A Fault Diagnosis Algorithm of Artificial Immune Network Model based on Neighborhood Rough Set Theory

This paper proposes a fault diagnosis algorithm of artificial immune network model based on neighborhood rough set theory. In the algorithm, the relationships between pruning threshold, the rates of mis-diagnosis, and missed diagnosis are discussed in the shape space. In addition, the fault mode boundaries, the fault mode inclusion relations, an observation index and an algorithm for adaptively adjusting pruning threshold are described. The simulation experiments show that the proposed fault diagnosis algorithm can identify the unknown and untrained fault modes, while keeping misdiagnosis rate and missed diagnosis rate low.