No abstract

No abstract

In large-scale stirred bioreactors, used for aerobic microbial production processes, the liquid-phase bulk mixing will strongly affect the overall oxygen transfer. This will result in oxygen concentration profiles in the liquid phase, especially in highly viscous fermentation broths.',* However, in a low viscous fermentation broth oxygen concentration gradients can also be observed. These gradients are caused by the relatively high overall mixing time of the liquid phase, compared with the time constant for oxygen consumption and for oxygen t r a n~f e r .~ When a multicompartment model for the liquid flow and the oxygen transfer into the liquid phase is used, it is possible to predict the overall oxygen transfer capacity of the reactor quite a~curately.~ This is contrary to the use of empirical correlations such as those suggested by Van 't Riet.4 The two-compartment model used for such calculations as reported before3 now will be changed into a five-compartment model. This is acceptable only if the extra parameters can be predicted a priori. The number of compartments is determined by the number of impellers, mounted in the vessel, the part of the reactor below the lowest impeller, the reactor part between the impellers (if the distance between the impellers is larger than the impeller diameter) and the upper part of the reactor volume. So, in this case with two impellers, a five-compartment model will be used (Fig. 1). For the model calculations we assume that the gas is introduced into the lowest impeller region, and that the oxygen transfer in the bottom part of the vessel is negligible. The aim of the model is to predict the following reactor operating variables: 1) the overall oxygen transfer capacity of the reactor (OTR); 2) the local liquid dissolved oxygen concentrations, for estimation of bad aerated zones which can introduce negative effects for the microorganisms and as a base for reliable scaledown experiments to estimate those effects5; 3) the gas-phase exhaustion.

No abstract

In this article a dynamic model of a continuous working UASB reactor is described. It results from the integration of the fluid flow pattern in the reactor, the kinetic behavior of the bacteria (where inhibition and limitation were taken into account), and the mass transport phenomena between different compartments and different phases. The mathematical equations underlying the model and describing the important mechanisms were programmed and prepared for computations and simulations by computer. The settler efficiency has to be over 99% to prevent the reactor from wash-out. When the settler efficiency is over 99%, the total sludge content of the reactor increases steadily, so the reactor is hardly ever in a steady state. This implies dynamic modeling. The model is able to predict the various observable and nonobservable or difficult to observe state variables, e.g., the sludge bed height, the sludge blanket concentration, the short-circuiting flows over bed and blanket, and the effluent COD concentration as a function of the hydrodynamic load, COD load, pH, and settler efficiency. The optimal pH value is between 6.0 and 8.0; fatty acid shock loadings are difficult to handle outside this optimal pH range.