Real-Time Turbine Engine Diagnostic System

Our innovation is a real-time turbine engine diagnostic system that is capable of simultaneously monitoring for, distinguishing between, and classifying the source and type of sensor, engine and facility faults. This capability, in turn, enables the enginecontrol system or test engineers to take appropriate risk mitigation actions to protect personnel and equipment, maximize operating life, and ensure test or mission success. We will show that reliable turbine engine diagnostic systems can be economicallyconfigured and maintained using tools that automate the development, analysis, calibration and production of the real-time monitoring software. We will further show that this software detects and diagnoses engine component and sensor failures with veryhigh reliability within a fixed processing time for all engine operating conditions. The diagnostic software tools will include: a turbine engine model for parameter estimation; advanced statistical event and fault detection procedures; a Bayesian beliefnetwork diagnostic decision management procedure; and automated online engine model calibration procedures. These new algorithms will be integrated with elements of an existing data quality validation software application to create the engine diagnosticsystem prototype and demonstrate feasibility of the innovation.Requirements for real-time diagnostics and fault detection exist in a wide range of military, aerospace and industrial control and safety systems where unexpected process interruptions due tocomponent or sensor failures are unsafe or uneconomical. Expert Microsystems has commercialized an innovative SBIR-derived data quality validation software product addressing elements of this market. The proposed software will extend this capability toprovide comprehensive diagnosis and discrimination of component and sensor failures for turbine-driven machinery. Further, the software will automate the production of these diagnostic modules and will provide an economical capability to integrate thediagnostic module with the user's process control equipment. While turbine-driven equipment forms our initial target market, the methods and tools embodied in the software are extraordinarily general in their applicability to a wide variety of industrial,utility, business, medical, transportation, financial, and biological processes and equipment thereby providing a very large potential market and improved economic competitiveness over the long term.