SBIR-STTR Award

Current Joint Operations in the Middle East have highlighted the need for increased system reliability and reduced petroleum consumption as both a cost reduction and force protection mechanism in the tactical battlefield. Williams-Pyro, Inc., is proposing to develop the Generator Fault Investigation Technology (GenFIT) system to perform diagnostics and prognostics on diesel generators with the end result of helping the Army achieve its HI-Power goals through reducing generator down time, improving fuel efficiency, and reducing emissions. The GenFIT system will be able to easily integrate with deployed TQGs to provide diagnostic and prognostic information to maintenance personnel, reducing the time required to service these generators. Many existing condition-based maintenance (CBM) systems are extremely complex, relying on neural networks and pattern recognition algorithms that need large amounts of equipment-specific training data. In contrast, Williams-Pyro is proposing to use a top-down approach to develop a first order diagnostics and prognostics methodology based on heuristic models derived from an understanding of diesel generator operating principals, observed generator performance values, and known generator parameters. The technology developed will be able to identify long-term generator performance degradation as well as reduce fuel consumption and emissions through proper maintenance and operation of the generator.

Keywords:
Hi-Power, Condition-Based Maintenance (Cbm), Diagnostics / Prognostics, Tactical Quiet Generator (Tqg), Rule-Based Heuristic Algorithms, Generator Retrofit, Acoustic / Vibrati

Williams-Pyro has established the feasibility of its Generator Fault Investigation Technology (GenFIT) system architecture for heuristic prognosis of tactical quiet generators (TQGs). The GenFIT architecture will provide real-time, measurement-based health status and fuel use rates of the TQG as a system. Williams-Pyro addressed two tasks to verify our approach for improved diagnostic and prognostic methodologies of diesel engine generators. The first task focused on finding methods for extracting the Discrete Event Set (DES) from sensor data. The DES approach is unique. It works on different types and sizes of diesel engines with little or no modification because it relies on the exact sequence of events that must occur for any diesel engine to run. The DES is extracted from sensor data using discrete wavelet transforms, and is compared to a baseline model of the four cycle diesel engine. The second task involved evaluating traditional sending unit data from instrument panel gauge. This type of evaluation provides additional details which the operator cannot see, even if they were to monitor the gauge continually. Successfully completing these two tasks has addressed any technical uncertainty about our approach to the challenging problem of automating the evaluation of diesel engine generator conditions.

Keywords:
Genfit, Diagnostics, Prognostics, Tqg, Diesel Engine Maintenance, Hi-Power, Fuel Efficiency