There are some analytical programs and numerical modeling software, including COMSOL, 3DSIM, ABAQUS, MARC, ANSYS (using stress analyses), and many other 3D or finite element based methods, which have already been used to find a relation between the parameters of AM processes (e.g. hatch distance, scan speed, and device power) and the properties of final products. It seems that previous researchers could already partially simulate some characteristics of manufactured products by the help of design of experiments/DOE methods (e.g. TAGUCHI) as well as microstructural studies (considering droplet geometry, phases characteristics, grain size, defects, etc.), thermal evaluations (CTE, heat capacity, conductivity, molten pool temperature distribution, etc.), or stress analyses. However, there is still no comprehensive method to precisely evaluate the influence of AM parameters and also the magnitude of their impact on the mechanical properties of the products.
Please let me know what method you think is the most complete, practical, and reliable approach which can best simulate the final properties of the additively-manufactured products using the process parameters.
The mechanic properties involve many aspects of materials so that it is hard to establish single modeling for comprehensive evaluation. It is better to specify which property is the object of your research. Concerning modeling efforts, stress and strain have been successfully assessed, but for the resultant mechanical properties, maybe we can not obtain a scale value without the discussion of microstructure. Thus, modeling the microstructure at first and then applying an analytical model that states the relationship between microstructure and property would be an indirect route affording your purpose. So phase-field or cellular automaton may be the core section in your modeling section. For connecting the AM parameters and formation conditions of microstructures, multi-physics modeling tool (e.g., COMSOL) would be applied at the first stage.
Although I raised the above question in a general manner, my focus is mainly on manufacturing NiTi parts using SLM process.
Attached, you can find the initial framework (it's surely incomplete and requires more details) which came to my mind addressing the process-structure-property relationship.
The mechanic properties involve many aspects of materials so that it is hard to establish single modeling for comprehensive evaluation. It is better to specify which property is the object of your research. Concerning modeling efforts, stress and strain have been successfully assessed, but for the resultant mechanical properties, maybe we can not obtain a scale value without the discussion of microstructure. Thus, modeling the microstructure at first and then applying an analytical model that states the relationship between microstructure and property would be an indirect route affording your purpose. So phase-field or cellular automaton may be the core section in your modeling section. For connecting the AM parameters and formation conditions of microstructures, multi-physics modeling tool (e.g., COMSOL) would be applied at the first stage.
Dear Dr. Yufan Zhao
Thanks for you response.
As our first step, we are going to predict the Fatigue properties based on SLM parameters, the most applicable feature which must be considered for biomedical implants. It seems we need numerical modelings in both areas you have already stated to simulate the fatigue life and properties; i.e., stress analyses and microstructural studies.
Dear Mahdi,
I think the two main issues regarding modelling of metal AM process are its multi-physics and multi-scale nature. In this respect, not only you need to think of different physics affecting each other, but also you need to consider their domain of occurrence.
Typically there are three distinct zones identified for metal AM models : micro-scale, which deals with micro-structure and even grain growth, meso-scale which takes into account the melt pool dynamics and finally macro-scale, which mostly covers thermo-mechanical analysis that predicts residual stresses and resultant deflections. Also the last one has a sub-field, called part-scale models that are computationally more reasonable while enabling one to predict deflections of a real-size sample.
You need to find a smart way to study the effect of process conditions on the micro-structures and also grain's orientation and morphology or maybe porosity evolution. Then based on that (micro and meso-scale models) you can make a Representative Volume Element (RVE) that you can obtain your average engineering properties based on, such as average Young's modulus, average Poisson's ratio, average tangent modulus, etc.
Then you can apply this obtained averaged mechanical behavior to your fatigue model, which is going to be in the macro-scale.
Best regards,
Mohamad
Dear Dr. Mohamad Bayat
So thanks for your helpful comments.
If I understood you correctly, it seems we should firstly model the effect of process parameters on the mentioned micro/meso-scale features. Afterward, by designing some experiments in different process conditions and collecting the mechanical properties of manufactured parts, we can find the relations between "parameter & those features" and therefore "parameters & properties" which this makes us capable of tailoring the properties of new products based on AM parameters.
(like what the attachment says; source: A. Laukkanen, "Multiscale Materials Modeling for Metal Additive Manufacturing, 2016)
Best regards,
Mahdi
Hi again,
Yes, for the first step you need to do a parametric study on your meso/micro models, if you want to go for a pure modelling approach. Something like what we did recently [1]. Then you can use simulations or experiments to find the average mechanical properties. By simulations you need to make an RVE out of the data from meso and micro models, while with the experiments you can simply perform some tensile tests and there is no need of any numerical model until this step. Then you can apply that mechanical model obtained from the RVE to your fatigue model, which is much bigger in size I guess.
[1] Bayat, Mohamad, Sankhya Mohanty, and Jesper Henri Hattel. "A systematic investigation of the effects of process parameters on heat and fluid flow and metallurgical conditions during laser-based powder bed fusion of Ti6Al4V alloy." International Journal of Heat and Mass Transfer 139 (2019): 213-230.
Best regards,
Mohamad
- Badges
- Science topic