Engine Data Mining Software Provided New Capability for Air Force Engineers

Applications

• Military and Commercial Significance:
• Researchers fully developed and demonstrated the new Web-based search capability, which allows users to log in to the newly implemented inspection database and access data from any networked computer, run customized data mining agents through a Web browser, and receive reports containing desired information. Managers can define an inspection time interval, such as Monday through Friday, and obtain reports pertaining to that period. They can also receive reports on demand or have the software automatically forward updated reports. For example, the new data mining capability can provide inspection reports on parts that cause the most frequent inspection errors or on a specific engine part’s length of inspection.
• Scientists originally developed the Intelligent Agent Architecture search program for use with data collected from the Eddy Current Inspection System (ECIS) stations under AFRL’s Engine Rotor Life Extension (ERLE) program. Under ERLE data management activities, the inspection database collects and records all inspection data from the ECIS engine inspections. This includes parts information, probe information, coordinates, inspection time, and raw inspection data. This-sort-and-search method provides managers with a tool to quantify and analyze the efficiency of the 31 inspection stations at Oklahoma City Air Logistics Center, Tinker Air Force Base, Oklahoma.
• The Intelligent Agent Architecture accommodates integration of other data sources. For example, ECIS data may be integrated with the Retirement for Cause database, which stores data associated with an engine part’s flaw. By integrating a part’s flaw information with its coordinate data and incorporating a data visualization technique, the engine owners can create parts diagrams that identify the part’s flaw locations. Similarly, this existing data may be coupled with future data sources, such as engine usage data, to generate a detailed depiction of an engine’s condition. The Intelligent Agent Architecture will allow integration of all data sources to enable searches for trends ranging from overall engine inspection processes to a specific engine’s remaining life.