Predicting process control parameters for fabricating an object using deposition
Process control parameters are predicted to fabricate an object using deposition. An input design geometry is provided for the object. A training data set includes past post-build physical inspection data for a plurality of objects that comprise at least one object that is different from the object to be physically fabricated; and training data generated through a repetitive process of randomly choosing values for each of multiple process control parameters and scoring adjustments to the multiple process control parameters as leading to either undesirable or desirable outcomes, the outcomes based respectively on the presence or absence of defects detected in a fabricated object arising from the process control parameter adjustments. A machine learning algorithm is trained using the provided training data set and a predicted optimal set of the multiple process control parameters is generated for initiating and performing the deposition process to fabricate the object.
1 . An additive manufacturing system, the system comprising:
a robotic additive manufacturing deposition system capable of performing a deposition process using instructions, and the robotic additive manufacturing deposition system comprising a sensor capable of generating sensor data; and
one or more processors, wherein the one or more processors are configured to:
receive a toolpath, the toolpath capable of being utilized to fabricate an object using the deposition process;
receive sensor data from the sensor, wherein the sensor data is provided as an input to a classifier, wherein:
the classifier comprises a set of parameters; and
the classifier is capable of outputting a real-time classification of detected object defects based upon the input to the classifier and the set of parameters;
generate instructions for performing an additive manufacturing process to fabricate the object, wherein the instructions are based on the real-time classification of detected object defects; and
control the robotic additive manufacturing deposition system based on the generated instructions to fabricate the object using the deposition process.
2 . The additive manufacturing system of claim 1 , wherein the sensor data is real-time sensor data.
3 . The additive manufacturing system of claim 1 , wherein the sensor data is generated based on observation of a deposition site, wherein the deposition site is a location at which the robotic additive manufacturing deposition system is depositing material.
4 . The additive manufacturing system of claim 1 , wherein the sensor data comprises one or more object properties of the object.
5 . The additive manufacturing system of claim 1 , wherein the one or more processors are further configured to perform a noise removal technique to the sensor data before the sensor data is provided as an input to the classifier.
6 . The additive manufacturing system of claim 1 , wherein the classifier is characterized in that the set of parameters were determined using a machine learning method.
7 . The additive manufacturing system of claim 6 , wherein the machine learning method is a convolutional neural network.
8 . The additive manufacturing system of claim 1 , wherein the instructions are generated based on a predicted set of process control parameters.
9 . The additive manufacturing system of claim 8 , wherein the predicted set of process control parameters are adjusted in real-time based on one or more outputs from the classifier.
10 . The additive manufacturing system of claim 1 , wherein the classifier further comprises activation functions.
11 . The additive manufacturing system of claim 8 , wherein the predicted set of process control parameters were determined based on an output of a second classifier, wherein the second classifier comprises a second set of parameters, and wherein the second set of parameters of the second classifier were determined using a machine learning process based on training data generated through a repetitive process of randomly choosing values for each of at least one of process control parameter adjustments and scoring adjustments to the process control parameters as leading to either undesirable or desirable outcomes, the desirability of outcomes determined based respectively on a presence or absence of defects detected in a fabricated object arising from the process control parameter adjustments.
12 . The additive manufacturing system of claim 1 , wherein the set of parameters of the classifier were determined using a machine learning process based upon training data comprising past process simulation data, past process characterization data, past in-process inspection data, and past post-build inspection data, for a plurality of objects that comprise at least one second object that is different from the first object.
13 . The additive manufacturing system of claim 1 , wherein the set of parameters of the classifier were determined using a machine leaning process based on training data, the training data comprising past sensor data generated for at least one past object, the at least one past object different from the first object.
14 . An additive manufacturing system, the system comprising:
a robotic additive manufacturing deposition system capable of performing a deposition process using instructions, and the robotic additive manufacturing deposition system comprising a sensor capable of generating sensor data; and
one or more processors, wherein the one or more processors are configured to:
receive a toolpath, the toolpath capable of being utilized to fabricate an object using the deposition process;
receive sensor data from the sensor, wherein the sensor data is provided as an input to a model, wherein:
the model comprises a set of model parameters; and
the model is capable of outputting a real-time set of process parameters based upon the input to the model and the set of model parameters;
generate a predicted set of process control parameters using the model;
generate instructions for performing an additive manufacturing process to fabricate the object, wherein the instructions are based on at least the set of process control parameters; and
control the robotic additive manufacturing deposition system based on the generated instructions to fabricate the object using the deposition process.
15 . The additive manufacturing system of claim 14 , wherein the sensor data is real-time sensor data.
16 . The additive manufacturing system of claim 14 , wherein the model is characterized in that the model parameters were determined using a machine learning method.
17 . The additive manufacturing system of claim 16 , wherein the machine learning method is a support vector machine.
18 . The additive manufacturing system of claim 14 , wherein the model further comprises activation functions.
19 . The additive manufacturing system of claim 14 , wherein the set of model parameters were determined using a machine learning process based on training data generated through a repetitive process of randomly choosing values for each of at least one of process control parameters adjustments and scoring adjustments to the process control parameters as leading to either undesirable or desirable outcomes, the desirability of outcomes determined based respectively on a presence or absence of defects detected in a fabricated object arising from the process control parameter adjustments.
20 . The additive manufacturing system of claim 14 , wherein a set of classifier parameters of a classifier were determined using a machine learning process based upon training data comprising past process simulation data, past process characterization data, past in-process inspection data, and past post-build inspection data, for a plurality of objects that comprise at least one second object that is different from the first object and wherein the classifier is capable of outputting a real-time classification of detected object defects based upon an input to the classifier and the set of classifier parameters.
21 . The additive manufacturing system of claim 14 , wherein the set of model parameters of the model were determined using a machine learning process based on training data, the training data comprising past sensor data generated for at least one past object, the at least one past object different from the first object.