Commercial scale casting process including optimization via multi-fidelity optimization

1. A process comprising:

a first casting modelling stage producing resulting casting parameters;

a second casting modelling stage performed using the resulting casting parameters of the first casting modelling stage and of higher fidelity than the first casting modelling stage, wherein:

relative to the second casting modeling stage, the first casting modelling stage involves: a simpler furnace model comprising a simpler geometric model of the furnace; simplification or omission of part features;

coarser mesh; and

simpler geometric approximation of gating features; and

the first casting modelling stage and the second casting modelling stage are performed using one or a connected group of computer systems;

in parallel with the second casting modelling stage, a physical casting trial using the resulting casting parameters of the first casting modelling stage; and

evaluating the physical casting trial that used the resulting casting parameters of the first casting modelling stage;

in response to the evaluating yielding an unsatisfactory result:

the one or a connected group of computer systems expanding a design of experiment (DOE); and

forking back to within the first casting modelling stage to iterate the first casting modeling stage, the second casting modelling stage, the physical casting trial, and the evaluating using the expanded design of experiment; and

in response to the evaluating yielding a satisfactory result, result after the expanding the design of experiment and after the forking back, performing commercial scale production of castings using the casting parameters of the first casting modelling stage.

2. The process of claim 1 wherein:

the first casting modelling stage includes a simulation having a lower fidelity than a higher fidelity simulation included in the second casting modelling stage.

3. The process of claim 2 wherein:

the first casting modelling stage includes multiple runs of the lower fidelity simulation;

and the second casting modelling stage includes only a single run of the higher fidelity simulation.

4. The process of claim 2 wherein:

the first casting modelling stage includes an evaluation after the lower fidelity simulation with a negative result of the evaluation forking back to before the lower fidelity simulation; and

the second casting modelling stage includes only a single run of the higher fidelity simulation.

5. The process of claim 1 wherein:

the first casting modelling stage includes an evaluation after the lower fidelity simulation with a negative result of the evaluation forking back to before the lower fidelity simulation; and

the second casting modelling stage includes only a single run of the higher fidelity simulation.

6. The process of claim 5 wherein:

the forking back is to a point dependent on the physical casting trial, second casting modelling stage, and first casting modelling stage; and

the forking back leads to repetition of the second casting modelling stage and the physical casting trial.

7. The process of claim 5 wherein:

the forking back causes the first casting modelling stage to run with an expanded process parameter range relative to the prior run; and

the expanded process parameter range is created algorithmically by a rules-based modification of a prior run process parameter range.

8. The process of claim 7 wherein:

the rules-based modification of a prior run process parameter range expands the process parameter range based on a magnitude of the unsatisfactory result and one or more associated process parameter sensitivities.

9. The process of claim 5 wherein:

the forking back is to a point dependent on the nature of the unsatisfactory result.

10. The process of claim 9 wherein:

the first casting modelling stage includes running a solidification model;

in response to a first said unsatisfactory result, the forking back is to a point after the solidification model; and

in response to a second said unsatisfactory result, the forking back is to a point before the solidification model.

11. The process of claim 5 wherein:

the first casting modelling stage includes running a solidification model;

the first casting modelling stage includes a determination of variables and regions of interest; and

the forking back is to a point after the solidification model but at or before the determination of variables and regions of interest.

12. The process of claim 5 wherein:

the first casting modelling stage includes running a solidification model;

the first casting modelling stage includes an optimization and sensitivity analysis of defect risk; and

the forking back is to a point after the solidification model but at or before the optimization and sensitivity analysis of defect risk.

13. The process of claim 1 wherein:

the first casting modelling stage includes running a solidification model;

the first casting modelling stage includes an optimization and sensitivity analysis of defect risk; and

the forking back is to a point after the solidification model but at or before the optimization and sensitivity analysis of defect risk.

14. The process of claim 1 wherein, relative to the second casting modeling stage, the first casting modelling stage involves one or more of:

omission of a casting

core; lower part count

per mold; smaller run

time; and fewer

iterations.

15. The process of claim 1 wherein, relative to the second casting modeling stage, the first casting modelling stage involves simplistic approximation of gating features.

16. The process of claim 1 wherein:

the evaluating comprises comparing results of the second casting modelling stage with results of the physical casting trial; and

the comparing comprises determining presence in the physical casting trial of defects not predicted by the second casting modelling stage.

17. The process of claim 1 wherein:

the evaluating includes the one or a connected group of computer systems:

determining need for engineering review by:

comparing the expanded DOE to original DOE bounds;

calculating a distance metric between the expanded DOE and the original DOE bounds; and

comparing the distance metric to predetermined limits.

18. The process of claim 1 wherein:

the one or a connected group of computer systems comprises one or more servers and one or more workstations; and

the one or more servers and one or more workstations have respective scripting language interpreters running respective scripts configured to:

compute simulation parameters corresponding to experimental points spanning a process parameter range;

execute simulations according to the computed simulation parameters;

analyze the relative risk of defects predicted by each executed simulation; and

perform an optimization and sensitivity analysis of defect risk in order to determine optimized casting process parameters.

19. The process of claim 1 wherein:

for each iteration of the physical casting trial, a 3D printer prints a wax pattern using the casting parameters;

for each iteration of the physical casting trial, a shell line robot shells the wax pattern using the casting parameters to form a mold; and

for each iteration of the physical casting trial, a casting furnace casts using the casting parameters.