Application of On-Line Estimation Techniques to Fermentation Processes

This paper presents some recent industrial scale site test results of a new estimation algorithm for the on-line determination of biomass. The technique uses measurements from established instruments such as carbon dioxide analysers and flow transducers. The key development is an adaptive estimation relationship for inferring infrequently measured process outputs, from other more rapidly sampled secondary outputs. The structure of the estimator is not application dependent and its parameters can be continuously estimated and updated. As a result, slow variations in plant or disturbance characteristics can be tracked. In contrast to model-based estimators (eg. Kalman filtering methds), the method proposed requires minimal effort for estimator design. Production scale bioreactor tests demonstrate the capabilities of the new estimator.