Deep learning-based tensile strength prediction in fused deposition modeling

Zhang, Jianjing; Wang, Peng; Gao, Robert X.

doi:10.1016/j.compind.2019.01.011

Cited by 116 publications

(47 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This is achieved through a black-box and as such a drawback of this exists in there being limited ability for identification of possible causal relationships [42]. The use of this technique was deemed suitable for the sample size of (120) due to previous applications of ANNs in FDM property prediction and for similar sample sizes (144) [29]. IBM SPSS 24 was used to generate a predictive model via use of a multi-layer perceptron neural network.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

The creation of a neural network based capability profile to enable generative design and the manufacture of functional FDM parts

Goudswaard

Hicks

Nassehi

2021

Int J Adv Manuf Technol

View full text Add to dashboard Cite

In order to manufacture functional parts using filament deposition modelling (FDM), an understanding of the machine’s capabilities is necessary. Eliciting this understanding poses a significant challenge due to a lack of knowledge relating manufacturing process parameters to mechanical properties of the manufactured part. Prior work has proposed that this could be overcome through the creation of capability profiles for FDM machines. However, such an approach has yet to be implemented and incorporated into the overall design process. Correspondingly, the aim of this paper is two-fold and includes the creation of a comprehensive capability profile for FDM and the implementation of the profile and evaluation of its utility within a generative design methodology. To provide the foundations for the capability profile, this paper first reports an experimental testing programme to characterise the influence of five manufacturing parameters on a part’s ultimate tensile strength (UTS) and tensile modulus (E). This characterisation is used to train an artificial neural network (ANN). This ANN forms the basis of a capability profile that is shown to be able to represent the mechanical properties with RMSEP of 1.95 MPa for UTS and 0.82 GPa for E. To validate the capability profile, it is incorporated into a generative design methodology enabling its application to the design and manufacture of functional parts. The resulting methodology is used to create two load bearing components where it is shown to be able to generate parts with satisfactory performance in only a couple of iterations. The novelty of the reported work lies in demonstrating the practical application of capability profiles in the FDM design process and how, when combined with generative approaches, they can make effective design decisions in place of the user.

show abstract

Section: Methodsmentioning

confidence: 99%

“…ANNs were also used by Garzon-Hernandes et al who present a two-stage thermal sintering method for predicting the mechanical performance of ABS samples [15]. Zhang et al applied ANNs to predict mechanical properties based upon three process parameters with additional thermal data from the printing process [29].…”

Section: Capability Profilingmentioning

confidence: 99%

The creation of a neural network based capability profile to enable generative design and the manufacture of functional FDM parts

Goudswaard

Hicks

Nassehi

2021

Int J Adv Manuf Technol

View full text Add to dashboard Cite

show abstract

“…The GRU requires fewer parameters, which makes training processing of the network faster, while LSTM provides better performance for some applications. [31]…”

Section: Rnnmentioning

confidence: 99%

Deep Learning for Medical Materials: Review and Perspective

Qi¹,

Liu²

2021

ES Mater. Manuf.

View full text Add to dashboard Cite

Deep learning and similar computational research approaches have cooperated with materials research for years. Especially in the last few years, with the fast evolution of machine learning and deep learning algorithms, a novel branch for material research is presented to be recognized, learned, practiced, adapted and perfected. Different conventional computational modeling methods, the deep learning approach assists material science and engineering from the aspect of data processing and analysis, rather than simulate the reality. However, mistakes always exist in all less explored concepts and developments. Here, in order to offer a better instruction of deep learning for materials scientists, several recent works on deep learning based material research were reviewed. The application of deep learning in material research is introduced and discussed. As an example of both general material and advanced material, medical material was discussed for the impact and future of deep learning methods. At last, we provided several insights on future work for both deep learning scientists, data scientists and materials scientists.

show abstract

“…The AI tool could accurately predict the tensile strength of the developed system. In a separate study, Zhang et al 198 used a long short-term memory network and layer-wise relevance propagation to predict the tensile strength of the FFF printed part. The model was trained with the data generated from experimental testing of components printed with different process parameters.…”

Section: Artificial Intelligence (Ai) and Integrated Process-structurmentioning

confidence: 99%

Review on process model, structure-property relationship of composites and future needs in fused filament fabrication

Papon

Haque

2020

Journal of Reinforced Plastics and Composites

View full text Add to dashboard Cite

This paper presents the state-of-the-art of additive manufacturing of composites for processing functional, load-bearing components. A general overview of different additive manufacturing methods is provided, and specific attention is focused on fused filament fabrication-based composites processing. Different process modeling strategies are summarized, and key aspects of these models are discussed. Significant results such as thermal and fluid flow characteristics, effects of nozzle geometry on melt flow, fiber orientation, bead spreading, and solidification, the formation of residual stresses, and deformation behavior are discussed from computational modeling perspective. The scientific advancement, model limitations, and future modeling needs are prescribed reviewing the current works. A general overview of material development in nano-micro-macro-scale reinforcement is also presented. Different length-scales of reinforcement has its own challenges and promises. The continuous fiber reinforcement has a great potential for being the next-generation composites manufacturing technology. However, the challenges in reducing the void content, better bonding between the fiber–matrix, and layer-layer adhesion, and process uncertainty are some of the key areas yet to advance. Based on the current limitations on computational modeling, materials development, and process modeling studies, future research needs and recommendations are provided.

show abstract

Deep learning-based tensile strength prediction in fused deposition modeling

Cited by 116 publications

References 29 publications

The creation of a neural network based capability profile to enable generative design and the manufacture of functional FDM parts

The creation of a neural network based capability profile to enable generative design and the manufacture of functional FDM parts

Deep Learning for Medical Materials: Review and Perspective

Review on process model, structure-property relationship of composites and future needs in fused filament fabrication

Contact Info

Product

Resources

About