Detail publikace

ANN Inverse Analysis in Stochastic Computational Mechanics

Originální název

ANN Inverse Analysis in Stochastic Computational Mechanics

Anglický název

ANN Inverse Analysis in Stochastic Computational Mechanics

Jazyk

en

Originální abstrakt

An approach of inverse analysis is proposed to obtain parameters and their statistics of a computational model in order to achieve the best agreement with experimental data. The inverse analysis is based on the coupling of a stochastic simulation and an artificial neural network (ANN). The identification parameters play the role of basic random variables with a scatter reflecting the physical range of possible values. A novelty of the approach is the utilization of the efficient small-sample simulation method Latin Hypercube Sampling (LHS) used for the stochastic preparation of the training set utilized in training the neural network. Once the network has been trained, it represents an approximation consequently utilized in a following way: To provide the best possible set of model parameters for the given experimental data. The paper focuses the attention on the statistical inverse analysis of material model parameters where statistical moments of input parameters have to be identified based on experimental data (histograms of response). The efficiency of the approach is shown using numerical examples from civil engineering computational mechanics. The examples of fracture energy experiments of concrete beams and a reinforced concrete frame solved by the nonlinear fracture mechanics tool for the objective failure modeling of structures with significant nonlinear effects are shown. Means and standard deviations of fracture-mechanical material parameters (like modulus of elasticity, fracture energy, etc.) are the subjects of identification.

Anglický abstrakt

An approach of inverse analysis is proposed to obtain parameters and their statistics of a computational model in order to achieve the best agreement with experimental data. The inverse analysis is based on the coupling of a stochastic simulation and an artificial neural network (ANN). The identification parameters play the role of basic random variables with a scatter reflecting the physical range of possible values. A novelty of the approach is the utilization of the efficient small-sample simulation method Latin Hypercube Sampling (LHS) used for the stochastic preparation of the training set utilized in training the neural network. Once the network has been trained, it represents an approximation consequently utilized in a following way: To provide the best possible set of model parameters for the given experimental data. The paper focuses the attention on the statistical inverse analysis of material model parameters where statistical moments of input parameters have to be identified based on experimental data (histograms of response). The efficiency of the approach is shown using numerical examples from civil engineering computational mechanics. The examples of fracture energy experiments of concrete beams and a reinforced concrete frame solved by the nonlinear fracture mechanics tool for the objective failure modeling of structures with significant nonlinear effects are shown. Means and standard deviations of fracture-mechanical material parameters (like modulus of elasticity, fracture energy, etc.) are the subjects of identification.

BibTex


@inbook{BUT55475,
  author="David {Lehký} and Drahomír {Novák}",
  title="ANN Inverse Analysis in Stochastic Computational Mechanics",
  annote="An approach of inverse analysis is proposed to obtain parameters and their statistics of a computational model in order to achieve the best agreement with experimental data. The inverse analysis is based on the coupling of a stochastic simulation and an artificial neural network (ANN). The identification parameters play the role of basic random variables with a scatter reflecting the physical range of possible values. A novelty of the approach is the utilization of the efficient small-sample simulation method Latin Hypercube Sampling (LHS) used for the stochastic preparation of the training set utilized in training the neural network. Once the network has been trained, it represents an approximation consequently utilized in a following way: To provide the best possible set of model parameters for the given experimental data. The paper focuses the attention on the statistical inverse analysis of material model parameters where statistical moments of input parameters have to be identified based on experimental data (histograms of response). 
The efficiency of the approach is shown using numerical examples from civil engineering computational mechanics. The examples of fracture energy experiments of concrete beams and a reinforced concrete frame solved by the nonlinear fracture mechanics tool for the objective failure modeling of structures with significant nonlinear effects are shown. Means and standard deviations of fracture-mechanical material parameters (like modulus of elasticity, fracture energy, etc.) are the subjects of identification.",
  address="Nova Science Publishers",
  booktitle="Artificial Intelligence: New Research",
  chapter="55475",
  edition="1",
  howpublished="print",
  institution="Nova Science Publishers",
  year="2009",
  month="october",
  pages="323--350",
  publisher="Nova Science Publishers",
  type="book chapter"
}