Test Automation to conduct Design of Experiments (DOE)

Link test scripting and data-based modeling to improve accuracy and efficiency

ETAS continues to introduce product features in their Measurement, Calibration and Diagnostic (MCD) software tools to allow users to leverage the strengths of the individual tools.  ETAS INCA-FLOW (version 4.7) has new test bench methods which link test scripting to data-based modeling and optimization via ETAS ASCMO (Advanced Simulation for Calibration Modeling and Optimization).

This enables calibration engineers to improve accuracy and efficiency during the development of optimized powertrain systems:

  • Reduces effort to create multiple variable design-of-experiments (DOE) tests.
  • Error-proofs testing by removing manual input DOE calibration settings.|
  • Reduces the number of tests required to achieve accurate modeling by reporting and monitoring model accuracy as testing proceeds in real time.

ETAS ASCMO data-based models can be applied in several ways:

  • Optimization of calibrations with data-based models using ETAS ASCMO modeling and optimization.
  • Creation of a virtual sensor in the ETAS INCA calibration and measurement platform using an ETAS ASCMO data-based model.
  • Implementation of a plant model in INCA-FLOW using the MiL-sim method and an ETAS ASCMO model.

How it used to be done – the old process

Engineers previously used several software applications to design and analyze test data.   The test would be created in one application, the test settings would be transferred to another application, and finally, modeling and analysis of the test results would be performed in yet another application.  After completion of the test, the engineer struggled with data outliers either due to incorrect manual calibration entries at the test cell or due to non-robust hardware.  Rather than understanding the variance, data was removed and the data was made to fit the model   Retesting the suspect test points was not possible due to limited test resources.  In most cases, the test hardware had been removed from the test dynamometer prior to completion of the data review.

How it should be done – the new process

Static Test Planning with ETAS ASCMO guides the engineer in designing the DOE by querying the expected system complexity (high, medium, low) and returning a proposed quantity of tests.  The engineer selects the proposal or chooses a desired number of tests.  The space-filling designed tests are arranged into groups or blocks.  For example, a block may be a group of tests that can be completed in a scheduled work shift.  Additional test points are added.  The repetitive points are used to check for drift in exhaust gas emission measurements or to detect robustness of the hardware.  The engineer adds the INCA-FLOW method ‘ASCMO Load ExpeDes’ which loads the DOE test plan.  Next, the engineer adds the ASCMO Train Model method which will call the ASCMO Static Modeling and Optimization routine.  The engineer has now created an INCA-FLOW test script that will execute the DOE test plan and model the data as the test is in progress.

Video Illustration:

 

We’d love to show you!

If you have any questions, just comment below. If you are interested in a demonstration or an evaluation, please contact us.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.