Subproject E4

Numerical prognosis for time-dependent alteration of textile strengthened structures

Director | Research Staff | Goals | Methods | Results | Publications

Director

4th Founding periode

Prof. Dr.-Ing. habil. Michael Kaliske
Prof. Dr.-Ing. Wolfgang Graf
Institute for Structural Analysis

3rd Founding period

Dr.-Ing. Michael Beer

Research Staff

Dr.-Ing. Steffen Freitag

Goals

Textile strengthened structures are subject to long-term changes. It must be clarified, how changes of material properties affect the serviceability and the load-carrying capacity. Experiments can generally only be conducted over short time periods with respect to the service life or life time of a structure.

In this subproject numerical procedures for a long-term prognosis of changes of material properties of textile strengthened structures are developed, which operate on the basis of measurement series that are only moderate in length. In this context a distinction is made between real-time data and time-condensed data.

For real-time data, i.e. for measurement series which are gained from experiments under natural environmental conditions, a prognosis procedure is developed on the basis of neural networks. Network type and network architecture are specified in such a way that nonlinear time-dependent measurement series of material properties can be reproduced. As a basis, a multi-layer perceptron network with sigmoid activation functions and feed-forward architecture is selected. The prediction of structural responses under time varying stresses is carried out with recurrent neural networks. Constructional and layout details as well as the operation mode are defined according to particular requirements. The measurement series are used for training the networks. The trained networks then directly generate the prognoses.

For time-condensed data, i.e. for measurement series which are gained from experiments under artificial accelerating (time-condensing) environmental conditions, a solution based on time-condensing procedures with acceleration functions is developed. Starting point is the probability distribution for selected material properties under accelerating environmental conditions, which is obtained from the measured data. The corresponding distribution under natural environmental conditions is to be determined. The dependencies between the given and the sought distribution functions are described with the aid of acceleration functions, which can be estimated from the experimental data. For these purposes, the available samples of only limited extent are numerically extended in size with a novel procedure in a preprocessor. For the analysis of the sought dependencies, again, neural networks are designated.

Both neural networks and time-condensing procedures are applicable very flexibly due to their mathematical bases. They operate model-free and extract information directly from the observed data. A subjective model specification for concrete cases is not required.

top of page

Methods

In this subproject the strategy of developing numerical computational algorithms on a general theoretical basis is pursued to cover a broad spectrum of applications. Fundamentals are probability theory, mathematical statistics, and the theory of neural networks. With the aid of a hybrid prognosis procedure, data analysis and numerical prognosis are combined. The prognosis algorithms are developed and tested for a computer-based application. The new developments make use of experiences from pertinent preparatory work, in particular, in the numerical simulation of stochastic processes with neural networks and in the numerical extension of samples.

top of page

Results

The developed numerical methods and algorithms for long-term prognoses are transferred into software, which are used for the solution of the example problems.

Neural network based prediction

  • Prediction of the creep behaviour of textile strengthened structures
    PDF-file (115 kB - in German)

Time-condensing procedures

  • Lifetime prediction of textile strengthened structures
    PDF-file (33 kB)

top of page

Publications

2010

  • Freitag, S.; Steinigen, F.; Graf, W.; Kaliske, M.: Numerical Long-Term Simulation of TRC Strengthened RC Structures. In: Brameshuber, W. (ed.): Proceedings of the International RILEM Conference on Material Science (MatSci) - Volume I, 2nd International Conference of Textile Reinforced Concrete (ICTRC), Aachen, RILEM Publications S.A.R.L., Bagneux, 2010, pp. 319-329
  • Graf, W.; Kaliske, M.; Sickert, J.-U.; Pannier, S.; Freitag, S.: Neural Networks and Imprecise Probability Concepts for the Design of Industry-Sized Structures. In: Khalili, N.; Valliappan, S.; Li, Q.; Russell, A. (Eds.): Proceedings of the 9th World Congress on Computational Mechanics (WCCM), Sydney, 2010, pp. 72-73
  • Beck, J.L.; Graf, W.; Katafygiotis, L. (eds.): Computer-Aided Civil and Infrastructure Engineering 25 (2010) 5 – Special Issue on "Computational Intelligence in Structural Engineering and Mechanics"
  • Freitag, S.; Graf, W.; Kaliske, M.: Prediction of Time-Dependent Structural Responses with Recurrent Neural Networks. Proceedings in Applied Mathematics and Mechanics 10 (2010), pp. 155-156 - doi:10.1002/pamm.201010070
  • Freitag, S.: Modellfreie numerische Prognosemethoden zur Tragwerksanalyse. Dissertation, Veröffentlichungen - Institut für Statik und Dynamik der Tragwerke, Heft 19, TU Dresden, 2010
  • Freitag, S.; Graf, W.; Kaliske, M.: Identification and prediction of time-dependent structural behavior with recurrent neural networks for uncertain data. In: Beer, M.; Muhanna, R.L.; Mullen, R.L. (eds.): Proceedings of the 4th International Workshop on Reliable Engineering Computing (REC 2010), Singapore, 2010. Singapore : Research Publishing Services, 2010, pp. 577-596 – doi:10.3850/978-981-08-5118-7_026
  • Graf, W.; Freitag, S.; Kaliske, M.; Sickert, J.-U.: Recurrent Neural Networks for Uncertain Time-Dependent Structural Behavior. Computer-Aided Civil and Infrastructure Engineering 25 (2010) pp. 322-333 – doi:10.1111/j.1467-8667.2009.00645.x
  • Graf, W.; Freitag, S.; Sickert, J.-U.: Uncertain structural processes and neural network application. In: 4th European Conference on Computational Mechanics (ECCM 2010), Paris, 2010. – CD-ROM
  • Freitag, S.:Modellfreie numerische Prognosemethoden zur Tragwerksanalyse. Dissertation, Dresden: Fakultät Bauingenieurwesen, Technische Universität Dresden, 2010.

2009

  • Graf, W.; Jenkel, C.; Pannier, S.; Sickert, J.-U.; Steinigen, F.: Numerical structural monitoring with the uncertainty model fuzzy randomness. International Journal of Reliability and Safety 3 (2009) 1/2/3, pp. 218-234
  • Freitag, S.; Graf, W.; Kaliske, M.; Sickert, J.-U.: Prediction of Structural Behaviour with Recurrent Neural Networks for Fuzzy Data. In: Topping, B.H.V.; Tsompanakis, Y. (eds.): Proceedings of the First International Conference on Soft Computing Technology in Civil, Structural and Environmental Engineering, Funchal, 2009. Stirlingshire : Civil-Comp Press, 2009, Book of Abstracts, paper 28, Volltext (20 S.), CD-ROM – doi:10.4203/ccp.92.28
  • Oeser, M.; Freitag, S.: Modeling of materials with fading memory using neural networks. International Journal for Numerical Methods in Engineering 78 (2009) 7, pp. 843-862 – doi:10.1002/nme.2518
  • Freitag, S.; Beer, M.; Graf, W.; Kaliske, M.: Lifetime prediction using accelerated test data and neural networks. Computers & Structures 87 (2009) pp. 1187-1194 – doi:10.1016/j.compstruc.2008.12.007
  • Freitag, S.; Graf, W.; Kaliske, M.: Prognose des Langzeitverhaltens von Textilbeton-Tragwerken mit rekurrenten neuronalen Netzen. In: Curbach, M. (Hrsg.), Jesse, F. (Hrsg.): Textile Reinforced Structures : Proceedings of the 4th Colloquium on Textile Reinforced Structures (CTRS4) und zur 1. Anwendertagung, Dresden, 3.-5.6.2009. SFB 528, Technische Universität Dresden, D–01062 Dresden : Eigenverlag, 2009, S. 365-376 URN: urn:nbn:de:bsz:14-ds-1244048026002-79164
  • Freitag, S.; Graf, W., Kaliske, M.: Prognose zeitveränderlicher Strukturantworten mit rekurrenten neuronalen Netzen. In: Dinkler, D.; Zilian, A. (Hrsg.): Forschungskolloquium Baustatik-Baupraxis, Falkenstein, 2009. S. 45
  • Freitag, S.; Kaliske, M.; Graf, W.: Time-dependent reliability assessment of structures using uncertain fractional rheological models. Proceedings in Applied Mathematics and Mechanics 9 (2009) pp. 225-226 – doi:10.1002/pamm.200910086

2008

  • Kaliske, M.; Graf, W.: Numerisches Tragwerksmonitoring und Prognose des Tragwerkverhaltens. In: Wagner, W. (Hrsg.): Baustatik-Baupraxis 10. Bericht, Universität Karlsruhe (TH), 2008, S. 183-192
  • Möller, B.; Beer, M.: Engineering computation under uncertainty capabilities of non-traditional models. Computers & Structures 86 (2008) 10, pp. 1024-1041 – doi:10.1016/j.compstruc.2007.05.041
  • Graf, W.; Möller, B.; Bartzsch, M.: Uncertain processes and numerical monitoring of structures In: Muhanna, R.; Mullen, R.L. (eds.): Proceedings of 3rd Internat. Workshop Reliable Engineering Computing (REC), Georgia Tec, Savannah, 2008. pp. 155-170
  • Beer, M.; Liebscher, M.: Designing robust structures a nonlinear simulation based approach. Computers & Structures 86 (2008) 10, pp. 1102-1122 – doi:10.1016/j.compstruc.2007.05.037
  • Freitag, S.; Graf, W.; Pannier, S.; Sickert, J.-U.: Reliability of structures under consideration of uncertain time-dependent material behaviour. In: Dubois, D. et al. (eds.): Adcances in Soft Computing 48. Soft Methods for Handling Variability and Imprecision. Berlin : Springer, 2008, pp. 383-390 4th Internat. Conference on Soft Methods in Probability and Statistics, Toulouse, 2008
  • Beer, M.: Quantification of imprecise statistical data. In: Chang-Koon, C. (ed.): Proceedings of the 4th International Conference on Advances in Structural Engineering and Mechanics - ASEM'08, Jeju, Korea, 2008. pp. 1381-1394
  • Beer, M.: Evaluation of inconsistent engineering data. In: Muhanna, R.; Mullen, R.L. (eds.): Proceedings of 3rd Internat. Workshop Reliable Engineering Computing (REC), Georgia Tec, Savannah, 2008. pp. 481-498

2007

  • Möller, B.; Reuter, U.: Uncertainty Forecasting in Engineering. Berlin : Springer, 2007
  • Beer, M.: Model-free Sampling. Structural Safety 29 (2007) pp. 49–65
  • Spanos, P.D.; Beer, M.; Red-Horse, J.: Karhunen-Loéve Expansion of Stochastic Processes with a Modified Exponential Covariance Kernel. ASCE Journal of Engineering Mechanics 133 (2007) 7, pp. 773-779
  • Freitag, S.; Beer, M.; Graf, W.: Lifetime prediction with a neural network application. In: 2nd GACM Colloquium on Computational Mechanics, TU Munich, 2007. Book of Abstracts, p. 92
  • Graf, W.; Möller, B.; Bartzsch, M.: Alteration of Structures as an Uncertainty Process. In: Xie, M.; Patnaikuni, I. (eds.): Proceedings of the 4th International Structural Engineering and Construction Conference, Melbourne, 2007. pp. 791-796
  • Freitag, S.; Beer, M.; Graf, W.; Kaliske, M.: Lifetime prediction with neural networks. In: Topping, B.H.V. (ed.): Proceedings of the 9th International Conference on the Application of Artificial Intelligence to Civil, Structural and Environmental Engineering, St. Julians, 2007. Book of Abstracts and CD-ROM, Paper 35, 19 pp.

2006

  • Freitag, S.; Beer, M.; Jesse, F.; Weiland, S.: Experimental Investigation and Prediction of long-term Behavior of Textile Reinforced Concrete for Strengthening. In: Hegger, J.; Brameshuber, W.; Will, N. (eds.): Textile Reinforced Concrete – Proceedings of the 1st International RILEM Conference, Aachen, 2006. RILEM, pp. 121-130
  • Freitag, S.; Graf, W.; Hoffmann, A.; Pannier, S.; Sickert, J.-U.; Steinigen, F.: Tragwerke aus Textilbeton – unscharfe numerische Simulation. In: Ruge, P.; Graf, W. (eds.): 10. Dresdner Baustatik-Seminar. Dresden, 2006, S. 123-132 – PDF-Datei, 544 kB
  • Graf, W.; Bartzsch, M.; Beer, M.; Liebscher, M.; Reuter, U.: Numerisches Tragwerksmonitoring und Prognose. In: Ruge, P.; Graf, W. (eds.) 10. Dresdner Baustatik-Seminar. Dresden, 2006, S. 147-156
  • Möller, B.; Beer, M.; Graf, W.; Sickert, J.-U.: Time-Dependent Reliability of Textile-Strengthened RC Structures under Consideration of Fuzzy Randomness. Computers and Structures 84 (2006) no. 8–9, pp. 585–603
  • Sickert, J.-U.; Möller, B.; Graf, W.; Freitag, S.: Time-dependent reliability of strengthened rc structures. In: Hegger, J.; Brameshuber, W.; Will, N. (eds.): Textile Reinforced Concrete – Proceedings of the 1st International RILEM Conference, Aachen, 2006. RILEM, pp. 265-274
  • Beyer, W.; Liebscher, M.; Beer, M.; Graf, W.: Neural Network Based Response Surface Methods – a Comparative Study. In: Proceedings of the 5th German LS-DYNA Forum 2006. DYNAmore GmbH, Ulm, 2006, K-II-29 – K-II-37
  • Beer, M.: Sampling Without Probabilistic Model. In: Muhanna, R.; Mullen, R.L. (eds.): 2nd NSF Workshop on Reliable Engineering Computing. Georgia Institute of Technology, Savannah, GA, USA, 2006, pp. 369–390, CD-ROM
  • Graf, W.; Möller, B.; Beer, M.: Zum Einfluß der Datenbasis auf Tragwerkssicherheit und Versagensrisiko. Wissenschaftliche Zeitschrift der Technischen Universität Dresden 55 (2006) Heft 3–4, S. 49–53

2005

  • Beer, M.; Spanos, P. D.: Simulation based structural reliability assessment involving imprecise data. In: Augusti, G. (ed.); Schueller, G. I. (ed.); Ciampoli, M. (ed.): Safety and Reliability of Engineering Systems and Structures, Proceedings of the 9th Int. Conference on Structural Safety and Reliability, ICOSSAR'05, Rome, 2005. Millpress, Rotterdam, 2005, CD-ROM, Doc. MS0704, pp. 1725-1732
  • Beer, M.; Spanos, P. D.: Neural network based Monte Carlo simulation of random processes. In: Augusti, G. (ed.); Schueller, G. I. (ed.); Ciampoli, M. (ed.): Safety and Reliability of Engineering Systems and Structures, Proceedings of the 9th Int. Conference on Structural Safety and Reliability, ICOSSAR'05, Rome, 2005. Rotterdam : Millpress, 2005, CD-ROM, Doc. 017, pp. 2179-2186
  • Beer, M.: Simulation of Fuzzy Random Variables. In: Bathe, K. J. (ed.): Third M.I.T. Conference on Computational Fluid and Solid Mechanics, Cambridge, USA, 2005. Compilation of Abstracts, 31
  • Beer, M.; Spanos, P. D.: Neural Networks in Process Simulation. In: Bathe, K. J. (ed.): Third M.I.T. Conference on Computational Fluid and Solid Mechanics, Cambridge, USA, 2005. Compilation of Abstracts, 32

2004

  • Beer, M.: Sample-Induced Simulation of Fuzzy Randomness. In: Wojtkiewicz, S.; Red-Horse, J.; Ghanem, R. (eds.): 9th ASCE EMD/SEI/GI/AD Joint Specialty Conference on Probabilistic Mechanics and Structural Reliability, Albuquerque, NM, 2004. CD-ROM, Doc. 08_103, 6 pp.
  • Beer, M.; Spanos, P. D.: A Neural Network Approach for Representing Realizations of Random Processes. In: Wojtkiewicz, S.; Red-Horse, J.; Ghanem, R. (eds.): 9th ASCE EMD/SEI/GI/AD Joint Specialty Conference on Probabilistic Mechanics and Structural Reliability Albuquerque, NM, 2004. CD-ROM, Doc. 04_104, 6 pp.

About this page

Massivbau
Last modified: Aug 14, 2024