This paper describes the application of a machine vision sensor in a neural network-based supervisory control system for a microbial cultivation. The vision sensor was capable of counting the number of microbial cells and acquiring other microbial growth information like sporulation activities from the sample medium. The information was useful for classifying the state of the microbial cultivation and could be used for control and determination of nutrient addition. A neural network supervisory unit was used to tune PID controllers in order to obtain an optimal cellular growth environment throughout the process. The process was simulated with a neural network-based program and tested with a bench scale reactor. Promising results were obtained from both the computer simulation and experimental tests. The results suggest that application of machine vision sensors and neural network-based supervisory controls are attractive for microbial cultivations.
- Fuzzy logic
- Machine vision
- Microbial cultivation control
- Neural network
ASJC Scopus subject areas
- Applied Microbiology and Biotechnology