Netfabb sim-presentation

2017 Temperature and Distortion Simulation for Laser Powder Bed Fusion PartsDivy Kishor TiwaryTechnical Specialist, Additive Solutions▸Originally CUBES, by Pan Computing LLC, Founded in 2012▸Physics based, multi-scale, process modeling software for thermomechanical processes(additive manufacturing, welding, thermalforming)▸Extensive model validation (Industrial Partners, America Makes, Penn State,Government)▸Acquired by Autodesk in March 2016▸Commercialized as Netfabb Simulation and integrated into Netfabb Thermo-Mechanical Simulation of Additive ManufacturingSimulation Utility for NetfabbBuild failure is a common problem in AMRecoater BladeInterferenceThermomechanical modeling can be used to avoid costly trial and error iterationsProcess ParametersFinal ProductHeat TransferTemperature HistoryDeformationMechanicalResponse Manual Optimization Iterations(will be automatic in the future)Inputs:-Machine process parameters-Material properties Outputs:-Deformation -Stress-TemperatureA Sequentially coupled thermomechanical model is used to calculate temperature, stress, & distortionHeat transfer:Mechanical response:Process ParametersHeat Transfer Temperature HistoryMechanical Response Stress & DeformationPart: 14 ft x 1.5 ft x 0.75 ftRun time: 1 day, 16 core machineThe Netfabb Sim solver uses custom Hex-8 ElementsProject PanAbaqusd dxkdT dx+Q =ρC pdT dt(1)T x,t =T 0+b T xt (2)k =k 0+b k T (3)C p =C 0+b C T(4)Q =ρC p b T x −b k b T 2t 2(5)q s =−k dTdxn=−kb T tn(6)Several key technologies were invented to facilitate the simulation of moving heat source problemsNovel fast and accurate Hex8 elements Layer by layer octree H-adaptivityCustom Finite Elements allow for run times over 1000x faster than general purpose solversLine where results are comparedAbaqus Netfabb ConvergedAbaqus Netfabb ConvergedAbaqus Netfabb ConvergedCoarser MeshRun-time for converged solution:•Abaqus: 208s•Netfabb Simulation: 0.23sThe modeling approach was successfully applied to simulate DED processesMoving source:▸Abaqus/ANSYS (4 elements / radius):▸9 billion nodes▸ 2 000 000 000 terabytes* of data for displacement results ▸Netfabb Simulation (1 element / radius):▸140 million nodes▸500 000 terabytes of dataSimulating Powder Bed Fusion parts adds computational challenge4 inches*10% of the whole internet, circa 2007Multi-scale modeling allows for fast simulation of entire partsDetailed fine-scale process parameter model (PRM) Input: Process parameters (Power, scan speed, layer thickness, etc.)Part-scale geometric model*Input: Geometry (STL)*4 hour runtime on a 14 core CPU PRM fileNetfabb Simulation is capable of analyzing full buildvolumesPart:275 mm x 204 mm x 44.1 mmRun time:11 hours20 core machineIn situ measurements give additional insight into the in-process physicsDVRTThermocoupleDAQThe measurements show the interaction of the individual deposition layersD i s p l a c e m e n t (m m )The model shows close agreement with experimental resultsPart-level simulation gives close agreement with post-process distortion measurementsExperimental data courtesy of America Makes, GE GRC, and United Technologies Research Center Part: 43 mm x 23 mm x 54 mmRun time: 30 min on 4 core machineThe simulations also predict common failure modesDIRECT MODELING OF POWDER LOSSES IMPROVES THE ACCURACY OF DISTORTION PREDICTIONSRESIDUAL STRESSES AND DISTORTION AFTER WIRE-CUTTING CAN BE ACCURATELY CALCULATEDWithout part scale plasticity With part scale plasticityExperimental data courtesy of America Makes, GE GRC, and United Technologies Research CenterThe distortion predictions can be used to output acompensated pre-formUnits in mmNominal Geometry Print Scan Nominal GeometrySimulation results Compensated GeometryExperimental data courtesy of America Makes, GE GRC, and United Technologies Research CenterCompensated Geometry Print ScanCompensation dramatically reduces distortion levels G4: Canonical 2With compensationWithout compensationLack of Fusion and Hot SpotsThermal images illustrate the temperature evolution during the buildMoving adaptivity can dramatically speed up simulations (thermal only)Run time: 3797 s Run time: 462 sPeak temperature can be tracked over all time steps on the refined meshThe simulation tracks the fraction of elements above or below threshold peak temperatures15.7 % of elements below 1250 o C:5.6 % of elements above 2600 o C:Multi-scale modeling can be used to predict defects on the entire buildMultiscale analysis to predict defects on the entire buildFraction below 1250 O C Fraction above 2600 O CCommercially Available as of v2018.1Conclusions1.The moving source model shows good agreement withexperimental in situ temperature and distortionmeasurements2.Results from small moving source analyses can bemapped onto larger geometries, allowing for fast andaccurate full part-scale simulations3.The simulated distortion predictions can be used topredict regions of likely recoater blade interference and support structure failure4.Direct modeling thermal losses to powder improves theaccuracy of the thermo-mechanical model5.Modeled distortion can be used to produce acompensated geometry, mitigating distortion prior to the first test buildAutodesk, Netfabb, and the Autodesk logo are registered trademarks or trademarks of Autodesk, Inc., and/or its subsidiaries and/or affiliates in the USA and/or other countries. All other brand names, product names, or trademarks belong to their respective holders. Autodesk reserves the right to alter product and services offerings, and specifications and pricing at any time without notice, and is not responsible for typographical or graphical errors that may appear in this document.© 2017 Autodesk. All rights reserved.。