J

40 fc

Fig. 28.3. Water content control during a fermentation run with the 200-kg bioreactor. Key: (--) water content set point; (- • -) measured water content of the bed. The vertical bars represent the volume of water added

10 15 20 25 30

Fig. 28.4. DMC strategy applied to the 200-kg bioreactor. Better temperature control is achieved when two manipulated variables are moved simultaneously, which can be seen by comparing this figure with Fig. 28.2. Key: (__) Bed temperature set point; (—) maximum bed temperature; (•) average bed temperature; (---) minimum bed temperature

10 15 20 25 30

Fig. 28.4. DMC strategy applied to the 200-kg bioreactor. Better temperature control is achieved when two manipulated variables are moved simultaneously, which can be seen by comparing this figure with Fig. 28.2. Key: (__) Bed temperature set point; (—) maximum bed temperature; (•) average bed temperature; (---) minimum bed temperature

28.3.2 Model-Based Evaluation of Control Strategies

An intermittently-mixed forcefully-aerated bioreactor, presented in Fig. 25.1, was modeled using the program presented in Chap. 25, the equations of which are shown in Fig. 25.2. Mixing was triggered when the outlet gas water activity fell below a set point. However, unlike the case study in Chap. 25, in which the inlet air conditions were maintained constant during the fermentation, in the present case study a control scheme was implemented to control either or both of the temperature and humidity of the inlet air, based on the average of the temperatures measured at different heights within the bed (Fig. 28.5). The success of the control scheme was evaluated on the basis of the fermentation profile for the average biomass concentration within the bed.

The present case study highlights the main points of interest that were identified in the work of von Meien et al. (2004). Readers interested in a deeper analysis should consult the original paper.

Figure 28.6 shows simulations done with a PID (Proportional-IntegralDerivative) controller, using two different strategies:

• Humidity control. In this case the relative humidity of the inlet air is varied by the controller, while the temperature is maintained at 38°C. Figure 28.6(a)) shows that the average temperature in the bed varies significantly from the optimum of 38°C throughout the fermentation.

• Temperature control. In this case the temperature of the inlet air is varied by the controller, while the relative humidity is maintained constant. In different simulations the constant relative humidity is maintained at 80% (Fig. 28.6(b)), 90% (Fig. 28.6(c)), and 99% (Fig. 28.6(d)). In this case it is possible to control the average temperature of the bed much better, in other words, the deviations from the optimum temperature are smaller.

The effect of the better temperature control in the case in which the inlet air temperature is manipulated is clear: The growth profiles obtained with "temperature control" (Fig. 28.6(f)) are closer to the optimum than the growth profile obtained with "humidity control" (Fig. 28.6(e)). Note that the relative humidity of the inlet air has no effect on the predicted growth performance in the case of "temperature control". Therefore, it is best to maintain the air saturated since this is easier to achieve in practice than attaining a particular relative humidity set point (see Chap. 29).

Figure 28.7 shows simulations done with a DMC (dynamic matrix control) controller. DMC is a form of Model Predictive Control, which was discussed in Sect. 27.2.3. Again "temperature control" and "humidity control" strategies are compared, with "temperature control" being superior, as it was in the case of PID control (von Meien et al. 2004). Figure 28.7 also shows that DMC control presents an interesting challenge. As discussed in Sect. 27.2.3, model predictive control

Fig. 28.5. Bioreactor and control scheme for the case study of control of an intermittently-agitated forcefully-aerated bioreactor. The bioreactor simulated is 2.0 m high, with an air flux at the inlet of 0.06 kg dry air s"1 m-2. In practice it might be impractical to measure the temperature at many different heights within the bed. In this case a single measurement at a half of the overall bed height will probably be sufficient

Fig. 28.5. Bioreactor and control scheme for the case study of control of an intermittently-agitated forcefully-aerated bioreactor. The bioreactor simulated is 2.0 m high, with an air flux at the inlet of 0.06 kg dry air s"1 m-2. In practice it might be impractical to measure the temperature at many different heights within the bed. In this case a single measurement at a half of the overall bed height will probably be sufficient

Was this article helpful?

0 0

Post a comment