Figure 3.2 Schematic representation of
stages in a data acquistion system.
concomitantly, the window is re-centered around the new entry, and so on. This
technique assures that all relevant changes—including those not intended—of any
considered variable are trapped and that the dynamics of all signal trajectories are fully
documented. The only inconveniences with this data treatment are that one needs to time-
stamp every data entry individually and one ends up with non-equidistant data-vectors.
The benefit is a data-to-archive reduction usually by between 10−^2 and 10−^4 and the
assurance that no important data are lost.
Derived Variables
Some very important process variables are not measured on-line but can be easily
calculated from on-line measurements usually available in fermentation systems. They
are typically dilution rate (D), oxygen (OTR) and carbon dioxide (CTR) transfer rates,
the respiratory quotient (RQ) and the volumetric mass transfer coefficient kLa. The
calculation of OTR is particularly noise sensitive especially at low biomass
concentrations due to the small differences in oxygen concentrations between inlet—and
outlet-gas. Hence special care should be taken on noise filtering. As reported in the work
of Locher et al. (1993), many times due to unaccounted effects, erroneous experimental
procedures and sensor calibration problems, dramatic errors in the calculation of derived
variables may occur. It was shown that neglecting changes in the atmospheric pressure
may result in errors for RQ evaluation larger than 100%. Obviously, errors in derived
variables are the direct consequence of errors attached to on-line measurements.
Depending also on the way they are calculated, numerical error propagation may have
significant effects. Many other derived variables may be monitored on-line. When some
nutrients fed to the bioreactor are kept at constant concentrations by feedback control, the
corresponding consumption rate ri may be approximated by the relationship ri=D(ci−cFi)
D being the dilution rate (D, the ratio of feed rate to volume,=F/V), ci the concentration in
the medium and cFi the concentration in the input feed rate F. When two quantities are
kept constant the corresponding yield is easily evaluated also from the ratio between the
two rates. These variables may give important information about the micro-organism’s
metabolic state.
Multiphase bioreactor design 66