Skip to main content
SpringerLink
Log in

Applying machine learning to predict the tensile shear strength of bonded beech wood as a function of the composition of polyurethane prepolymers and various pretreatments

  • Original
  • Published:
Wood Science and Technology Aims and scope Submit manuscript

Abstract

The present study showed that machine learning can be used to predict the tensile shear strength of bonded beech wood as a function of the composition of polyurethane prepolymers and various pretreatments. A comprehensive experimental data set was used for training and testing the algorithms support vector machines, random forest and artificial neural networks. Within the framework of the experiments, the structure–property relationships of 1C PUR prepolymers were analyzed by systematical variation of the structural parameters urea and urethane group content, cross-link density, ethylene oxide content, and the functionality via isocyanate (NCO) or polyether component. The bonded wood joints were tested according to DIN EN 302-1. Prior to testing, the shear test specimens were pretreated according to procedures A1 and A4, five temperature steps (5, 40, 70, 150 and 200 °C) and two alternating climates. The complete data set (N = 2840) was preprocessed and split into a training set and a test set using tenfold cross-validation. The performance of the algorithms was evaluated with the coefficient of determination (R2), root-mean-square error (RMSE) and mean absolute percentage error (MAPE). All machine learning algorithms revealed a high accuracy, but the artificial neural network showed the best performance with R2= 0.92, RMSE = 0.948 and a MAPE of 9.21. The work paves the way for future machine learning applications in the field of adhesive bonding technology and may enable a fast and effective development of new adhesives and enhance the efficiency of their application.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2

Similar content being viewed by others

References

  • Bardak S, Tiryaki S, Bardak T, Aydin A (2016a) Predictive performance of artificial neural network and multiple linear regression models in predicting adhesive bonding strength of wood. Strength Mater 48:811–824. https://doi.org/10.1007/s11223-017-9828-x

    Article  Google Scholar 

  • Bardak S, Tiryaki S, Nemli G, Aydın A (2016b) Investigation and neural network prediction of wood bonding quality based on pressing conditions. Int J Adhes Adhes 68:115–123. https://doi.org/10.1016/j.ijadhadh.2016.02.010

    Article  CAS  Google Scholar 

  • Basheer IA, Hajmeer M (2000) Artificial neural networks: fundamentals, computing, design, and application. J Microbiol Methods 43:3–31

    Article  CAS  Google Scholar 

  • Bishop CM (2006) Pattern recognition and machine learning (information science and statistics). Springer, New York

    Google Scholar 

  • Breiman L (2001) Random forests. Mach Learn 45:5–32. https://doi.org/10.1023/a:1010933404324

    Article  Google Scholar 

  • Burden F, Winkler D (2009) Bayesian regularization of neural networks. In: Livingstone DJ (ed) Artificial neural networks: methods and applications. Humana Press, Totowa, pp 23–42. https://doi.org/10.1007/978-1-60327-101-1_3

    Chapter  Google Scholar 

  • Cao M, Alkayem NF, Pan L, Novák D (2016) Advanced methods in neural networks-based sensitivity analysis with their applications in civil engineering. In: Rosa JLG (ed) Artificial neural networks—models and applications. InTech, Rijeka, p 13. https://doi.org/10.5772/64026

    Chapter  Google Scholar 

  • Chen H, Engkvist O, Wang Y, Olivecrona M, Blaschke T (2018) The rise of deep learning in drug discovery. Drug Discov Today 23(6):1241–1250. https://doi.org/10.1016/j.drudis.2018.01.039

    Article  PubMed  Google Scholar 

  • Clauß S, Allenspach K, Gabriel J, Niemz P (2010) Improving the thermal stability of one-component polyurethane adhesives by adding filler material. Wood Sci Technol 45:383–388. https://doi.org/10.1007/s00226-010-0321-y

    Article  CAS  Google Scholar 

  • Clauß S, Dijkstra DJ, Gabriel J, Kläusler O, Matner M, Meckel W, Niemz P (2011) Influence of the chemical structure of PUR prepolymers on thermal stability. Int J Adhes Adhes 31:513–523. https://doi.org/10.1016/j.ijadhadh.2011.05.005

    Article  CAS  Google Scholar 

  • DIN EN 386 (2001) Glued laminated timber—Performance requirements and minimum production requirements. Beuth, Berlin

    Google Scholar 

  • DIN EN 302-1 (2004) Adhesives for load-bearing timber structures—Test methods-Part 1: Determination of bond strength in longitudinal tensile shear strength. Beuth, Berlin

    Google Scholar 

  • DIN EN 1995-1-1 (2010) Eurocode 5: Design of timber structures—Part 1-1: General-Common rules and rules for buildings. Beuth, Berlin

    Google Scholar 

  • DIN EN 14080 (2011) Timber structures—Glued laminated timber and glued solid timber-Requirements. Beuth, Berlin

    Google Scholar 

  • DIN EN 15425 (2015) Adhesives-One component polyurethane (PUR) for load-bearing timber structures—Classification and performance requirements. Beuth, Berlin

    Google Scholar 

  • Foresee FD, Hagan MT (1997) Gauss-Newton approximation to Bayesian learning. In: Proceedings of international conference on neural networks (ICNN’97), vol 1933, pp 1930–1935. https://doi.org/10.1109/icnn.1997.614194

  • Hajmeer MN, Basheer IA, Najjar YM (1997) Computational neural networks for predictive microbiology II. Application to microbial growth. Int J Food Microbiol 34:51–66

    Article  CAS  Google Scholar 

  • Hinton GE, Salakhutdinov RR (2006) Reducing the dimensionality of data with neural networks. Science 313:504–507. https://doi.org/10.1126/science.1127647

    Article  CAS  PubMed  Google Scholar 

  • Hinton G, Deng L, Yu D et al (2012) Deep neural networks for acoustic modeling in speech recognition: the shared views of four research groups. IEEE Signal Process Mag 29:82–97. https://doi.org/10.1109/MSP.2012.2205597

    Article  Google Scholar 

  • Kira K, Rendell LA (1992) A practical approach to feature selection. In: Paper presented at the proceedings of the ninth international workshop on machine learning, Aberdeen, Scotland, UK

    Chapter  Google Scholar 

  • Kläusler O (2007) Untersuchung zur Auswirkung der Zusammensetzung von Polyurethan-Prepolymeren auf die Verklebungsgüte von Buchenholz. (Investigation of the effect of the composition of polyurethane prepolymers on the bonding quality of beech wood). Diploma Thesis, Hamburg University, Hamburg

  • Kläusler O, Rehm K, Elstermann F, Niemz P (2014) Influence of wood machining on tensile shear strength and wood failure percentage of one-component polyurethane bonded wooden joints after wetting. Int Wood Prod J 5:18–26. https://doi.org/10.1179/2042645313Y.0000000039

    Article  Google Scholar 

  • Kohavi R (1995) A study of cross-validation and bootstrap for accuracy estimation and model selection. In: Paper presented at the proceedings of the 14th international joint conference on Artificial intelligence, vol 2, Montreal, Quebec, Canada

  • Korley LTJ, Pate BD, Thomas EL, Hammond PT (2006) Effect of the degree of soft and hard segment ordering on the morphology and mechanical behavior of semicrystalline segmented polyurethanes. Polymer 47:3073–3082. https://doi.org/10.1016/j.polymer.2006.02.093

    Article  CAS  Google Scholar 

  • LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444. https://doi.org/10.1038/nature14539

    Article  CAS  Google Scholar 

  • Mansfield SD, Iliadis L, Avramidis S (2007) Neural network prediction of bending strength and stiffness in western hemlock (Tsuga heterophylla Raf.). Holzforschung 61:707. https://doi.org/10.1515/hf.2007.115

    Article  CAS  Google Scholar 

  • Noble PA, Almeida JS, Lovell CR (2000) Application of neural computing methods for interpreting phospholipid fatty acid profiles of natural microbial communities. Appl Environ Microbiol 66:694–699

    Article  CAS  Google Scholar 

  • Richter K, Schirle M (2002) Behaviour of 1 K PUR adhesives under increased moisture and temperature conditions. In: Teischinger, Stingl (eds) Lignovisionen, Proceedings of the international Symposium on Wood Based Materials. BOKU, Vienna, pp 153–158

    Google Scholar 

  • Richter K, Steiger R (2005) Thermal stability of wood-wood and wood-FRP bonding with polyurethane and epoxy adhesives. Adv Eng Mater 7:419–426. https://doi.org/10.1002/adem.200500062

    Article  CAS  Google Scholar 

  • Richter K, Pizzi A, Despres A (2006) Thermal stability of structural one-component polyurethane adhesives for wood—structure-property relationship. J Appl Polym Sci 102:5698–5707. https://doi.org/10.1002/app.25084

    Article  CAS  Google Scholar 

  • Schrödter A, Niemz P (2006) Investigation on the failure behaviour of glue joints at high temperatures and relative humidity. Holztechnologie 47:24–32

    Google Scholar 

  • Šebenik U, Krajnc M (2007) Influence of the soft segment length and content on the synthesis and properties of isocyanate-terminated urethane prepolymers. Int J Adhes Adhes 27:527–535. https://doi.org/10.1016/j.ijadhadh.2006.10.001

    Article  CAS  Google Scholar 

  • Snoek J, Larochelle H, Adams RP (2012) Practical bayesian optimization of machine learning algorithms. arXiv e-prints

  • Tiryaki S, Bardak S, Bardak T (2015) Experimental investigation and prediction of bonding strength of Oriental beech (Fagus orientalis Lipsky) bonded with polyvinyl acetate adhesive. J Adhes Sci Technol 29:2521–2536. https://doi.org/10.1080/01694243.2015.1072989

    Article  CAS  Google Scholar 

  • Vapnik VN (1995) The nature of statistical learning theory. Springer, New York

    Book  Google Scholar 

Download references

Acknowledgements

The authors would like to express their gratitude to the Bayer Material Science AG, Leverkusen, for providing the 1C PUR prepolymers. Furthermore, the authors would like to thank Prof. Dr. P. Niemz for his professional support and Martin Arnold as well as Dr. Mirko Luković for proofreading the manuscript.

Author information

Authors and Affiliations

  1. Laboratory for Cellulose and Wood Materials, Group WoodTec, Empa, Swiss Federal Laboratories for Materials Science and Technology, Überlandstr. 129, 8600, Dübendorf, Switzerland

    Mark Schubert

  2. Swiss Wood Solutions AG, c/o ETH Zürich, Stefano-Franscini-Platz 3, HIF E 29.1, 8093, Zurich, Switzerland

    Oliver Kläusler

Authors
  1. Mark Schubert
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Oliver Kläusler
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Mark Schubert or Oliver Kläusler.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Schubert, M., Kläusler, O. Applying machine learning to predict the tensile shear strength of bonded beech wood as a function of the composition of polyurethane prepolymers and various pretreatments. Wood Sci Technol 54, 19–29 (2020). https://doi.org/10.1007/s00226-019-01144-6

Download citation

  • Received:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00226-019-01144-6

Search

Navigation