AI (Artificial Intelligence) Gas Turbine Rotor Diagnostics
Abstract
With increased emphasis on improving the reliability and maintainability of gas turbine engines in the U.S. Air Force inventory and the need to operate aircraft from austere forward bases, new diagnostic tools are required. Such tools must provide reliable, consistent diagnoses and minimize training requirements for maintenance personnel. Knowledge-based diagnostic systems have the potential to meet these needs by improving the productivity of USAF maintenance personnel, affecting a standardized diagnostic approach at all maintenance facilities, and providing a wider dissemination of the benefits of accumulated USAF experience in gas turbine vibration diagnostics. Typical rotordynamic faults observed in gas turbine engines were surveyed to assess the extent of knowledge to diagnose rotordynamic faults. A set of generic rotordynamic faults was selected and a diagnostic strategy was developed for those faults. System configuration for a typical knowledge-based rotordynamic diagnostic system was defined based on the requirements identified during the survey. The diagnostic concepts developed were demonstrated using a laboratory test rig capable of having faults implanted in it. Five faults were selected: unbalance, misalignment, rub, increased support flexibility and accessory vibrations. The diagnostic logic for the five faults was implemented as a knowledge-based system (KBS) using the HARVEST fault tree analyzer. The KBS was interfaced to a data acquisition system to acquire vibration data for diagnosis. The integrated system was used successfully to diagnose all faults implanted in the test rig.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1986
- Accession Number
- ADA178996
Entities
People
- B. Aggarwal
- J. Tecza
- Jessica L. Giordano
- R. Brunner