Filtern
Dokumenttyp
- Vortrag (4) (entfernen)
Sprache
- Englisch (4)
Referierte Publikation
- nein (4) (entfernen)
Schlagworte
- Model calibration (4) (entfernen)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (4)
In materials and component research, artificial intelligence methodologies will lead to massive upheavals in the coming years. The processes of material development, material processing, lifetime prediction and material characterization will change significantly. By combining AI methods and new forms of knowledge representation, the data-based management of product life cycles will take on new qualities. To address this emerging field of research Fraunhofer IWM set up the online workshop »AI Methods for Fatigue Behavior Assessment and Component Lifetime Prediction«
PGD model with domain mapping of Bead-on-Plate weld simulation for wire arc additive manufacturing
(2022)
Numerical simulations are essential in predicting the behavior of systems in many engineering fields and industrial sectors. The development of accurate virtual representations of actual physical products or processes allows huge savings in cost and resources. In fact, digital twins would allow reducing the number of real, physical prototypes, tests, and experiments, thus also increasing the sustainability of the production processes and products’ lifetime. Standard numerical methods fail in providing real time simulations, especially for complex processes such as additive manufacturing applications.
This work aims to build up a reduced order model for efficient wire arc additive manufacturing simulations by using the proper generalized decomposition (PGD) [1,2] method. Model order reduction is a popular concept to decrease the computational effort, where each evaluation of the reduced forward model is faster than evaluations using classical methods, even for complex models. The simulation of a moving heat source leads to a hardly separable parametric problem, which is solved by a new mapping approach [3]. Using this procedure, it is possible to create a simple separated representation of the forward model.
In this contribution, a PGD model is derived for the first part of wire arc additive manufacturing: bead-on-plate weld. An excellent agreement with a standard finite element method is shown. The reduced model is further used in a model calibration set up, speeding up calibrations and ultimately leading to an optimized real-time simulation.
Numerical simulations are essential in predicting the behavior of systems in many engineering fields and industrial sectors. The development of accurate virtual representations of actual physical products or processes (also known as digital twins) allows huge savings in cost and resources. In fact, digital twins would allow reducing the number of real, physical prototypes, tests, and experiments, thus also increasing the sustainability of production processes and products’ lifetime. Standard numerical methods fail in providing real time simulations, especially for complex processes such as additive manufacturing applications.
This work aims to use a reduced order model for efficient wire arc additive manufacturing simulations, calibrations and real-time process control. Model reduction, e.g. the proper generalized decomposition [1,2] method, is a popular concept to decrease the computational effort. A new mapping approach [3] was applied to simulate a moving heat source with the proper generalized decomposition. Using this procedure even complex models can be simulated in real-time. The physical model is later on calibrated with the use of a stochastic model updating process and the reduced order model, leading to an optimized real-time simulation.
In this contribution, a proper generalized decomposition model for a bead-on-plate wire arc additive manufacturing is presented. It is also coupled with a stochastic model updating process identifying the heat source characteristics as well as the boundary conditions of the transient thermal problem, whereas the heat source shape is simulated using a Goldak heat source
For extrusion-based 3D concrete printing, the early age mechanical behavior is influenced by various time dependent phenomena: structural build-up, plasticity as well as viscosity. The structural build-up is governing the stability and early-age strength development of the fresh printable cementitious materials and with that influencing the printability, buildability, and open time of the printing process. Generally, it is influenced by a number of factors, i.e. composition of the printable material, printing regime, and ambient conditions (temperature, humidity, etc.). There are several approaches to model the structural build-up of cementitious materials. All models are based on a time-dependent internal structural parameter describing the flocculation state, which is assumed to be zero after mixing and increases with time. The approaches differ in the definition of the time dependency (linear, exponential, bi-linear). Usually, the parameters are defined for a specific material composition without considering the influence of ambient conditions.
In this contribution, the bi-linear structural build-up model [Kruger et al., Construction and Building Materials 224, 2019] is extended by the temperature influence. Temperature changes will occur in real life printing processes due to changing ambient conditions (summer, winter, day, night) as well as the printing process (pressure changes etc.) and have a significant impact on the structural build-up process: an increase of the temperature leads to a faster dissolution of cement phases, accelerates hydration and boosts the Brownian motion. For that reason, the model parameters are simulated as temperature dependent using an Arrhenius function. Furthermore, the proposed extended model is calibrated based on measurement data using Bayesian inference. A very good agreement of the predicted model data with the measured control data was reached. Additionally, the structural build-up model is integrated into a viscoelastic and elastoplastic mechanical model, simulating the whole mechanical behavior during layer deposition.