COMPUTATIONAL SOFT SENSOR FOR FUNGAL BIOFILTRATION PROCESS

Based on the data for the experimental measurements of the state on the fungal biofilter some process variables: the CO2 concentration and the elimination capacity (CE) have been estimated using a differential neural observer scheme via the pressure difference data. This scheme is developed in two parts, the first is the dynamical neural network structure and the second is compound by the observer structure, this type of computational sensor is called soft sensor. The good performance for the estimate states is shown by the CE and CO2 dynamical evolutions versus their estimate states on graphical way.