Last modified: 2021-05-29
Abstract
Selective laser melting (SLM) is an additive manufacturing (AM) technique which fabricates 3D objects layer-by-layer by selectively melting powder with laser. As variety of metals and polymers are available in AM, many applications are possible especially in medical and aerospace arears. However, geometry of final products highly depends on manufacturing process, especially on a building direction. The geometrical imperfections have been observed in variety of structures made of metals and polymers. Such imperfection leads to large variability in the mechanical properties of the product. Thus, the prediction of mechanical properties is an important task in order to assure the quality of additively manufactured products. Our study discusses the geometrical accuracy of AM products.
The numerical simulation has potential in quality assurance to predict mechanical behaviors of additively manufactured products. The key point of the predictive simulation is to anticipate geometrical imperfections before manufacturing. The predictive models should be generated even for untested manufacturing conditions. From this requirement, the correlation between a process parameter and geometrical imperfections is a matter of great interest. Thus, our study aims at interpolating within the statistical database of geometrical measurement with respect to the process parameter. Focusing on a macroscopic distortion of additively manufactured circular holes, our study finds out the correlation with the building direction.
In this work, a prediction method using finite element method (FEM) was developed to investigate additively manufactured open-hole tensile (OHT) specimens of nylon, PA2200, considering geometrical imperfection associated with a building direction. Focus was given on a circular hole in components, which appears often in the light-weight component design. The imperfect circular hole has simply been modeled as ellipsoid, and representative geometrical parameters were defined. By measuring the specimens, it was found out that the length of the shortest ligament linearly depended on sine function of building direction with a correlation coefficient of 0.94. This extremely high correlation with sine function of a process parameter is a new finding in our study. The proposed method was validated by comparison with image-based analysis as well as experimental test results of OHT specimens. Finally, the constructed database of geometrical parameters could be interpolated in order to predict the distorted geometry and to perform probabilistic finite element analyses for the case of different and unmanufactured building directions.