Fault detection and diagnosis in batch and fed-batch bioreactor system using PCA, PCR, PLS with GLR method.
ChiranjiviBotre, M. StefanyAngarita Gómez, MajdiMansouri, Mohamed N. Nounou, Hazem N. Nounouand M. Nazmul Karim
Multivariate statistical methods are powerful tools capable to handle huge, noise and highly correlated data sets. Principal Component Analysis (PCA), Partial least squares (PLS) as well as Principal Component Regression, linear and non-linear approaches, will be used to detect and predict faults in a batch and a fed-batch bioreactor system. Generalized likelihood ration (GLR) statistical method is used to detect the fault and it has been shown to have good fault detection abilities. Due to inherent dynamics in batch processes, multiblock and multiway methods have to be applied in which only historical data sets of past successful batches are needed. Principal Components are kept to extract the majority of variance and reduce the dimensionality of the original data set, making easier the analysis and fault detection. As soon as the fault is detected, it is necessary to find, the variable that presents the major deviation from its expected value therefore, a contribution plot for diagnosis is required.Experimental data is of batch production of carotenoids via fermentation to describe glucose consumption, metabolic products formation and depletion, and the carotenoid production in the Saccharmoycescerevisiae strain mutant SM14 with 20 g/l glucose as the carbon source. Experiments are performed to validate the results obtained from models.
Keywords: PCA, PCR,GLR, fault detection, batch process, statistical process control.
- MacGregor, J.F., and T. Kourti. “Statistical Process Control of Multivariate Processes.” Control Engineering Practice 3, no. 3 (March 1995): 403–14.
- FouziHarrou, Mohamed N. Nounou, Hazem N. Nounou and MudduMadakyaru, “Statisticalfaultdetectionusing PCA-based GLR hypothesistesting” Journal of Loss Prevention in the Process Industries 26 (2013) 129e139.
- S. Joe Qin, “Statistical process monitoring: basics and beyond” Journal ofChemometrics 2003; 17: 480–502.