Data driven shrinkage compensation

1. A method for data driven shrinkage compensation to improve fabrication of objects using an additive manufacturing process, the method comprising:

subdividing, by a polygon subdivider, polygons in a three-dimensional file modeling an object into facets;

calculating, by an axis dimension calculator, dimensions of the object modeled in the three-dimensional file from an x-directional strand disposed between two facets of a first predetermined facet pair, a y-directional strand disposed between two facets of second predetermined facet pair, and a z-directional strand disposed between two facets of a third predetermined facet pair, the object being formed from at least some of the polygons;

predicting, by a dimension change predictor, dimensional changes in the x-directional strand, the y-directional strand, and the z-directional strand as a result of the fabrication of the object using the additive manufacturing process based on a shape shrinkage model; and

correcting, by a dimension change compensator, x-coordinate data, y-coordinate data, and z-coordinate data of at least one facet of the predetermined facet pairs to compensate for the predicted dimensional changes in the x-directional strand, the y-directional strand, and the z-directional strand.

2. The method of claim 1 , wherein said correcting step corrects the x-coordinate data, y-coordinate data, and z-coordinate data of both facets of at least one of the predetermined facet pairs.

3. The method of claim 2 , wherein said correcting step equally moves, by one half of a dimensional change, each of the facets of the at least one of the predetermined facet pairs.

4. The method of claim 1 , wherein said subdividing step further subdivides the polygons into vertices, and the method further comprises:

numbering each of the facets; and

numbering the vertices of each of the facets such that a same respective number is assigned to a same respective coordinate of vertices.

5. The method of claim 4 , further comprising:

setting sampling points on the polygons;

associating a facet number of a given facet with the x-coordinate data, y-coordinate data, and z-coordinate data of a sampling point included in the given facet.

6. The method of claim 5 , wherein each of the x-directional strand, the y-directional strand, and the z-directional strand are calculated from a respective pair of opposing sampling points.

7. The method of claim 6 , wherein a length of the x-strand is calculated from the respective pair of opposing sampling points with a same z-coordinate and y-coordinate, wherein a length of the y-strand is calculated from the respective pair of opposing sampling points with a same x-coordinate and z-coordinate, and wherein a length of the z-strand is calculated from the respective pair of opposing sampling points with a same x-coordinate and y-coordinate.

8. The method of claim 4 , wherein the sampling points are set using a ray intersection technique or point-in-polygon technique.

9. The method of claim 1 , further comprising deriving the shape shrinkage model from a test artifact, the test artifact having a plurality of shape variations.

10. The method of claim 9 , wherein the shape shrinkage model is derived using at least one of a kernel regression technique, a neural network, and a deep learning technique.

11. The method of claim 1 , wherein said deriving step comprises:

printing the test artifact using a three-dimensional printer;

acquiring dimensions of the test artifact as point cloud data using a three-dimensional scanner;

calculating, using the point cloud data, dimensions of the test artifact from an x-directional strand disposed between two facets of a first predetermined facet pair, a y-directional strand disposed between two facets of second predetermined facet pair, and a z-directional strand disposed between two facets of a third predetermined facet pair; and

generating the shape shrinkage model from dimensional changes in any of the x-directional strand, the y-directional strand, and the z-directional strand.

12. A non-transitory article of manufacture tangibly embodying a computer readable program which when executed causes a computer to perform the steps of claim 1 .

13. A system for data driven shrinkage compensation to improve fabrication of objects using an additive manufacturing process, the system comprising:

at least one processor operatively coupled to a memory; and

one or more tools implemented by the processor, including:

a polygon subdivider for subdividing polygons in a three-dimensional file modeling an object into facets;

an axis dimension calculator for calculating dimensions of the object modeled in the three-dimensional file from an x-directional strand disposed between two facets of a first predetermined facet pair, a y-directional strand disposed between two facets of second predetermined facet pair, and a z-directional strand disposed between two facets of a third predetermined facet pair, the object being formed from at least some of the polygons;

a dimension change predictor for predicting dimensional changes in the x-directional strand, the y-directional strand, and the z-directional strand as a result of the fabrication of the object using the additive manufacturing process based on a shape shrinkage model; and

a dimension change compensator for correcting x-coordinate data, y-coordinate data, and z-coordinate data of at least one facet of the predetermined facet pairs to compensate for the predicted dimensional changes in the x-directional strand, the y-directional strand, and the z-directional strand.

14. The system of claim 13 , wherein said dimension change compensator corrects the x-coordinate data, y-coordinate data, and z-coordinate data of both facets of at least one of the predetermined facet pairs.

15. The system of claim 14 , wherein said dimension change compensator equally moves, by one half of a dimensional change, each of the facets of the at least one of the predetermined facet pairs.

16. The system of claim 13 , wherein said polygon subdivider further subdivides the polygons into vertices, and the system further comprises a polygon labeler for numbering each of the facets, and numbering the vertices of each of the facets such that a same respective number is assigned to a same respective coordinate of vertices.

17. The system of claim 13 , wherein each of the x-directional strand, the y-directional strand, and the z-directional strand are calculated from a respective pair of opposing sampling points.

18. The system of claim 13 , wherein a length of the x-strand is calculated from a respective pair of opposing sampling points with a same z-coordinate and y-coordinate, wherein a length of the y-strand is calculated from a respective pair of opposing sampling points with a same x-coordinate and z-coordinate, and wherein a length of the z-strand is calculated from a respective pair of opposing sampling points with a same x-coordinate and y-coordinate.

19. The system of claim 13 , further comprising a shape shrinkage model generator for deriving the shape shrinkage model from a test artifact, the test artifact having a plurality of shape variations.

20. The system of claim 19 , wherein the shape shrinkage model is derived using at least one of a kernel regression technique, a neural network, and a deep learning technique.