MLS, a machine learning system for engine fault diagnosis
In the research reported in this paper, efforts have been made to define and employ heuristic as well as algorithmic rules to conceptualize numerical data produced by normal and faulty jet and rocket engine behavior examples, employed in developing the machine learning system called MLS.
Abstract
Numerical data generated by a process usually represents concepts which could be utilized by other processes. However, inferring a concept from a large volume of numerical data is a complex task even for human experts. Though transforming numerical data into diagrams often reduces the complexity of understanding, a large volume of data generating hundreds of diagrams is difficult to comprehend for domain experts. Our analysis, however, indicates that human experts do not find it as difficult to outline the rules to comprehend data. In the research reported in this paper, efforts have been made to define and employ heuristic as well as algorithmic rules to conceptualize numerical data produced by normal and faulty jet and rocket engine behavior examples. These rules have been employed in developing the machine learning system called MLS. The input to MLS is examples which contain numerical data of normal and faulty engine behavior and which are obtained from an engine simulation program. This data includes sensor values in time, start and end points of the significant variations in the sensor values, and first and second derivatives of these variations. MLS first transforms the numerical data into discrete selectors. Partial descriptions formed by those selectors are then Permission to copy without fee all or part of this material is granted provided that the copies are not made or distributed .for direct commercial advantage, the ACM copyright notice and the title o f the publication and its date appear, and notice is given that copying is by permission of the Association for Comput ing Machinery. To copy otherwise, or to republish, requires a fee and /o r specfic permission. © ACM 1988 0-89791-271-3/88/0006/0721 $1.50 721 generalized or specialized to generate concept descriptions about faults. The concepts are represented in the form of characteristic and discriminant descriptions, which are stored in the knowledge base and are employed to diagnose faults. MLS has been successfully tested on jet engine as well as rocket engine