Refine
Document Type
- Conference Proceeding (51)
- Article (41)
- Doctoral Thesis (23)
- Master's Thesis (3)
- Bachelor Thesis (2)
- Periodical (2)
- Report (2)
- Book (1)
Institute
- Institut für Strukturmechanik (ISM) (40)
- In Zusammenarbeit mit der Bauhaus-Universität Weimar (18)
- Professur Angewandte Mathematik (9)
- Bauhaus-Institut für Geschichte und Theorie der Architektur und Planung (8)
- Professur Stochastik und Optimierung (7)
- F. A. Finger-Institut für Baustoffkunde (FIB) (6)
- Junior-Professur Computational Architecture (6)
- Institut für Europäische Urbanistik (5)
- Graduiertenkolleg 1462 (4)
- Professur Informatik im Bauwesen (4)
Keywords
- Angewandte Mathematik (65)
- Building Information Modeling (36)
- Angewandte Informatik (35)
- Computerunterstütztes Verfahren (35)
- Data, information and knowledge modeling in civil engineering; Function theoretic methods and PDE in engineering sciences; Mathematical methods for (robotics and) computer vision; Numerical modeling in engineering; Optimization in engineering applications (34)
- Strukturmechanik (31)
- Architektur (9)
- Städtebau (7)
- Stochastik (5)
- Architecture (2)
Year of publication
- 2015 (125) (remove)
A topology optimization method has been developed for structures subjected to multiple load cases (Example of a bridge pier subjected to wind loads, traffic, superstructure...). We formulate the problem as a multi-criterial optimization problem, where the compliance is computed for each load case. Then, the Epsilon constraint method (method proposed by Chankong and Haimes, 1971) is adapted. The strategy of this method is based on the concept of minimizing the maximum compliance resulting from the critical load case while the other remaining compliances are considered in the constraints. In each iteration, the compliances of all load cases are computed and only the maximum one is minimized. The topology optimization process is switching from one load to another according to the variation of the resulting compliance. In this work we will motivate and explain the proposed methodology and provide some numerical examples.
Nanostructured materials are extensively applied in many fields of material science for new industrial applications, particularly in the automotive, aerospace industry due to their exceptional physical and mechanical properties. Experimental testing of nanomaterials is expensive, timeconsuming,challenging and sometimes unfeasible. Therefore,computational simulations have been employed as alternative method to predict macroscopic material properties. The behavior of polymeric nanocomposites (PNCs) are highly complex.
The origins of macroscopic material properties reside in the properties and interactions taking place on finer scales. It is therefore essential to use multiscale modeling strategy to properly account for all large length and time scales associated with these material systems, which across many orders of magnitude. Numerous multiscale models of PNCs have been established, however, most of them connect only two scales. There are a few multiscale models for PNCs bridging four length scales (nano-, micro-, meso- and macro-scales). In addition, nanomaterials are stochastic in nature and the prediction of macroscopic mechanical properties are influenced by many factors such as fine-scale features. The predicted mechanical properties obtained by traditional approaches significantly deviate from the measured values in experiments due to neglecting uncertainty of material features. This discrepancy is indicated that the effective macroscopic properties of materials are highly sensitive to various sources of uncertainty, such as loading and boundary conditions and material characteristics, etc., while very few stochastic multiscale models for PNCs have been developed. Therefore, it is essential to construct PNC models within the framework of stochastic modeling and quantify the stochastic effect of the input parameters on the macroscopic mechanical properties of those materials.
This study aims to develop computational models at four length scales (nano-, micro-, meso- and macro-scales) and hierarchical upscaling approaches bridging length scales from nano- to macro-scales. A framework for uncertainty quantification (UQ) applied to predict the mechanical properties
of the PNCs in dependence of material features at different scales is studied. Sensitivity and uncertainty analysis are of great helps in quantifying the effect of input parameters, considering both main and interaction effects, on the mechanical properties of the PNCs. To achieve this major
goal, the following tasks are carried out:
At nano-scale, molecular dynamics (MD) were used to investigate deformation mechanism of glassy amorphous polyethylene (PE) in dependence of temperature and strain rate. Steered molecular dynamics (SMD)were also employed to investigate interfacial characteristic of the PNCs.
At mico-scale, we developed an atomistic-based continuum model represented by a representative volume element (RVE) in which the SWNT’s properties and the SWNT/polymer interphase are modeled at nano-scale, the surrounding polymer matrix is modeled by solid elements. Then, a two-parameter model was employed at meso-scale. A hierarchical multiscale approach has been developed to obtain the structure-property relations at one length scale and transfer the effect to the higher length
scales. In particular, we homogenized the RVE into an equivalent fiber.
The equivalent fiber was then employed in a micromechanical analysis (i.e. Mori-Tanaka model) to predict the effective macroscopic properties of the PNC. Furthermore, an averaging homogenization process was also used to obtain the effective stiffness of the PCN at meso-scale.
Stochastic modeling and uncertainty quantification consist of the following ingredients:
- Simple random sampling, Latin hypercube sampling, Sobol’ quasirandom sequences, Iman and Conover’s method (inducing correlation in Latin hypercube sampling) are employed to generate independent and dependent sample data, respectively.
- Surrogate models, such as polynomial regression, moving least squares (MLS), hybrid method combining polynomial regression and MLS, Kriging regression, and penalized spline regression, are employed as an approximation of a mechanical model. The advantage of the surrogate models is the high computational efficiency and robust as they can be constructed from a limited amount of available data.
- Global sensitivity analysis (SA) methods, such as variance-based methods for models with independent and dependent input parameters, Fourier-based techniques for performing variance-based methods and partial derivatives, elementary effects in the context of local SA, are used to quantify the effects of input parameters and their interactions on the mechanical properties of the PNCs. A bootstrap technique is used to assess the robustness of the global SA methods with respect to their performance.
In addition, the probability distribution of mechanical properties are determined by using the probability plot method. The upper and lower bounds of the predicted Young’s modulus according to 95 % prediction intervals were provided.
The above-mentioned methods study on the behaviour of intact materials. Novel numerical methods such as a node-based smoothed extended finite element method (NS-XFEM) and an edge-based smoothed phantom node method (ES-Phantom node) were developed for fracture problems. These methods can be used to account for crack at macro-scale for future works. The predicted mechanical properties were validated and verified. They show good agreement with previous experimental and simulations results.
The 20th International Conference on the Applications of Computer Science and Mathematics in Architecture and Civil Engineering will be held at the Bauhaus University Weimar from 20th till 22nd July 2015. Architects, computer scientists, mathematicians, and engineers from all over the world will meet in Weimar for an interdisciplinary exchange of experiences, to report on their results in research, development and practice and to discuss. The conference covers a broad range of research areas: numerical analysis, function theoretic methods, partial differential equations, continuum mechanics, engineering applications, coupled problems, computer sciences, and related topics. Several plenary lectures in aforementioned areas will take place during the conference.
We invite architects, engineers, designers, computer scientists, mathematicians, planners, project managers, and software developers from business, science and research to participate in the conference!
Some caad packages offer additional support for the optimization of spatial configurations, but the possibilities for applying optimization are usually limited either by the complexity of the data model or by the constraints of the underlying caad system. Since we missed a system that allows to experiment with optimization techniques for the synthesis of spatial configurations, we developed a collection of methods over the past years. This collection is now combined in the presented open source library for computational planning synthesis, called CPlan. The aim of the library is to provide an easy to use programming framework with a flat learning curve for people with basic programming knowledge. It offers an extensible structure that allows to add new customized parts for various purposes. In this paper the existing functionality of the CPlan library is described.
From the design experiences of arch dams in the past, it has significant practical value to carry out the shape optimization of arch dams, which can fully make use of material characteristics and reduce the cost of constructions. Suitable variables need to be chosen to formulate the objective function, e.g. to minimize the total volume of the arch dam. Additionally a series of constraints are derived and a reasonable and convenient penalty function has been formed, which can easily enforce the characteristics of constraints and optimal design. For the optimization method, a Genetic Algorithm is adopted to perform a global search. Simultaneously, ANSYS is used to do the mechanical analysis under the coupling of thermal and hydraulic loads. One of the constraints of the newly designed dam is to fulfill requirements on the structural safety. Therefore, a reliability analysis is applied to offer a good decision supporting for matters concerning predictions of both safety and service life of the arch dam. By this, the key factors which would influence the stability and safety of arch dam significantly can be acquired, and supply a good way to take preventive measures to prolong ate the service life of an arch dam and enhances the safety of structure.
Superplasticizers are utilized both to improve the fluidity during the placement and to reduce the water content of concretes. Both effects have also an impact on the properties of the hardened concrete. As a side effect the presence of superplasticizers affects the strength development of concretes that is strongly retarded. This may lead to an ecomomical drawback of the concrete manufacturing. The present work is aimed at gaining insights on the causes of the retarding effect of superplasticizers on the hydration of Portland cement. In order to simplify the complex interactions occurring during the hydration of Portland cement the majority of the work focuses on the interaction of superplasticizer and tricalcium silicate (Ca3SiO5 or C3S, the main compound of Portland cement clinker). The tests are performed in three main parts accompanied by methods as for example isothermal conduction calorimetry, electrical conductivity, Electron Microscopy, ICP-OES, TOC, as well as Analytical Ultracentrifugation.
In the first main part and based on the interaction of cations and anionic charges of polymers, the interactions between calcium ions and superplasticizers are investigated. As a main effect calcium ions are complexed by the functional groups of the polymers (carboxy, sulfonic). Calcium ions may be both dissolved in the aqueous phase and a constitute of particle interfaces. Besides these effects it is furthermore shown that superplasticizers induce the formation of nanoscaled particles which are dispersed in the aqueous phase (cluster formation). Analogous to recent findings in the field of biomineralization, it is reasonable to assume that these nanoparticles influence the crystal growth by their assembly process.
Based on the assumption that superplasticizers hinder either or both dissolution and precipitation and by that retard the cement hydration, the impact on separate reactions is investigated. On experiments that address the solubility of C-S-H phases and portlandite, it is shown that complexation of calcium ions in the aqueous phase by functional groups of polymers increases the solubility of portlandite. Contrary, in case of C-S-H solubility the complexation of calcium ions in solution leads to decrease of the calcium ion concentration in the aqueous phase. These effects are explained by differences in adsorption of polymers on C-S-H phases and portlandite. It is proposed that adsorption is stronger on C-S-H phases compared to portlandite due to the increased specific surface area of C-S-H phases. Following that, it is claimed that before polymers are able to adsorb on C-S-H phases the functional groups must be screened by calcium ions in the aqueous phase. It is further shown that data regarding the impact of superplasticizers on the unconstrained dissolution rate of C3S does not provide a clear relation to the overall retarding effect occurring during the hydration of C3S. Both increased and decreased dissolution rate with respect to the reference sample are detected. If the complexation capability of the superplasticizers is considered then also a reduced dissolution rate of C3S is determined. Despite the fact that the global hydration process is accelerated, the addition of calcite leads to a slower dissolution rate. Thus, a hindered unconstrained dissolution of C3S as possibly cause for the retarding effect still remains open for discussion. In the last section of this part, the pure crystallization of hydrate phases (C-S-H phases, portlandite) is fathomed. Results clearly show that superplasticizers prolong the induction time and modify the rate of crystal growth during pure crystallization in particular due to the complexation of ions in solution. But this effect is insufficient to account for the overall retarding effect. Further important factors are the blocking of crystal growth faces by adsorbed polymers and the dispersion of nanoscaled particles which hinders their agglomeration in order to build up crystals.
In the last main part of the work, the previously gathered results are utilized in order to investigate hydration kinetics. During hydration, dissolution and precipitation occur in parallel. Thereby, special attention is laid on the ion composition of the aqueous phase of C3S pastes and suspensions in order to determine the rate limiting step. All in all it is concluded that the retarding effect of superplasticizers on the hydration of tricalcium silicate is based on the retardation of crystallization of hydrate phases (C-S-H phases and portlandite). Thereby, the two effects complexation of calcium ions on surfaces and stabilization of nanoscaled particles are of major importance. These mechanisms may partly be compensated by template performance and increase in solubility by complexation of ions in solution. The decreased dissolution rate of C3S by the presence of superplasticizers during the in parallel occuring hydration process can only be assessed indirectly by means of the development of the ion concentrations in the aqueous phase (reaction path). Whether this observation is the cause or the consequence within the dissolution-precipitation process and therefore accounts for the retarding effect remains a topic for further investigations.
Besides these results it is shown that superplasticizers can be associated chemically with inhibitors because they reduce the frequency factor to end the induction period. Because the activation energy is widely unaffected it is shown that the basic reaction mechanism sustain. Furthermore, a method was developed which permits for the first time the determination of ion concentrations in the aqueous phase of C3S pastes in-situ. It is shown that during the C3S hydration the ion concentration in the aqueous phase is developed correspondingly to the heat release rate (calorimetry). The method permits the differentiation of the acceleration period in three stages. It is emphasized that crystallization of the product phases of C3S hydration, namely C-S-H phases and portlandite, are responsible for the end of the induction period.
In this study, an application of evolutionary multi-objective optimization algorithms on the optimization of sandwich structures is presented. The solution strategy is known as Elitist Non-Dominated Sorting Evolution Strategy (ENSES) wherein Evolution Strategies (ES) as Evolutionary Algorithm (EA) in the elitist Non-dominated Sorting Genetic algorithm (NSGA-II) procedure. Evolutionary algorithm seems a compatible approach to resolve multi-objective optimization problems because it is inspired by natural evolution, which closely linked to Artificial Intelligence (AI) techniques and elitism has shown an important factor for improving evolutionary multi-objective search. In order to evaluate the notion of performance by ENSES, the well-known study case of sandwich structures are reconsidered. For Case 1, the goals of the multi-objective optimization are minimization of the deflection and the weight of the sandwich structures. The length, the core and skin thicknesses are the design variables of Case 1. For Case 2, the objective functions are the fabrication cost, the beam weight and the end deflection of the sandwich structures. There are four design variables i.e., the weld height, the weld length, the beam depth and the beam width in Case 2. Numerical results are presented in terms of Paretooptimal solutions for both evaluated cases.