You are here

Diagnosis-Driven Prognosis for Decision Making

Award Information
Agency: National Aeronautics and Space Administration
Branch: N/A
Contract: NNX14CA25P
Agency Tracking Number: 144546
Amount: $115,419.00
Phase: Phase I
Program: SBIR
Solicitation Topic Code: A1.04
Solicitation Number: N/A
Timeline
Solicitation Year: 2014
Award Year: 2014
Award Start Date (Proposal Award Date): 2014-06-20
Award End Date (Contract End Date): 2014-12-19
Small Business Information
99 East River Drive
East Hartford, CT 06108-7301
United States
DUNS: 808837496
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Somnath Deb
 Principal Investigator
 (860) 761-9344
 sudipto@teamqsi.com
Business Contact
 Sudipto Ghoshal
Title: Business Official
Phone: (860) 761-9341
Email: sudipto@teamqsi.com
Research Institution
 Stub
Abstract

One cannot build a system-level Prognosis and Health Management (PHM) solution by cobbling together a bunch of existing prognostic techniques; it will have a very high rate of false-positive indications. On the other hand, if a system-level health management solution could identify the individual degradations and indictors associated with those degradations, and thereby decouple the problem into smaller pieces, the existing prognostic techniques could still be used to predict time to failure, and could therefore drive an effective Condition Based Maintenance and Decision Support System (CBM+).

Qualtech Systems, Inc. (QSI) and Vanderbilt University team seeks to develop a system-level diagnostic and prognostic process and a "sense and respond capability" which first uses error codes and discrete sensor values to correctly diagnose the system health including degradations and failures of sensors and components, and then invoke appropriate prognostic routines for assessment of remaining life and capability. Thus, QSI's Testability Engineering And Maintenance System (TEAMS) real-time reasoner will enable the use of many existing prognostics techniques in the broader context by decomposing the complex system into local datasets of degradations and associated sensor data sets, thereby limiting the problem-space for the prognostic techniques to their limited design scope. Indeed, it is well established in the contexts of parameter estimation and model-based fault identification (i.e., fault isolation and severity estimation) that feature selection and diagnosis, respectively, followed by parameter estimation provides major improvements in estimation performance (measured in terms of computational time as well as the standard deviations of the estimated parameters) when compared to full parameter estimation which provides biased estimates for all the parameters.

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government