LAUSR

Abstract In this work we propose a novel fault detection (FD) technique in order to enhance monitoring of biological processes. To do that, a new statistical FD method, that is based on combining the advantages of the double exponentially weighted moving average (EWMA), called Max-DEWMA, with those of the particle filtering (PF), and multiscale representation is developed. The advantages of PF-based multiscale (MS) Max-DEWMA (M-DEWMA) are threefold: (i) the dynamical multiscale representation is proposed to extract accurate deterministic features and decorrelate autocorrelated measurements; (ii) PF is proposed to estimate the states of biological processes; (iii) MS-M-DEWMA chart is able to detect smaller fault shifts in the mean/variances and enhance the monitoring of biological processes. The FD performance is studied using Cad System in E. coli (CSEC) model. PF-based MS-M-DEWMA is used to enhance FD of the CSEC model through monitoring some of the key variables involved in this model such as enzymes, lysine and cadaverine.