ASRI proposes to develop a programmable Dialectic Neural Network Engine Health Monitoring System ((DNNHNS) for on-line engine health management and fault detection capabilities.. The DNNHNS utilizes a new information processing architecture, the Dialectic Neural Network, coupled with a novel Topo Spectrogram for vibration signature representation to achieve more reliable and robust autonomous health monitoring and fault detection. The DNNHMS represents knowledge in a form that allows cognitive reasoning within a computer. The Topo Spectrogram converts a large amount of complex engine vibration signals into succinct image-like patterns containing dynamic features critical for diagnostic evaluation. When integrated with Dialectic Neural Networks, cognitive knowledge is stored as objects rather than cumbersome rules, which is highly suitable for autonomous health monitoring and fault detection. The significance of the proposed program is attributed to its enhanced on-line signature processing and representation capability along with the innovative information processing architecture for fault pattern detection and identification. Successful development of the DNNHMS can reduce catastrophic engine failure risks and will improve reliability of NASA's advanced propulsion systems. Phase I will demonstrate the feasibility and relative benefits of DNNHMS. In Phase II a PC-based DNNHMS hardware/software prototype will be demonstrated.
Potential Commercial Applications:A portable low-cost Engine Health Monitoring System has strong commercial application. In Particular, in the commercial transportation and power generation industries as well as in the manufacturing sector where production lines frequently rely on critical, active machinery. Implementing an effective health monitoring system in these commercial arenas would reduce the risks of catastrophic hardware losses and plant down-time. The commercial potential for an effective DNNHMS in the market place has been recognized by several of ASRI's non-defense customers. In particular, ASRI has assisted both Reynold's Metal Company and Seagate Technology in developing reliable machinery health monitoring capabilities to enhance the safety and reliability of their production line machinery.