Christopher Teubert

icon indicating a website link https://www.nasa.gov/content/diagnostics-prognostics icon indicating an email [email protected]

icon company @nasa icon location NASA Ames Research Center icon location Lead, Diagnostics and Prognostics Group

Results 24 comments of Christopher Teubert

Discrepancies in pump model

[NASA2012__a-comparison-of-filter-based-approaches-for-model-based-prognostics.pdf](https://github.com/nasa/prog_models/files/6971963/NASA2012__a-comparison-of-filter-based-approaches-for-model-based-prognostics.pdf) [Uploading NASA2014__distributed-prognostics-based-on-structural-model-decomposition.pdf…]()

Explore data-driven models

Started exploring this- working to implement LSTM State Transition Model

Check if each state exists when setting state_limits

could also be good to make sure state_limits has the function items() (i.e., is a dict or dict-like class)

Hyperparameter optimization

Consider using kerastuner

Research: investigate additional surrogate methods

See also: https://pymor.org

Battery Characterization Example

Make sure we give Ed credit with the release - he has contributed to the approach

Revamp plotting/figures

e.g., https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.plot.line.html

Update plots to be able to use an existing figure

Came out of discussions with @kjjarvis

Multiple Output layers

Suggested by @matteocorbetta

Release/v1.4

Completed release review. Participants: @matteocorbetta, @chetankul, @kjjarvis Notes: * Data-driven - bullet 1, interchangeable - something more specific like “interchangeable in sim, …” or “in use” * In API Reference-...