The Application Of AIS In Fault Detection

With the increasing complexity of industrial process, fault diagnosis has attracted much attention recently for safe and effective operation. Since the development of artificial intelligence has provided new inspiration to the research of fault diagnosis, the application of artificial immunity mechanism in fault diagnosis is researched in this paper and new fault diagnosis methods are developed.The main contributions can be summarized as follows:1. Current fault diagnosis methods are reviewed briefly and the development and application of the Artificial Immune System (AIS) are introduced.2. Negative Selection (NS) Algorithm based on the self/non-self discrimination is described. The construction of binary Negative Selection, real-valued Negative Selection and the real-valued Negative Selection with variable-radius detectors is provided and the characteristics of different algorithms are compared.3. A new fault detection method based on Negative Selection Algorithm is proposed for complicated process. Principal Component Analysis (PCA) is utilized to reduce the dimensionality of the data. The detectors for fault detection are generated by applying the real-valued Negative Selection (NS) algorithm with variable-radius detectors. The simulation results on the Tennessee Eastman Process show the effectiveness of the proposed method.4. A new fault diagnosis method is presented based on hybrid immune learning algorithm, which combines negative selection algorithm and conventional pattern classification algorithm. Because it is difficult to make a complete catalog of all the possible and probable anomalous situations in reality, NS is used to generate the unknown fault samples which are needed as a part of training data for a classification algorithm. The simulation results on the IRIS dataset classifying problem illustrate that the performance of the proposed method is better than that of the conventional classification methods.