involving specimens of two different sizes and shapes as a result of need for rapid test method for predicting long-term compressive strength of concrete. The prediction is done using multiple regression analysis and artificial neural networks. A comparison between two methods depicts that artificial neural networks can be used to predict the compressive strength of concrete effectively. The results are plotted as experimentally evaluated compressive strength versus predicted strength through both methods of analysis.     

Large-scale structural health monitoring using composite recurrent neural networks and grid environments

The demand for resilient and smart structures has been rapidly increasing in recent decades. With the occurrence of the big data revolution, research on data-driven structural health monitoring (SHM) has gained traction in the civil engineering community. Unsupervised learning, in particular, can be directly employed solely using field-acquired data. However, the majority of unsupervised learning SHM research focuses on detecting damage in simple structures or components and possibly low-resolution damage localization. In this study, an unsupervised learning, novelty detection framework for detecting and localizing damage in large-scale structures is proposed. The framework relies on a 5D, time-dependent grid environment and a novel spatiotemporal composite autoencoder network. This network is a hybrid of autoencoder convolutional neural networks and long short-term memory networks. A 10-story, 10-bay, numerical structure is used to evaluate the proposed framework damage diagnosis capabilities. The framework was successful in diagnosing the structure health state with average accuracies of 93% and 85% for damage detection and localization, respectively.     

多跨连续V形刚构桥设计分析

以天泉湖大桥（28＋5×40＋28）m连续V形墩为例,介绍分析了V形连续刚构桥的受力特点及高次超静定内力的解决方法,提出了通过合理设置铰的方法解决内力的观点。     

钢-混凝土连续组合梁桥负弯矩区抗裂设计

为进一步验证抗拔不抗剪连接件对组合梁桥负弯矩区抗裂的效果,并研究采用该连接件时负弯矩截面的设计方法,利用考虑滑移效应的组合梁负弯矩区受力模型,推导了采用抗拔不抗剪连接时组合梁负弯矩截面承载力的计算模型,计算结果与有限元分析结果吻合良好.通过理论推导,进一步证明了抗拔不抗剪连接件对负弯矩截面承载力的折减不大,且能有效延缓...     

Unconfined compressive strength prediction of soils stabilized using artificial neural networks and support vector machines

This study aims to improve the unconfined compressive strength of soils using additives as well as by predicting the strength behavior of stabilized soils using two artificial-intelligence-based models. The soils used in this study are stabilized using various combinations of cement, lime, and rice husk ash. To predict the results of unconfined compressive strength tests conducted on soils, a comprehensive laboratory dataset comprising 137 soil specimens treated with different combinations of cement, lime, and rice husk ash is used. Two artificial-intelligence-based models including artificial neural networks and support vector machines are used comparatively to predict the strength characteristics of soils treated with cement, lime, and rice husk ash under different conditions. The suggested models predicted the unconfined compressive strength of soils accurately and can be introduced as reliable predictive models in geotechnical engineering. This study demonstrates the better performance of support vector machines in predicting the strength of the investigated soils compared with artificial neural networks. The type of kernel function used in support vector machine models contributed positively to the performance of the proposed models. Moreover, based on sensitivity analysis results, it is discovered that cement and lime contents impose more prominent effects on the unconfined compressive strength values of the investigated soils compared with the other parameters.     

Response of a train moving on multi-span railway bridges undergoing ground settlement

This paper presents an incremental-iterative procedure to investigate the influence of ground settlement on dynamic interactions of train–bridge system. The train is simulated as a sequence of identical sprung mass units with equal intervals and the bridge system as a series of simple beams with identical properties. To resolve the train-induced vibrations of a beam structure undergoing support settlement, this study decomposes the total beam response into two parts: the static response due to vertical support settlement and the dynamic component caused by inertia effect of beam vibration. An exact solution for static displacement is presented by exerting the support displacements on the beam statically. Thus the remaining dynamic response of the vehicle/bridge coupling system is solved by Galerkin’s method and computed using an iterative approach with Newmark’s finite difference formulas. Numerical studies indicate that for the dynamic interactions of train–bridge system, the inclusion of ground settlement is generally small on the bridge response, but it can amplify drastically the vertical response of the moving train, especially for the concave-up settlement profile. This conclusion is of significance in aligning a rail route that has to cross a region with local land subsidence.     

Simplified vulnerability assessment of reinforced concrete circular piers in multi-span simply supported bridges

This paper proposes a simplified procedure for the seismic vulnerability analysis of multi-span simply supported bridges, in the case of single piers with solid circular sections. The proposed method can be applied whenever the seismic response of the whole bridge depends on the most critical pier. For an assigned limit state, the procedure determines the capacity curve of the critical pier as a function of three parameters (elastic stiffness, displacement at yielding, displacement at collapse), starting from the behaviour under combined axial stress and bending, and then taking into account the different possible collapse modes (shear failure; lap splice failure of the longitudinal bars; buckling) and the geometric nonlinearity. A significant numerical example is presented in which the traditional FEM solution is compared with the proposed simplified procedure.     

Some observations on biofouling prediction in pipelines using model trees and artificial neural networks versus logistic regression

Biofouling is the phenomenon of micro-organism attachments to wet surfaces. Complete understanding of the mechanisms and rates of biofilms creations are partially understood therefore forecasting their formations is difficult. This study is on biofouling predictions for water distribution systems pipelines using model trees, artificial neural networks, and logistic regression. The three methods were tested through base runs and sensitivity analysis runs using data from the experiment conducted by Simões et al. (2006). The results showed that none of the models were superior for all cases, therefore a single model could not be recommended. This leads to an important conclusion that utilising ‘low cost’ modelling methods such as logistic regression can be sufficient for providing reliable estimates for biofilm growth potential. ‘Low cost’ approaches should be applied prior to invoking expensive models such as data driven methods as the latter might not be needed.