NASA SBIR 2021-I Solicitation

Proposal Summary

Proposal Number:          21-1- A1.06-2874
Subtopic Title:
      Vertical Lift Technology and Urban Air Mobility
Proposal Title:
      Onboard Prognostics and Health Management for UAM using Machine Learning Techniques

Small Business Concern

Empirical Systems Aerospace, Inc.
3580 Sueldo Street, San Luis Obispo, CA 93401
(805) 275-1053                                                                                                                                                                                

Principal Investigator:

Clayton Green
3580 Sueldo Street, CA 93401 - 7338
(805) 275-1053                                                                                                                                                                                

Business Official:

Andrew Gibson
3580 Sueldo St., CA 93401 - 7338
(805) 704-1865                                                                                                                                                                                

Summary Details:

Estimated Technology Readiness Level (TRL) :                                                                                                                                                          
Begin: 2
End: 4
Technical Abstract (Limit 2000 characters, approximately 200 words):

During this Phase I effort ESAero will research and develop a prognostics and health management system (PHM) designed for electric motor inverters. The final PHM system will consist of a microcomputer or FPGA loaded with a fault detection algorithm connected to the sensors in an inverter. This stage will focus on refining the algorithm. Past ESAero PHM used deep autoencoder, but convolutional autoencoder, a recurrent neural network (RNN), long short-term memory (LSTM) units, and gated recurrent units (GRUs) will be investigated as alternatives and improvements. ESAero will add additional software functions of fault classification and remaining useful life. K-Nearest Neighbor (KNN), support vector machines (SVMs), and random decision forests are candidates for fault classification methods. For remaining useful life, ESAero will explore several physic model based and data-driven approaches. This effort will utilize ESAero’s large depository of X57 test data taking during prototype and acceptance testing. PHM will “operate” on the data and detect and predict faults recorded in the X57 tests. This will demonstrate PHM’s capability, performance, requirements, and reduce risks before entering hardware design. Planning ahead, ESAero will investigate the certification and regulations that will be applicable to PHM. ESAero will develop a risk mitigation plan to overcome regulatory barriers. The product of this research will lead to the development of PHM requirements for UAM inverters. In addition, ESAero will begin prototype component selection which will verify currently available hardware can meet these requirements.

Potential NASA Applications (Limit 1500 characters, approximately 150 words):

This effort will add understanding of health and remaining useful life (RUL) of inverters for electric UAM. The results of this effort will benefit NASA’s Advanced Air Mobility (AAM) National Campaign, NASA X57 Maxwell, and other NASA electric efforts. Determining reliability, RUL, and how to maintenance electrical components of electric systems has burdened regulators. Electric systems have no visual detection of ware.  A system that can manage the health and predict RUL will provide actionable data to technicians and regulators. 

Potential Non-NASA Applications (Limit 1500 characters, approximately 150 words):

PHM for inverters will provide health and remaining useful life (RUL) to electric components that previously would only be hours operated or operation to failure. PHM planned in this effort could be a small integrable board or a software addon to an inverter with enough computational power. This inverter PHM could later be incorporated in a aircraft level PHM health manage system.

Duration:     6

Form Generated on 04/06/2021 12:05:23