Neural networks-based negative selection algorithm with applications in fault diagnosis

In this paper, we first propose a novel neural networks-based negative selection algorithm (NSA). The principle and structure of our NSA are presented, and its training algorithm is derived. Taking advantage of neural networks training, it has the distinguished capability of adaptation, which is well suited for dealing with practical problems under time-varying circumstances. A new fault diagnosis scheme using this NSA is next introduced. Two illustrative simulations of anomaly detection in chaotic time series and inner raceway fault diagnosis of bearings demonstrate the efficiency of the proposed neural networks-based NSA.