Machine learning based on virtual (V) and real (R) data

1. A method comprising:

simulating, with computer-based software, a manufacturing process, to produce virtual data about the process, wherein the virtual data about the process comprises physics-based features that represent energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process;

identifying process parameters for a real-world version of the process,

providing a real-world sensor to sense a parameter associated with the real-world version of the process;

receiving sensor readings from the real-world sensor while the real-world version of the process is being performed;

training a machine-learning software model to predict a behavior of the real-world sensor in a subsequent real-world manufacturing process based on the virtual data about the process, the process parameters, and the sensor readings to produce a trained machine-learning software model, wherein the virtual data about the process provided to train the machine-learning software model to predict the behavior of the real-world sensor comprises high fidelity physics information generated based on a computer-aided design (CAD) model of an object to be manufactured using the process; and

predicting a behavior of the real-world version of the sensor in the subsequent real-world manufacturing process with the trained machine-learning software model.

2. The method of claim 1 , wherein simulating the process with the computer-based software application to produce the virtual data about the process comprises:

providing a computer-aided design (CAD) model of a part associated with the process; and

producing virtual data about a machine tool path for the process with computer-based software based on the CAD model.

3. The method of claim 2 , wherein the process is a manufacturing process and the machine tool path is for a machine to manufacture a part represented in the CAD model.

4. The method of claim 3 , wherein the manufacturing process is an additive manufacturing process, and wherein the machine comprises a build platform, a powder bed atop the build platform; and a heat source configured to deliver heat into various portions of the powder bed to create the part.

5. The method of claim 4 , wherein the virtual data about the machine tool path comprises power data, speed data, and pattern data for the heat source in the additive manufacturing process.

6. The method of claim 4 , wherein the process parameters for the machine to perform the real-world version of the manufacturing process comprise power data, speed data, and pattern data for the heat source in the additive manufacturing process.

7. The method of claim 4 , wherein the real world sensor is a photodiode sensor configured to sense light in the machine produced by the additive manufacturing process.

8. The method of claim 2 , wherein the process parameters are for a machine to perform the real-world version of the process based on the virtual data about the machine tool path.

9. The method of claim 1 , further comprising:

optimizing the process parameters using a trained version of the machine-learning software application.

10. The method of claim 9 , wherein optimizing the process parameters using the trained version of the machine-learning software model comprises:

adjusting the process parameters based on information from the machine-learning software model to produce a desired sensor reading at the real world sensor during the process.

11. The method of claim 1 , further comprising:

predicting a behavior of the real world sensor during the process with the machine-learning software model;

comparing the predicted behavior to an actual behavior of the real-world sensor during the process; and

detecting an anomaly in sensor behavior based on the comparison.

12. The method of claim 1 , wherein the process parameters for the real-world version of the process are parameters for entering into a machine configured to perform the process, the method further comprising:

entering the identified process parameters into a real-world machine configured to perform the process.

13. The method of claim 1 , wherein the only virtual data are the physics-based features that represent energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process.

14. The method of claim 1 , wherein the high fidelity physics information, comprises virtual data representing a temperature field, a melt pool dimension, material phase transformation data, mechanical distortion data, and residual stress and strain data associated with the process.

15. A system comprising:

a computer comprising:

a computer-based processor; and

a computer-based memory storing computer-executable instructions that, when executed by the computer-based processor, cause the computer-based processor to simulate a manufacturing process to produce virtual data about the process, wherein the virtual data about the process comprises physics-based features that represent energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process;

one or more machines configured to perform a real-world version of the process based on process parameters associated with the process;

a real-world sensor to sense a parameter associated with the real-world version of the process; and

a machine-learning software model trained to predict a behavior of the real-world sensor in a subsequent real-world manufacturing process based on the virtual data about the process, the process parameters, and sensor readings from the real-world sensor during the real-world-version of the process, wherein the virtual data about the process provided to train the machine-learning software model to predict the behavior of the real-world sensor comprises high fidelity physics information generated based on a computer-aided design (CAD) model of an object to be manufactured using the process.

16. The system of claim 15 , further comprising:

a computer-aided design (CAD) software application to produce a CAD model of a part associated with the process, and

wherein the computer-based processor simulates the process to produce the virtual data based at least in part on the CAD model.

17. The system of claim 16 , wherein the process is a manufacturing process and the virtual data about the process comprises virtual data about a machine tool path for the one or more machines.

18. The system of claim 17 , wherein the manufacturing process is an additive manufacturing process, and wherein the machine is a three-dimensional (3D) printer that comprises:

a build platform,

a powder bed atop the build platform; and

a heat source configured to deliver heat into various portions of the powder bed to create the part.

19. The system of claim 18 , wherein the virtual data about the machine tool path comprises power data, speed data, and pattern data for the heat source in the additive manufacturing process.

20. The system of claim 18 , wherein the process parameters for the machine to perform the real-world version of the manufacturing process comprise power data, speed data, and pattern data for the heat source in the additive manufacturing process.

21. The system of claim 18 , wherein the real world sensor is a photodiode sensor configured to sense light in the 3D printer produced by the additive manufacturing process.

22. The system of claim 21 , wherein the computer-based processor is further configured to optimize the process parameters for subsequent iterations of the process using a trained version of the machine-learning software application.

23. The system of claim 15 , wherein the computer-based processor is further configured to:

predict a behavior of the real world sensor during the process with the machine-learning software model;

compare the predicted behavior to an actual behavior of the real-world sensor during the process; and

detect an anomaly in sensor behavior based on the comparison.

24. A method comprising:

simulating, with computer-based software, an additive manufacturing process, to produce virtual data about the additive manufacturing process, wherein the virtual data about the additive manufacturing process comprises physics-based features that represent energy introduced into a part being manufactured per unit of time during the additive manufacturing process, and an ability to dissipate the energy introduced into the part during the additive manufacturing process;

entering process parameters into a real-world additive manufacturing machine to manufacture a real-world part using a real-world version of the additive manufacturing process;

providing a real-world sensor to sense a parameter associated with the real-world additive manufacturing machine performing the real-world version of the additive manufacturing process to produce the part;

receiving sensor readings from the real-world sensor while the real-world version of the process is being performed;

training a machine-learning software model to predict a behavior of the real-world sensor in a subsequent real-world manufacturing process based on the virtual data about the process, the process parameters, and the sensor readings to produce a trained machine-learning software model, wherein the machine comprises a build platform, a powder bed atop the build platform; and a heat source configured to deliver heat into various portions of the powder bed to create the part, wherein the virtual data about the process provided to train the machine-learning software model to predict the behavior of the real-world sensor comprises high fidelity physics information generated based on a computer-aided design (CAD) model of an object to be manufactured using the process;

subsequently manufacturing a real part using a process that includes additive manufacturing; and

predicting a behavior of a real-world version of the sensor in the subsequent manufacturing of the real part with the trained machine-learning software model.

25. A method comprising:

simulating, with computer-based software, a manufacturing process, to produce virtual data about the process, wherein the virtual data about the process comprises physics-based features that represent energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process;

identifying process parameters for a real-world version of the process, providing a real-world sensor to sense a parameter associated with the real-world version of the process;

receiving sensor readings from the real-world sensor while the real-world version of the process is being performed;

training a machine-learning software model to predict a behavior of the real-world sensor in a subsequent real-world manufacturing process based on the virtual data about the process, the process parameters, and the sensor readings to produce a trained machine-learning software model;

predicting a behavior of the real-world version of the sensor in the subsequent real-world manufacturing process with the trained machine-learning software model; and

periodically updating, in real time during the manufacturing process, values representing the energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process.

26. A system comprising:

a computer comprising:

a computer-based processor; and

a computer-based memory storing computer-executable instructions that, when executed by the computer-based processor, cause the computer-based processor to simulate a manufacturing process to produce virtual data about the process, wherein the virtual data about the process comprises physics-based features that represent energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process;

one or more machines configured to perform a real-world version of the process based on process parameters associated with the process;

a real-world sensor to sense a parameter associated with the real-world version of the process; and

a machine-learning software model trained to predict a behavior of the real-world sensor in a subsequent real-world manufacturing process based on the virtual data about the process, the process parameters, and sensor readings from the real-world sensor during the real-world-version of the process,

wherein the computer-based processor periodically updates, in real time during the manufacturing process, values representing the energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process.

27. A method comprising:

simulating, with computer-based software, an additive manufacturing process, to produce virtual data about the additive manufacturing process, wherein the virtual data about the additive manufacturing process comprises physics-based features that represent energy introduced into a part being manufactured per unit of time during the additive manufacturing process, and an ability to dissipate the energy introduced into the part during the additive manufacturing process;

entering process parameters into a real-world additive manufacturing machine to manufacture a real-world part using a real-world version of the additive manufacturing process;

providing a real-world sensor to sense a parameter associated with the real-world additive manufacturing machine performing the real-world version of the additive manufacturing process to produce the part;

receiving sensor readings from the real-world sensor while the real-world version of the process is being performed;

training a machine-learning software model to predict a behavior of the real-world sensor in a subsequent real-world manufacturing process based on the virtual data about the process, the process parameters, and the sensor readings to produce a trained machine-learning software model, wherein the machine comprises a build platform, a powder bed atop the build platform; and a heat source configured to deliver heat into various portions of the powder bed to create the part;

subsequently manufacturing a real part using a process that includes additive manufacturing; and

predicting a behavior of a real-world version of the sensor in the subsequent manufacturing of the real part with the trained machine-learning software model; and

periodically updating, in real time during the manufacturing process, values representing the energy introduced into a part being manufactured per unit of time during the manufacturing process, and an ability to dissipate the energy introduced into the part during the manufacturing process.