Neural network assisted multiscale analysis for the elastic properties prediction of 3D braided composites under uncertainty

Balokas, Georgios; Czichon, Steffen; Rolfes, Raimund

doi:10.1016/j.compstruct.2017.06.037

Cited by 109 publications

(39 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Twenty years have passed and we are yet to see such broader uptake but there is a more realistic hope that the time is now right due to pervasive computing and powerful computers, and the whole big data and internet of things revolution. Recent applications of machine learning and neural networks in structural engineering are mostly related to prediction and modelling of elastic properties of materials [inter alia 16,17], compressive and bond strength of concrete [e.g. 18,19], buckling load [20][21][22], development of cementitious composites [23], and the refinement finite element models [e.g.…”

Section: Machine and Deep Learning In Structural And Civil Engineeringmentioning

confidence: 99%

Machine Learning for Sustainable Structures: A Call for Data

et al. 2019

View full text Add to dashboard Cite

Buildings are the world's largest contributors to energy demand, greenhouse gases (GHG) emissions, resource consumption and waste generation. An unmissable opportunity exists to tackle climate change, global warming, and resource scarcity by rethinking how we approach building design. Structural materials often dominate the total mass of a building; therefore, a significant potential for material efficiency and GHG emissions mitigation is to be found in efficient structural design and use of structural materials. To this end, environmental impact assessment methods, such as life cycle assessment (LCA), are increasingly used. However, they risk failing to deliver the expected benefits due to the high number of parameters and uncertainty factors that characterise impacts of buildings along their lifespans. Additionally, effort and cost required for a reliable assessment seem to be major barriers to a more widespread adoption of LCA. More rapid progress towards reducing building impacts seems therefore possible by combining established environmental impact assessment methods with artificial intelligence approaches such as machine learning and neural networks. This short communication will briefly present previous attempts to employ such techniques in civil and structural engineering. It will present likely outcomes of machine learning and neural network applications in the field of structural engineering andmost importantly-it calls for data from professionals across the globe to form a fundamental basis which will enable quicker transition to a more sustainable built environment.

show abstract

Section: Machine and Deep Learning In Structural And Civil Engineeringmentioning

confidence: 99%

Machine Learning for Sustainable Structures: A Call for Data

et al. 2019

View full text Add to dashboard Cite

show abstract

“…It is capable of giving predictions for a new case of input within the range of the examples with which it was trained in a very short period of time. Thus they are widely used in a variety of problems that need to be simplified, including composite materials' technological problems [13][14]. In the present work, 2 ANNs were developed using MatLab NN Tool [15] as seen in Fig.…”

Section: Artificial Neural Networkmentioning

confidence: 99%

Prediction of mechanical properties of porous CFRP specimens by ANNs and X-ray CT data

2018

View full text Add to dashboard Cite

Carbon fiber reinforced plastics (CFRPs) have evolved into the primary material for several lightweight structures. However, despite their extensive use and the quality amelioration, CFRPs remain susceptible to a variety of manufacturing defects, the most common of which are the pores. Predictive tools capable of correlating the mechanical properties of CFRP parts with the characteristics of defects as derived from non-destructive testing (NDT) techniques or even further with the manufacturing parameters could serve as an effective tool for the quality control of CFRP structural parts. In the present paper, the characteristics of pores as evaluated by X-ray Computed Tomography (CT) have been correlated with the matrixdominated mechanical properties of unidirectional porous CFRP specimens using Artificial Neural Networks (ANN). Thirty (30) porosity scenarios have been created and given as input to the numerical model. That multi-scale numerical model, which had been validated experimentally, has been used for training the ANN model. The predictions of the ANN agree very well with results from mechanical tests. Moving one step forward, a second ANN has been developed to correlate the autoclave pressure directly with the mechanical properties of the CFRP specimens. The validity of the latter ANN depends on the accuracy of the relation between the autoclave pressure and the characteristics of the pores. The present work represents a step towards the development of effective quality control tools for composite materials.

show abstract

“…Instead, surrogate model approaches, which are constructed based on a limited set of actual input/output data points, are a suitable method when dealing with such complex problems. Several methods of application of surrogate models have been reported in the literature for evaluation of uncertainties in composite laminates, such as Kriging method [15,16], radial basis function [17], polynomial chaos expansion (PCE) [18][19][20], and artificial neural network (ANN) [21,22]. State-of-the-art reviews on the surrogate models for evaluating the uncertainty in structural responses of composite laminates can be found in [23].…”

Section: Introductionmentioning

confidence: 99%

Uncertainty analysis of mechanical behavior of flax fiber composites

Ravandi¹,

Banu²,

Noorian³

2020

Preprint

View full text Add to dashboard Cite

The natural origin of the fibers combined with random production flaws results in significant uncertainties in the properties of natural fiber reinforced composites. A probabilistic assessment can help to characterize the uncertainties and evaluate the reliability of natural fiber composites, enabling their use in engineering designs. Toward this end, this study aims to quantify the uncertainties in the tensile strength and frequency response of a unidirectional flax/epoxy composite due to the variability of various input parameters, including the fiber material properties and manufacturing flaws. Based on the available data in the literature, the non-deterministic input variables were divided into normal and uniform variables using a statistical test. A computationally efficient response surface approach based on the polynomial chaos expansion was adopted to conduct the uncertainty analysis with multi-type uncertain variables. Moreover, the results were validated by the direct Monte Carlo simulation to demonstrate the accuracy and efficiency of the surrogate model.

show abstract

Neural network assisted multiscale analysis for the elastic properties prediction of 3D braided composites under uncertainty

Cited by 109 publications

References 44 publications

Machine Learning for Sustainable Structures: A Call for Data

Machine Learning for Sustainable Structures: A Call for Data

Prediction of mechanical properties of porous CFRP specimens by ANNs and X-ray CT data

Uncertainty analysis of mechanical behavior of flax fiber composites

Contact Info

Product

Resources

About