Transcription
Tutorial: Scalable Models for Efficient Design in EMC and SI ApplicationsScalable Compact Models for Efficient Design in EMC and SI ApplicationsF. Ferranti*, T. Dhaene*, L. Knockaert*, G. Antonini**, A. Ciccomancini Scogna†* Department of Information Technology (INTEC), Ghent University – IBBT** Dipartimento di Ingegneria Elettrica e dell'Informazione, Università degli Studi dell'Aquila†CST of AmericaDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
OutlineIntroductionScalable MacromodelsNumerical examples EMC example SI exampleConclusionsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
OutlineIntroductionScalable MacromodelsNumerical examples EMC example SI exampleConclusionsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
telecomelectronicsfluid dynamicsinputoutputout nt of Information Technology – Internet Based Communication Networks and Services (IBCN)
AutomotiveChemistryElectronicsSimulation ModelDesignvariableswidth, temperature,angle, frequency, .pAerodynamicsFluent , HSPICE , CST ,Comsol , Abaqus , .Costly2MetallurgyResponsevariableslift, S-parameters,pressure, stress, .Adaptive sevariablesDesignvariablesDistributed ComputingCheapScalable/Parametric ModelNeural network, Kriging, SVM, rational function, CAD/CAM/CAEEnvironmentDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
AutomotiveChemistryElectronicsSimulation ModelDesignvariableswidth, temperature,angle, frequency, .pAerodynamicsFluent , HSPICE , CST ,Comsol , Abaqus , .Costly2MetallurgyResponsevariableslift, S-parameters,pressure, stress, .Adaptive sevariablesDesignvariablesDistributed ComputingCheapScalable/Parametric ModelNeural network, Kriging, SVM, rational function, CAD/CAM/CAEEnvironmentDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Design process several decisions materials geometrical dimensions shape constraints space cost performanceDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Simulators implementation of models describe systems behavior help designersMeasurements post tuning verification help designersDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
A typical design process requires design space optimizationdesign space explorationsensitivity analysis multiple simulations (measurements) different design parameters values (e.g. layout features)Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
A typical design process requires Multiple simulations (measurements) computationally expensive (time and memory) Can we do better? Yes By scalable macromodelsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
OutlineIntroductionScalable MacromodelsNumerical examples EMC example SI exampleConclusionsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
real worldmodelingdiscretisation bysimulatorsmodelPDEU(s,g)sX AX BUY CX DUY(s,g)Model-driven PMORU(s,g)H(s,g)Y(s,g)scalable macromodelDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
real worldmodelingdiscretisation bysimulatorsmodelPDEU(s,g)sX AX BUY CX DUdataY(s,g)simulationsData-driven PMORModel-driven PMORU(s,g)H(s,g)Y(s,g)scalable macromodelDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
real worldmodelingmeasurementsdiscretisation bysimulatorsmodelPDEdataU(s,g)sX AX BUY CX DUY(s,g)Data-driven PMORModel-driven PMORU(s,g)H(s,g)Y(s,g)scalable macromodelDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
PMOR conceptsTwo design space grids are used in the modeling process estimation grid validation gridDesign spaceg ( g ( n ) ) nN 1Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
real worldmodelingmeasurementsdiscretisation bysimulatorsmodelPDEU(s,g)sX AX BUY CX DUdataY(s,g)simulationsData-driven PMORModel-driven PMORU(s,g)H(s,g)Y(s,g)scalable macromodelDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
VECTFITMULTIPORTRationalFittingMethodsSYSTEM MATRICESInterpolationSTABILITYLeast SquaresTRANSFER FUNCTIONSPASSIVITYPOLES/RESIDUESNOISY DATAdata-driven PMORaccuracyH(s,g)efficiencystability guaranteedpassivity guaranteedscattered dataDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
VECTFITMULTIPORTRationalFittingMethodsSYSTEM MATRICESInterpolationSTABILITYLeast SquaresTRANSFER FUNCTIONSPASSIVITYPOLES/RESIDUESNOISY DATAdata-driven PMORPQ( g )p 1 s a p ( g )H( s, g ) H(s,g)H( s, g ) C(g )( sI A (g )) 1 B(g ) D(g )scattered dataDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
nSTABILITYLeast SquaresTRANSFER FUNCTIONSPASSIVITYNOISY DATAdata-driven PMORPQ( g )p 1 s a p ( g )H( s, g ) H(s,g)H( s, g ) C(g )( sI A (g )) 1 B(g ) D(g )scattered dataDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Design spaceg ( g ( n ) ) nN 1Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Features each design space cell has its own model local approach independent from a specific state-space realization stability and passivity guaranteed over the design space suitable to robust adaptive sampling different flavoursDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
OutlineIntroductionScalable MacromodelsNumerical examples EMC example SI exampleConclusionsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
VECTFITMULTIPORTSYSTEM MATRICESRationalFittingMethodsInterpolationLeast Squaresdata-driven PMORPQ(g )p 1 s a p (g )H ( s, g ) H(s,g)H( s, g ) C(g )( sI A (g )) 1 B(g ) D(g )scattered dataDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
3D example: EnclosureL12.1cmW30cm30.1cmDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Plane wave excitationDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Speed-up 33750 xDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
W 1.8 cmL 15 cmDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
W 1.8 cmL 15 cmDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
OutlineIntroductionScalable MacromodelsNumerical examples EMC example SI exampleConclusionsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
nSTABILITYLeast SquaresTRANSFER FUNCTIONSPASSIVITYdata-driven PMORPQ(g )p 1 s a p (g )H ( s, g ) H(s,g)H( s, g ) C(g )( sI A (g )) 1 B(g ) D(g )scattered dataDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
3D example: PCBR antipads radius7.7mm6.6mmD distance betweensignalviasandground vias (centerto center)Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
3D example: PCBDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
Speed-up 4650 xDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
D 1.8525 mmR 0.543 mmDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
D 1.8525 mmR 0.543 mmDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
D 1.8525 mmR 0.543 mmDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
OutlineIntroductionScalable MacromodelsNumerical examples EMC example SI exampleConclusionsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
telecomelectronicsfluid dynamicsinputoutputout nt of Information Technology – Internet Based Communication Networks and Services (IBCN)
AutomotiveChemistryElectronicsSimulation ModelDesignvariableswidth, temperature,angle, frequency, .pAerodynamicsFluent , HSPICE , CST ,Comsol , Abaqus , .Costly2MetallurgyResponsevariableslift, S-parameters,pressure, stress, .Adaptive sevariablesDesignvariablesDistributed ComputingCheapScalable/Parametric ModelNeural network, Kriging, SVM, rational function, CAD/CAM/CAEEnvironmentDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
AutomotiveChemistryElectronicsSimulation ModelDesignvariableswidth, temperature,angle, frequency, .pAerodynamicsFluent , HSPICE , CST ,Comsol , Abaqus , .Costly2MetallurgyResponsevariableslift, S-parameters,pressure, stress, .Adaptive sevariablesDesignvariablesDistributed ComputingCheapScalable/Parametric ModelNeural network, Kriging, SVM, rational function, CAD/CAM/CAEEnvironmentDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Scalable macromodelsU(s,g)Multiple design variablesH(s,g)Y(s,g)scalable macromodelCompact modelsEfficient design activities (excellent speed-ups) Multiple simulations (measurements) Design space optimization, exploration, sensitivity analysisDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Scalable macromodelsU(s,g)Time-domain simulations Non-linear drivers and receiversH(s,g)scalable macromodelStochastic modeling impact of manufacturing tolerancesModels from measurements Y(s,g)noise to handleApplications in different domainsDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Department of Information Technology – Internet Based Communication Networks and Services (IBCN)
QuestionsContact info: francesco.ferranti@ugent.beDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Recent publicationsF. Ferranti, L. Knockaert, T. Dhaene, "Passivity-Preserving Parametric Macromodeling byMeans of Scaled and Shifted State-Space Systems", IEEE Trans. on Microwave Theory andTechniques, vol. 59, no. 10, pp.2394-2403, October 2011.F. Ferranti, T. Dhaene, L. Knockaert, G. Antonini and A. Ciccomancini Scogna, "ScalableCompact Models for Fast Design Optimization of Complex Electromagnetic Systems",International Journal of RF and Microwave Computer-Aided Engineering, vol. 22, no. 1, pp.20-29, January 2012.F. Ferranti, M. Nakhla, G. Antonini, T. Dhaene, L. Knockaert, A. E. Ruehli, “Interpolationbased Parameterized Model Order Reduction of Delayed Systems”, IEEE Trans. on MicrowaveTheory and Techniques, vol. 60, no. 3, pp. 431-440, March 2012.Contact info: francesco.ferranti@ugent.beDepartment of Information Technology – Internet Based Communication Networks and Services (IBCN)
Scalable Compact Models for Efficient Design in EMC and SI Applications Tutorial: Scalable Models for Efficient Design in EMC and SI Applications F. Ferranti *, T. Dhaene*, L. Knockaert*, G. Antonini**, . Data-driven PMOR data simulations Model-driven PMOR sX AX BU Y CX DU
