Object model dimensions for additive manufacturing

Abstract

In an example, a method includes generating a training dataset for an inference model to generate dimensional modifications to apply to object models to compensate for departures from model dimensions in objects generated using additive manufacturing based on those object models. Generating the training dataset may include acquiring, for each of a plurality of generated objects, (i) an indication of a first dimensional inaccuracy and a second dimensional inaccuracy, wherein the first and second dimensional inaccuracies are acquired in a direction of a first axis and relate to respective first and second object dimensions; and (ii) an indication of object placement within a fabrication chamber during object generation.

1 . A method for generating objects with an additive manufacturing apparatus by fusing build material within a fabrication chamber, the method comprising:
 generating a plurality of training objects in different positions within the fabrication chamber;   measuring a dimension of each of the generated training objects; and   based on the measuring, generating a training dataset for an inference model to generate dimensional modifications to apply to an object model to compensate for departures from model dimensions associated with a position of an object in the fabrication chamber, wherein generating the training dataset comprises:   acquiring, for each of the plurality of generated training objects:
 an indication of a first dimensional inaccuracy and a second dimensional inaccuracy, wherein the first and second dimensional inaccuracies are acquired in a direction of a first axis and relate to respective first and second object dimensions; and 
 an indication of a position of the training object within the fabrication chamber during object generation. 
   
     
     
         2 . A method according to  claim 1  further comprising using the training dataset to generate the inference model including at least one of a scaling factor and an offset factor to apply to the object model. 
     
     
         3 . A method according to  claim 2  further comprising:
 applying the inference model to the object model to determine a modified object model; and 
 generating the object from the modified object model. 
 
     
     
         4 . A method according to  claim 1  further comprising acquiring an indication of a third dimensional inaccuracy and a fourth dimensional inaccuracy, wherein the third and fourth dimensional inaccuracies are acquired in a direction of a second axis, which is orthogonal to the first axis and relate to respective third and fourth object dimensions. 
     
     
         5 . A method according to  claim 1  wherein the plurality of generated training objects comprises at least one of:
 at least one instance of a first training object and at least one instance of a second training object, wherein the first and second training objects have different length:width:height ratios; and 
 a plurality of instances of a third training object and a plurality of instances of a fourth training object, wherein the third and fourth training objects have different solid proportions and different instances of each training object are generated in different positions within the fabrication chamber. 
 
     
     
         6 . A method according to  claim 1  further comprising generating a plurality of training datasets relating to at least one of: different environmental conditions, different object generation apparatus, different object generation material compositions, different object cooling profiles, and different print modes. 
     
     
         7 - 15 . (canceled) 
     
     
         16 . A tangible non-transitory machine readable medium for generating objects with an additive manufacturing apparatus by fusing build material within a fabrication chamber, the medium having instructions that, when executed:
 cause the additive manufacturing apparatus to generate a plurality of training objects in different positions within the fabrication chamber; and   cause a processor to generate a training dataset for an inference model to generate dimensional modifications to apply to an object model to compensate for departures from model dimensions associated with a position of an object in the fabrication chamber, wherein the instructions to cause the processor to generate the training dataset comprise instructions to cause the processor to:   acquire, for each of the plurality of generated training objects:
 an indication of a first dimensional inaccuracy and a second dimensional inaccuracy, wherein the first and second dimensional inaccuracies are acquired in a direction of a first axis and relate to respective first and second object dimensions; and 
 an indication of a position of the training object within the fabrication chamber during object generation; and 
   cause the processor to generate the inference model using the training dataset.   
     
     
         17 . A medium according to  claim 16 , further comprising instructions to:
 cause a processor to apply the inference model to the object model to determine a modified object model; and   cause the additive manufacturing apparatus to generate the object from the modified object model.   
     
     
         18 . A medium according to  claim 17 , wherein the inference model includes at least one of a scaling factor and an offset factor. 
     
     
         19 . A medium according to  claim 16  residing on the additive manufacturing apparatus. 
     
     
         20 . An additive manufacturing apparatus comprising:
 a fabrication chamber;   a print bed in the fabrication chamber;   a build material distribution system to layer build material over the print bed;   a printhead to distribute a fusing agent on build material layered over the print bed;   an energy source to apply energy to build material treated with the fusing agent; and   processing circuitry to:
 generate a plurality of training objects on the print bed in different positions within the fabrication chamber; 
 based on dimensional inaccuracies in the training objects, generate a training dataset for an inference model to generate dimensional modifications to apply to an object model to compensate for departures from model dimensions associated with a position of an object in the fabrication chamber; 
 generate the inference model using the training dataset; 
 apply the inference model to the object model to determine a modified object model; and 
 generate the object from the modified object model. 
   
     
     
         21 . An apparatus according to  claim 20  wherein the processing circuitry is to generate the training dataset by acquiring, for each of the plurality of generated training objects:
 an indication of a first dimensional inaccuracy and a second dimensional inaccuracy, wherein the first and second dimensional inaccuracies are acquired in a direction of a first axis and relate to respective first and second object dimensions; and 
 an indication of a position of the training object within the fabrication chamber during object generation.