WOLFRAM

Wolfram Language & System Documentation Center

SystemModelSurrogate[…]

represents the symbolic surrogate model obtained from SystemModelSurrogateTrain.

Details
Details and Options Details and Options
Examples  
Basic Examples  
Scope  
Properties  
Applications  
See Also
Tech Notes
Related Guides
Related Links
History
Cite this Page

SystemModelSurrogate

SystemModelSurrogate[…]

represents the symbolic surrogate model obtained from SystemModelSurrogateTrain.

Details

  • SystemModelSurrogate is generated by the function SystemModelSurrogateTrain and provides ways of exploring, diagnosing and extracting training results.
  • A SystemModelSurrogate[…] object sms can be used to retrieve properties through sms["property"].
  • sms["Properties"] gives a list of available properties. Additional details about the properties are listed on the page for SystemModelSurrogateTrain.
  • Typical properties include:
  • "Net"surrogate model as a neural net
    "ParametricFunction"surrogate model as a parametric function
    "SystemModel"surrogate model as a system model
  • Use sms[{p1,…}] to get the values of properties pi.

Examples

open all close all

Basic Examples  (1)

Compute a surrogate model:

Wolfram Language code: surrogate = SystemModelSurrogateTrain[AffineStateSpaceModel[{{Cos[x2], (-a)*x2 + Cos[x1^2]}, {{0}, {1}}, {x2}, {{0}}}, {x1, x2}], <|"Outputs" -> x2, "ParameterValues" -> {a -> CenteredInterval[0.2, 0.1]}|>]

Extract the parametric function:

Wolfram Language code: surrogate["ParametricFunction"]

Retrieve a list of properties:

Wolfram Language code: surrogate[{"Net", "RRMSEPlot"}]

Scope  (1)

Properties  (1)

Find the list of available properties:

Wolfram Language code: model = NonlinearStateSpaceModel[{{u - x1 x2, u x2 + a Cos[x1]}, {x1}}, {x1, x2}, u]
Wolfram Language code: spec = <|"Outputs" -> x1, "ParameterValues" -> {a -> Around[0.5, 0.1]}, "Inputs" -> {u -> Cos}|>;
Wolfram Language code: surrogate = SystemModelSurrogateTrain[model, spec]
Wolfram Language code: surrogate["Properties"]

Extract the parametric function:

Wolfram Language code: fun = surrogate["ParametricFunction"]

Extract the neural net:

Wolfram Language code: surrogate["Net"]

Create a SystemModel from the surrogate with a given model name:

Wolfram Language code: sm = surrogate["SystemModel", "SurrogateModelName" -> "MySurrogate", GeneratedAssetLocation -> $TemporaryDirectory]

Extract validation information like the maximum absolute error plot:

Wolfram Language code: surrogate["MAEPlot"]

Extract the root mean squared error plot:

Wolfram Language code: surrogate["RMSEPlot"]

Extract the relative root mean squared error plot:

Wolfram Language code: surrogate["RRMSEPlot"]

Compare the surrogate with the original model for a specific value of the parameter:

Wolfram Language code: Show[SystemModelPlot[model, x1, <|"ParameterValues" -> {a -> 0.5}, "Inputs" -> {u -> Cos}|>], ListLinePlot[fun[0.5], PlotStyle -> Directive[StandardRed, Dashed], PlotRange -> All], PlotRange -> All]

Applications  (1)

Start with a model for battery discharging:

Wolfram Language code: model = \!\(\*GraphicsBox[«8»]\);

Estimate the voltage in a battery for a specific range of the open circuit voltage at maximum state of charge:

Wolfram Language code: simulation = <|"Outputs" -> "battery1.v", "SimulationInterval" -> 300, "SimulationCount" -> 50|>;
Wolfram Language code: uncertainty = <|"ParameterValues" -> {"cellData1.OCVmax" -> Interval[{5, 10}]}|>;
Wolfram Language code: surrogate = SystemModelSurrogateTrain[model, Join[uncertainty, simulation]]

Compare the surrogate with the original model for a specific value of the open circuit voltage:

Wolfram Language code: Show[SystemModelPlot[model, "battery1.v", 300, <|"ParameterValues" -> {"cellData1.OCVmax" -> 7.5}|>], SystemModelPlot[surrogate, "battery1.v", 300, <|"ParameterValues" -> {"cellData1.OCVmax" -> 7.5}|>, PlotStyle -> StandardRed], PlotRange -> All]

Extract the neural net:

Wolfram Language code: net = surrogate["Net"]

Deploy a Manipulate that plots the net for various values of the parameter:

Wolfram Language code: CloudDeploy[Manipulate[ListLinePlot[net[{v}, WorkingPrecision -> "Real64"], PlotRange -> All], {{v, 7.5}, 5, 10}], Permissions -> "Public"]
Wolfram Research (2026), SystemModelSurrogate, Wolfram Language function, https://reference.wolfram.com/language/ref/SystemModelSurrogate.html.

Text

Wolfram Research (2026), SystemModelSurrogate, Wolfram Language function, https://reference.wolfram.com/language/ref/SystemModelSurrogate.html.

CMS

Wolfram Language. 2026. "SystemModelSurrogate." Wolfram Language & System Documentation Center. Wolfram Research. https://reference.wolfram.com/language/ref/SystemModelSurrogate.html.

APA

Wolfram Language. (2026). SystemModelSurrogate. Wolfram Language & System Documentation Center. Retrieved from https://reference.wolfram.com/language/ref/SystemModelSurrogate.html

BibTeX

@misc{reference.wolfram_2026_systemmodelsurrogate, author="Wolfram Research", title="{SystemModelSurrogate}", year="2026", howpublished="\url{https://reference.wolfram.com/language/ref/SystemModelSurrogate.html}", note=[Accessed: 12-June-2026]}

BibLaTeX

@online{reference.wolfram_2026_systemmodelsurrogate, organization={Wolfram Research}, title={SystemModelSurrogate}, year={2026}, url={https://reference.wolfram.com/language/ref/SystemModelSurrogate.html}, note=[Accessed: 12-June-2026]}