An enhanced dynamic artificial immune system based on simulated vaccine for early fault diagnosis with limited data

Artificial immune system (AIS) shows better performance with less training data in the process safety and risk engineering. However, the traditional AIS based fault diagnosis model is invalid for new process with no history data. Meanwhile, the data-based method has the limitation. To improve the process safety of chemical system under the extreme absence of data, an enhanced dynamic artificial immune system based on simulated vaccine and correlation coefficient methods (SV-CCDAIS), has been proposed. First, simulated vaccine was used to obtain the simulated data from simulation software or modelling for online diagnosis of actual process. Second, dynamic time warping was used to align the normal process data to solve the problem of time dimension misalignment. Third, different dynamic correlation measurement methods were proposed. Finally, a complete diagnostic and autonomous updating process was designed. The proposed method was applied to an ethanol-water separation start-up process, results confirmed that the proposed method exhibited 1.5 time higher fault diagnosis accuracy compared to the convolutional neural network method based on dynamic kernel principal component analysis, in addition, and the average time of fault diagnosis is 5 s, which is shorter than comparative method. (c) 2022 Published by Elsevier Ltd on behalf of Institution of Chemical Engineers.

Automated summary

An enhanced dynamic artificial immune system based on simulated vaccine and correlation coefficient methods (SV-CCDAIS) is proposed to improve the process safety of chemical systems with extreme absence of data. By using simulated vaccine, dynamic time warping, and different dynamic correlation measurement methods, the proposed method achieves higher fault diagnosis accuracy and shorter diagnosis time compared to the comparative method.