FUZZY MODELLING IN BIOTECHNOLOGICAL PROCESSES: ENHANCING OF PARAMETER OBSERVABILITY

摘要

Biotechnological processes show an high level of complexity. The approach to their description cannot be limited to the assembling of simple mechanisms which account for main phenomena. A macroapproach has been proposed based on the evaluation of key parameters which directly or indirectly account for the relevant phenomena. The key parameter must satisfy some constrains in order to be eligible, the main one is the observability that involve different considerations about achievability, feasibility, practicability, time, cost, availability and simplicity. This paper propose a simple but effective methodology to improve observability. A secondary fuzzy algorithm based on off line measurements is set up to evaluate the low observable parameter. The secondary fuzzy algorithm is then inferred serially with the main fuzzy algorithm and contains a predictive evaluation of the parameter. It is important underline that only an initial fuzzification and a final defuzzification need reducing the uncertainty contained in these steps. The methodology has been applied to the description of ethanol production by Saccharomyces cerevisiae entrapped in Ca-alginate.