Multi-objective optimization of wooden structures with polymorphic uncertain parameters
Abstract of the planned project content, taking into account the results from the first project phase
The main objectives of the SPP 1886 is the supply of novel structural design concepts and of models to account for polymorphic uncertain data and information. A fundamental aspect in structures with natural materials (e.g. timber) is the underlying uncertainty. Several types of uncertainties are apparent and not negligible, e.g. for material parameters, geometrical variables, loads and bearing conditions.
In context of imprecise probability, the different types of uncertainty characteristics are combined to yield polymorphic uncertainty. In this project, aleatoric uncertainty is described by randomness and epistemic uncertainties are taken into account as fuzzy quantities for the parametrization of randomness. This polymorphic uncertainty approach is referred to as fuzzy probability based randomness. Dependencies between and within the uncertain variables are considered. Spatial intra-dependencies are simulated by multivariate fuzzy probability based random fields and inter-dependencies are regarded as fuzzy interactions and crosscorrelations of random variables. The design of wooden structures is supported by optimization strategies. Uncertain design variables and a-priori variables are considered. Multiple competing optimization objectives, for instance maximization of bearing capabilities and minimization of the result uncertainty, are aimed at. Uncertain design variables are geometrical parameters, which influence global failure modes, e.g. the load capacity in the sense of bearing load, temperature and moisture. Several optimization objectives are investigated, e.g. ultimate load, critical displacements and acceptable dead load. Resulting from the first project phase, a framework for multi-objective optimization considering polymorphic uncertainty models with several dependencies is available for the second project phase. Both, Finite Element Methods and analytical solutions are applied, while glulam beams, timber boards and timber joints are focused as model problems. The objectives are supplemented and renewed for the second phase.
The development of novel methods and advanced algorithms is planned. Following the previous objectives, time dependencies as a further type of intra-dependency are in the major focus in the second phase of the project. The four major objectives are summarized subsequently: 1. multi-objective optimization aiming at decision making in design of wooden structures; 2. polymorphic uncertainty analyses with design variables and a-priori parameters; 3. time-dependent polymorphic uncertainty models with spatially depending intra-dependencies, parameter inter-dependencies and processes; 4. efficient processing of uncertainty analysis and multi-objective optimization for the applicability to real world problems.
Essential project goals and objectives
- Development of polymorphic uncertainty modeling considering time dependency as enhancement to the spatial depending intra-dependencies and parameter inter-dependencies.
- Development of polymorphic uncertainty analyses for the simulation and propagation of time dependency in timber structures.
- Development of multi-objective optimization strategies for decision making in design of timber structures considering time dependent polymorphic uncertainties.
- Improvements in numerical efficiency by advanced coupling of parallelization methods and surrogate models in polymorphic uncertainty analyses.
- Combination of the developments of both phases in this project for applications of the structural analysis framework on representative engineering tasks.
- Schietzold, F. N.; Graf, W.; Kaliske, M.:
Design Optimization of a Glulam Beam with Polymorphic Uncertain Parameters.
3rd International Conference on Uncertainty Quantification in Computational Sciences and Engineering (UNCECOMP2019), Crete, 2019 - Schietzold, F. N.; Schmidt, A.; Dannert, M. M.; Frau, A.; Fleury, R. M. N.; Graf, W.; Kaliske, M.; Könke, C.; Lahmer, T.; Nackenhorst, U.:
Development of fuzzy probability based random fields for the numerical structural design (GAMM - Mitteilungen), Wien 2019 Link - Schietzold, N.; Graf, W.; Kaliske, M.:
Polymorphic uncertainty modeling for optimization of timber structures.
In:Ferson, S. (ed.),Proceedings of 8th International Workshop on Reliable Engineering Computing (REC), pp. 63-70, University of Liverpool, 2018 - Schietzold, N.; Graf, W.; Kaliske, M.:
Polymorphic uncertainty modeling for optimization of timber structures GAMM-Tagung, TU München, 2018, Proceedings Applied Mathematics and Mechanics (PAMM) 18 (2018) 4 pp. - Leichsenring, F.; Jenkel, C.; Graf, W.; Kaliske, M.:
Numerical simulation of wooden structures with polymorphicuncertainty in material parameters, special issue onComputing with polymorphic uncertain data (REC 2016),International Journal of Reliability and Safety,Vol. 12 (2018) No. 1/2, pp. 24-45DOI: 10.1504/IJRS.2018.10013787 - Kaliske, M.; Jenkel, C; Leichsenring, F.; Graf, W.:
Numerical simulation of wooden structures with polymorphicuncertain data. In: Bucher, C. et al. (eds.),Proceedings 12thICOSSAR, pp. 2521-2530, TU Vienna, 2017