Neural Networks: Some Successful Applications in Computational Mechanics (original) (raw)
Related papers
Soft computing techniques in probabilistic seismic analysis of structures
2006
Earthquake-resistant design of structures using probabilistic analysis is an emerging field in structural engineering. The objective of this chapter is to investigate the efficiency of soft computing methods when incorporated into the solution of computationally intensive earthquake engineering problems. Two methodologies are proposed in this work where limitstate probabilities of exceedance for real world structures are determined. Neural networks based metamodels are used in order to replace a large number of time-consuming structural analyses required for the calculation of a limit-state probability. The Rprop algorithm is employed for the training of the neural networks; using data obtained from appropriately selected structural analyses.
Neural Nets and Structural Safety: Applications and Ideas
1994
Abstract The paper describes the first achievements and the future developments of the application of neural networks to structural engineering at ISMES. Two main fields of applications are presented: neural nets in design, with reference to the design of arch dams, and in data interpretation, with reference to the management of structural safety.
Monte Carlo analysis of structural systems using neural networks
Monte Carlo simulation is increasingly being used in the analysis of large and complex structural systems for the assessment of the uncertainty spread and the reliability. A major handicap for the popularization of this technology is the large number of deterministic evaluations needed to such purposes, inasmuch as linear or nonlinear finite element solvers are required for each output sample calculation. In order to simplify this task neural networks are evaluated in this paper as a partial surrogate of the deterministic solver. The neural networks are trained with the input/output pairs resulting from a few number of finite element simulations, and are henceforth used in a Monte Carlo context. It is shown that when employed in this way, neural networks constitute a promising tool for a drastic reduction of the computational cost needed by a Monte Carlo simulation in this field of application. Three types of networks have been selected for the study, two of which correspond to supervised and the other one to hybrid learning procedures. The paper compares the network designs in their more relevant aspects, which are the training speed and accuracy, the extrapolation ability and the accuracy of the estimated probabilities.
Advances in Engineering Software, 2008
The objective of this paper is to investigate the efficiency of soft computing methods, in particular methodologies based on neural networks, when incorporated into the solution of computationally intensive engineering problems. Two types of applications have been considered, namely parameter (flaw) identification and probabilistic seismic analysis of structures. Artificial neural networks (ANNs) based metamodels are used in order to replace the time-consuming repeated structural analyses. The back-propagation algorithm is employed for training the ANN, using data derived from selected analyses. The trained ANN is then used to predict the values of the necessary data. The numerical tests demonstrate the computational advantages of the proposed methodologies.
2016
One of the efficient methods in seismic resistant design of medium height buildings is applying seismic isolation systems at the base of the structures to mitigate the response of structure. In this study an effective numerical reliability-based optimization technique is presented for the optimum design of isolation system under random time history earthquake loading. Friction Pendulum System (FPS) as one of the popular types of seismic isolation devices is considered to protect delicate equipment installed on the floor of a specific concrete building. So the object is to minimize the probability of failure of the base-isolated building subjected to design performance criteria in terms of the story acceleration of the superstructure. Due to stochastic nature of variables such as input ground motion; a novel method is proposed to predict the reliability of the supposed structure using artificial neural networks (ANN). The reliability of the system in the format of probability of fail...
Computer Methods in Applied Mechanics and Engineering, 1996
This paper examines the application of Neural Networks (NN) to the reliability analysis of complex structural systems in connection with Monte Carlo Simulation (MCS). The failure of the system is associated with the plastic collapse. The use of NN was motivated by the approximate concepts inherent in reliability analysis and the time consuming repeated analyses required for MCS. A Back Propagation algorithm is implemented for training the NN utilising available information generated from selected elasto-plastic analyses. The trained NN is then used to compute the critical load factor due to different sets of basic random variables leading to close prediction of the probability of failure. The use of MCS with Importance Sampling further improves the prediction of the probability of failure with Neural Networks.
Engineering Structures
The fragility curve is defined as the conditional probability of failure of a structure, or its critical components, at given values of seismic intensity measures (IMs). The conditional probability of failure is usually computed adopting a log-normal assumption to reduce the computational cost. In this paper, an artificial neural network (ANN) is constructed to improve the computational efficiency for the calculation of structural outputs. The following aspects are addressed in this paper: (a) Implementation of an efficient algorithm to select IMs as inputs of the ANN. The most relevant IMs are selected with a forward selection approach based on semi-partial correlation coefficients; (b) Quantification and investigation of the ANN prediction uncertainty computed with the delta method. It consists of an aleatory component from the simplification of the seismic inputs and an epistemic model uncertainty from the limited size of the training data. The aleatory component is integrated in the computation of fragility curves, whereas the epistemic component provides the confidence intervals; (c) Computation of fragility curves with Monte Carlo method and verification of the validity of the log-normal assumption. This methodology is applied to estimate the probability of failure of an electrical cabinet in a reactor building studied in the framework of the KARISMA benchmark.
7th ECCOMAS Thematic Conference on Computational Methods in Structural Dynamics and Earthquake Engineering, Crete - Greece, 24/06/2019-26/06/2019, 2019
The implementation of methods which belong to the field of Artificial Intelligence such as the Artificial Neural Networks (ANN) based methods is continuously increased in many scientific and technological applications. As regards the civil engineering applications, the investigation for the utilization of these methods has led to very promising results. More specifically, the experimental application of ANN-based methods for the seismic vulnerability assessment of structures has proved that they can be utilized as alternative methods in parallel of the well-documented existing methods. However, despite their promising results there is no wide acceptance of ANNs as computational tools for the prediction of the seismic damage level of structures. This can possibly be attributed mainly to the fact that the vast majority of civil engineers who investigate methods for structures' seismic vulnerability assessment has no the minimum required background about the abilities and the utilization of the ANN-based methods. The current paper attempts to present a different point of view of the ANN-based methods and to prove that the research for their further implementation can be approachable by civil engineers provided that the corresponding formulation is defined in certain stages. In the framework of the current paper it is also proved that by the utilization of ANNs the definition of relatively simple equations for the preliminary estimation of the seismic damage level of R/C buildings in near-real time is possible.
Stochastic Finite Element Simulation of Uncertain Structures Subjected to Earthquake
Shock and Vibration, 2000
In present study, the stochastic finite element simulation based on the efficient Neumann expansion technique is extended for the analysis of uncertain structures under seismically induced random ground motion. The basic objective is to investigate the possibility of applying the Neumann expansion technique coupled with the Monte Carlo simulation for dynamic stochastic systems upto that extent of parameter variation after which the method is no longer gives accurate results compared to that of the direct Monte carlo simulation. The stochastic structural parameters are discretized by the local averaging method and then simulated by Cholesky decomposition of the respective covariance matrix. The earthquake induced ground motion is treated as stationary random process defined by respective power spectral density function. Finally, the finite element solution has been obtained in frequency domain utilizing the advantage of Neumann expansion technique.