User Tools

Site Tools


sensitivity_analysis_toolbox_example_-_external_model

Sensitivity analysis toolbox example - External model

Overview

The following example demonstrates how to perform a sensitivity analysis of an external model using the importing/export functionality of the sensitivity toolbox. The sensitivity analysis of an external model is performed in the following steps:

  1. Create parameters in Ecolego corresponding to the model inputs in the external model
  2. Assign probability density functions to the parameters in Ecolego
  3. Generate probabilistic values for the parameters (Sensitivity toolbox→input samples→generate)
  4. Export the samples of the parameters (Sensitivity toolbox→Input samples→export)
  5. Simulate samples of model outputs externally and put the samples in a tab delimited textfile
  6. Import the model outputs (Sensitivity toolbox→outputs→import)
  7. Inspect the sensitivity analysis indices (Sensitivity toolbox→results→create correlation table)

The model

The model in this example is the three parameter ishigami model

y = sin(x1)+7.0*sin(x2)^2.0+0.1*x3^4.0*sin(x1)


The model will not be created in Ecolego, but evaluated in excel as a external model.

Generation of parameter samples in Ecolego

Create the three parameters x1,x2,x3 in Ecolego and assign them the distribution unif(min=-3.14,max=3.14).

Opening the Sensitivity analysis perspective

Open the Sensitivity Analysis toolbox perspective by clicking SensitivityToolboxPerspective24 in the Ecolego toolbar

Selection of probabilistic parameters

SensitivityToolboxParameterSelectionTabGo to Settings→Parameter selection to open the Parameter selection view. Here, the parameters for which the sensitivity analysis should be run are selected.\\
\\
Select the three parameters x1,x2,x3 by moving them to the right in the parameter selection view


Selection of Sensitivity analysis method

SensitivityToolboxMethodTabGo to Settings→Method to open the Sensitivity analysis method view. Here, the sensitivity analysis method and its settings is selected. Note that only the probabilistic method can be used with external models.\\
\\
Select the method Probabilistic with the following settings:\\
\\
Base sample size: 1000\\
Sampling type: Latin Hypercube\\
Calculate second order indices: Unchecked\\
Seed: Auto\\


Generating and exporting parameter samples

Go to Analysis→Inputs samples→Generate. This will generate samples for the parameters.

Go to Analysis→Inputs samples. Select Analysis→Inputs samples→Export. In the dialog that appears, select all variables and press OK. Then select which file to export the samples to (for example “IshigamiInputs.txt”).

Evaluating the external model

Open the file with the exported samples in Excel. Evalute the following expression in a separate column of length 1000 using the samples for x1,x2,x3:

y = sin(x1)+7.0*sin(x2)^2.0+0.1*x3^4.0*sin(x1)


Export the resulting column from excel as a new text file named IshigamiOutputs.txt.

Sensitivity analysis

To perform the sensitivity analysis, both the input samples and outputs samples that were used for the model evaluation are first imported into ecolego.

Import the input samples by selecting Analysis→Inputs→Import and selecting the file containing the samples generated for the model inputs (IshigamiInputs.txt).

Import the output samples by selecting Analysis→Outputs→Import and selecting the file containing the samples generated for the model outputs (IshigamiOutputs.txt).

Go to Analysis→Result. Select the output y to open a correlation table to inspect the sensitivity indices.

sensitivity_analysis_toolbox_example_-_external_model.txt · Last modified: 2023/03/06 13:37 by mina