Anaerobic digestion process parameter identification and marginal confidence intervals by multivariate steady state analysis and bootstrap

There are a few works related to on line steady state detection algorithms and less with parameter estimation. This work used Principal Component Analysis (PCA) for reduction of the dimension of the data space and producing independent variables, allowing the application of the multivariate Cao and Rhinehart algorithm in steady state detection. Once steady states were detected, model parameters can be calculated by steady state mass balance equations and non linear fit of data. Bootstrap method was used in order to approximate the parameters estimators distribution and confidence boundaries for the kinetic model. These methodologies were applied to a case of anaerobic wastewater digestion process where four different organic loading rates (OLR) were applied.