| EnviaTec | Investigation of possibilities for optimizing |
|
Market situation required a company from chemical industry to significantly increase production capacity. In connection with this a wastewater emergence changed in amount and quality was to be expected. With the help of dynamical simulations carried out with STOAT it was to be determined, how the subsequent wastewater treatment facility would cope with the new loads and whether there wouldhave to be investions into enhancement of the treatment facility. The calculations were based on the following scenarios:
The treatment facility to be simulated had been built as classical activated sludge treatment with upstream denitrification. Center of the facility were the denitrification tank (B 2), the aeration tank (B 3) and the secondary settlement tank (B 4). It is possible to add external c-sources (here: acetic acid) at the influent of the denitrification tank. Due to a proper working ammonia stripper installed at the production facility there is hardly any nitrogen at the influent of the treatment facility. This also makes internal recirculation of nitrate-rich medium unnecessary. As a result the B2 was more or less without use at the time of the investigation. The balancing tank (B1) is used to buffer the stream coming from the production facility with regard to amount and concentration. At 650m³ the volume of B1 is large enough to accept short stoppages of the chemical facility without impairing operation of the treatment facility. Modeling facility operation was based upon data from the second half of 2003. With the exception of a short shutdown of the chemical facility at the end of october, this timespan is characterized by mostly undisturbed operation. The balancing effect of the balancing tank during this phase is very clearly shown by the following plot (operational data).
Simulating the behaviour of the balancing tank was done using a PID controller (see following figure).
Using STOAT, the time behaviour of system states as well as outlet values of different process steps can be examined in detail. Fig. 3 is an example for presentation of important values of the final effluent, in this case the actual state during the second half of 2003. The abscissa shows simulated time in hours, while the ordinate shows the corresponding values of flow and composition. Below the figure you will find a summary of the values. As seen from this figure, the high COD outlet values especially during the first half of the simulation period are a consequence of comparably high quantities of dissolved, biologically degradable COD. This indicates, that the facility did not reach the optimum during this phase. The curves depicting ammonium and nitrate are only slightly above zero and can therefore hardly be seen in the figure. The extremely low nitrogen concentration is a result of the already mentioned wastewater pre-treatment close to production. Still, the values given below the figure point to problems with nitrification of ammonia.
COD curves for the worst case scenario (C) are similar to this plot, but are on average higher by ca. 30 mg/l and have more pronounced peaks. The outlet values for ammonia were of special interest to the customer, as the permit value is especially low here at 2 mg/l. To have a closer look at the nitrogen fraction within the final effluent for the worst case scenario (C), all other parameters are not shown in fig. 4.
The turquoise curve in Fig. 4 shows the development of ammonia concentration. The permit value of 2 mg/l is exceeded several times over extendet periods of time, although the mean value lies below. The seemingly absurd negative values result from the use of Activated Sludge Model Nr. 1 for simulation of the activated sludge process. This model does not take into consideration lack of the nutrients nitrogen and phosphorus. Even if there is not enough bio-available nitrogen at the influent of the biological step, simulated growth of heterotrophic and autotrophic organisms is unaffected. The TKN used for building up biomass results in negative concentrations in the effluent. In other words, the negative ammonia values show a disturbance of the C-N-P ratio to the debit of bio-available nitrogen. The constantly neglectably low nitrate concentrations end-of-the-pipe in combination with high ammonia values are a strong indication of the absence of nitrification. The reason for this can be deduced from a closer look at the processes in B2 and B3. For this purpose 3D plots of the concentrations of autotrophic and heterotrophic organisms as well as ammonia and temperature over time were generated (Fig. 5 to Fig. 8). For the plots the flow of the wastewater is along the stage axis from left front to right back; «Stage 1» represents the prepended denitrification, «Stage 2» the aeration step (nitrification). Time elapses along its axis from right front to left back. The respective values are shown on the vertical axis.
It follows from Fig. 5, that the concentration of the autotrophic organisms remains zero for almost the whole simulation period in both tanks. The reasons for this can in turn be seen from ammonia (Fig. 6) and temperature (Fig. 7) curves. As shown in Fig. 6, there repeatedly is scarcity of ammonia, which results in the autotrophs dying. Even when there is enough NH4-N in the influx again, they cannot grow again fast enough due to their slow growth rate. The frequent change between ammonium scarcity and ammonium supply, in connection with an unfavourable temperature profile, especially during the first half of the simulation period, leads to the unsuitable conditions shown in Fig. 5.
The high temperature in B2 and B3 during the first half of the simulation period (july, august and september) apparently also adversely affects growth of the heterotrophs. This follows from comparison of Fig. 7 and Fig. 8. Together this also explains, why such high concentrations of dissolved, bio-degradable COD in the efflux of the treatment facility have been calculated for this timespan (Fig. 3).
The results of the stoat simulation and conclusions from those for the different scenarios after increase in production capacity where presented to the staff running the production facility. As a first measure, it was proposed to guide process wastewater efflux through a heat exchanger to constantly keep the temperature below 36 degree Celsius.
The partly insufficient ammonia concentration at the influent of the treatment facility is a consequence of ammonia stripping. It was therefore possible to compensate the TKN deficit by directly feeding a wastewater stream extracted before the stripper into B1. According the staff this too could be implemented rather easily and was therefore accepted as a second optimization. The consequences of these planned changes were checked with an additional simulation run. The input files were adapted in the following way: the temperature was fixed to 35 degree Celsius and the NH4-N concentration was raised by including an additional load matched to the optimal BOD-TKN ratio of 100 : 5. Originally the NH4-N concentration at the influent of the treatment facility was on average 5.2 mg/l. This was increased by 67.5 mg/l to 72.7 mg/l, i.e. 13 times as much (mean values given).
The changes in the efflux of the treatment facility due to the previously explained optimizations are visible from Fig. 9: the outlet values have been significantly improved compared to the original scenario C. One major change for the COD curve is the disappearance of the large peak in the first half of the simulation period. It was possible to lower the mean concentration of dissolved, bio-degradable COD from 80.4 to 47.1 mg/l, the maximum value from 189.4 to 58.0 mg/l. Analogous changes were achieved for total COD.
Most important result is, though, the expected improvement regarding ammonium. After a short settling phase at the beginning, the curve runs well-balanced within a concentration range maintaining a comfortable safety margin to the permit value. The nitrate values, which also maintain a constant level after the short settling phase, show, that now a robust nitrification is taking place (Fig. 10). The difference to Fig. 4 can hardly be overlooked. A look at the population of autotrophic and heterotrophic microorganisms in B2 and B3 shows the resulting improvements so to speak from the inside: The population of the autotrophs remains stable over the whole simulation period, their concentration increasing along the flow path of the wastewater (Fig. 11). This allows for the effective nitrification. Accordingly, after a short settling phase, the ammonia concentration of the B2 influx is degraded to values constantly clearly below 2 mg/l in the B3 efflux. That the ammonia concentration at the influent of B2 is only roughly half the concentration set via the input file, results from “diluting” with the almost ammonium-free activated sludge return (30 m³/h).
Compared to the situation of scenario C without the use of a heat exchanger Fig. 8, the concentration of heterotrophs is higher especially during the first half of the simulation period (Fig. 12). This leads to a significantly improved degradation of the substrate and therefore lower COD values in the efflux of the facility. These improvements in degradation are also reflected in the sum of the loads of the final effluent over the whole simulation period (see following table).
As already mentioned, the results of this simulation were discussed with the staff of the operating company. The proposed changes were realised within a few weeks. Since then the facility is running without any complaints. This example, too, shows how dynamical simulation allows for a significantly improved understanding of the highly complex processes in biological wastewater treatment facilities. This is the basis for optimizing facilities and processes and saving investments as well as operating costs. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Top of the side | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
