Increased load and more stringent discharge limits A Case Study at Käppala WWTP

Transcription

1 Increased load and more stringent discharge limits A Case Study at Käppala WWTP A. Thunberg*, T. Palmgren* * Käppala Association, Box 3095, SE Lidingö, Sweden andreas.thunberg@kapppala.se, torsten.palmgren@kappala.se Abstract. Because of an expected imposition of stricter limits on wastewater discharges greater emphasis must be placed on how to meet such demands cost effectively. At the Käppala wastewater treatment plant (WWTP) two different unit processes have been proposed to improve the treatment of the plant cost effectively with a low environmental impact; degasification of mixed liquor suspended solids (MLSS) or a fixed-film activated-sludge process. The degasification technology presents a way of increasing the hydraulic and organic capacity with only minor changes of the existing process. By subjecting the MLSS of the biological tanks to a vacuum before it enters the secondary clarifiers the sludge settling characteristics are greatly improved. With improved settling characteristics the solids loading rate can be increased which allows for higher concentrations of MLSS in the biological tanks. The organic capacity is thereby increased without changing the existing process configuration. The degasification of the mixed liquor is performed in a vacuum tower placed between the biological tanks and the secondary clarifiers. In order to achieve the necessary pressure drop of 0.95 bar the water is lifted 9.5 m above the water surface where it is degassed. The degasification technology has been implemented to several WWTPs world wide but is still fairly unknown in the Nordic countries. The effect on the settling characteristics depends on a large number of parameters. The process performance must hence be tested in a full scale test which will be conducted at the Käppala WWTP during The second unit process proposed is a fixed-film activated-sludge process which will enhance the activated sludge process by providing a greater biomass concentration in the tanks and increase the capacity of a given volume. By calculating the capital, operational and maintenance costs for both of the unit processes the vacuum technology has proven to be the most cost effective way for the Käppala Association to increase the capacity and meet stricter limits on wastewater discharges. In this paper the two technologies are compared economically in a scenario where stricter limits have been imposed. The results show that stricter discharge limits will likely lead to a large increase of the marginal treatment costs, i.e. the cost for the extra amount of nutrients removed. However, the total costs per inhabitant will decrease at the Käppala WWTP as an effect of large investments already having been made and an increasing number of inhabitants in the connected municipalities. Keywords Degasification of MLSS, MBBR, discharge limits. Background The Käppala WWTP is situated on the island Lidingö outside of Stockholm and receives sewage from eleven municipalities north of Stockholm. The plant is built in a densely populated area under ground to save space for housing and an expansion of the plant is only possible downwards. The recipient of the Käppala WWTP is the Stockholm archipelago in the Baltic Sea. The status of the sea is in general poor with heavy algae blooming during summer. To improve the situation several countries surrounding the Baltic Sea have come together in an intergovernmental cooperation called the Baltic Sea Action Plan, or BSAP. According to this agreement Sweden must decrease the load of phosphorous and nitrogen to the sea. Parallel with the ongoing work in the BSAP agreement the EU water framework directive has been implemented in Sweden to regulate the ecological status of lakes, rivers and coastal areas. The Swedish water authority has decreed that the ecological status of the recipient of Käppala WWTP should reach good status by the year The status is now moderate to poor. In order to reach the goals set by the BSAP and the EU water framework directive stricter limits on wastewater discharges are expected. Besides stricter discharge limits the load to the plant is also expected to increase over the coming decades. The Stockholm region is now expanding rapidly and about 2 million people are living in 26 municipalities in the county of Stockholm. In 2030 this number is estimated to be 2.4 million people and

2 of these will be connected to the Käppala WWTP. With an industrial load of approximately people equivalents (p e) the total load would amount to almost p e. The design load of the plant is p e and the current load p e. However, the design capacity of some parameters will likely be exceeded when p e is reached, Table 1. The nitrogen and BOD 7 load (biological oxygen demand) are not far from design load and with stricter limits greater capacity will be needed. The current discharge limits are 10 mg/l of total nitrogen, 8 mg/l of BOD 7 and 0.3 mg/l of total phosphorous as one year mean values. Table 1 Design, current and future load situation at the Käppala WWTP. Design Current p e p e Average flow 2.5 m 3 /s 1.8 m 3 /s 2.6 m 3 /s 3.4 m 3 /s Instantaneous peak flow 6 m 3 /s 6 m 3 /s 7 m 3 /s 9 m 3 /s BOD 7 load 42 ton/d 37 ton/d 49 ton/d 63 ton/d Nitrogen load 7 ton/d 6 ton/d 8.4 ton/d 10.8 ton/d Phosphorous load 1.4 ton/d 0.9 ton/d 1.1 ton/d 1.4 ton/d Because of this the Käppala Association has decided to investigate how to increase the plant performance in a cost effective way and with minimum environmental impact. During 2009 and 2010 a large number of physical unit operations and unit processes were investigated for the future plant configuration. Two unit processes were chosen as a first and second hand choice. The first hand choice, the degasification technology, presents a cost effective way of improving the capacity of the plant without any bigger changes of the existing process and with a low environmental impact. The technology must however first be tested at the Käppala WWTP as its performance depends much on site specific conditions. During 2012 and 2013 a full scale test is planned for evaluation. If the performance of the degasification technology proves to be less than required a fixed-film activated-sludge process is the second hand choice. To compare the two unit processes thorough calculations of plant performance, capital and operational costs have been performed for four different future scenarios. In Table 2 the conditions that were used in each scenario are shown. Table 2 Scenarios used for comparison of the degasification technology and fixed-film activated-sludge. Scenario 1 Scenario 2 Scenario 3 Scenario 4 Discharge of N-tot (mg/l) Discharge of P-tot (mg/l) Discharge of BOD 7 (mg/l) Load (p e) Plant description The Käppala WWTP was put in operation in It then consisted of 6 parallel treatment lines with a nominal capacity of p e. Between the plant was expanded with 5 additional lines and modified for biological phosphorous and nitrogen removal according to the University of Cape Town treatment process (UCT-process). In Figure 1 the flow diagram is shown. For more information about the history of Käppala WWTP see also Bengtsson & Palmgren (1999) and Manhem & Palmgren (2003). The preliminary treatment comprises of fine bar screens and aerated grit chambers followed by primary treatment with primary settling tanks. Secondary treatment and nutrient removal is performed with predenitrification and enhanced biological phosphorous removal (EPBR) according to the UCT-process. No external carbon is added to the secondary treatment. Following the secondary clarifiers are 30 dual media sand filters for tertiary treatment. Effluent phosphorous concentration is controlled by precipitation with ferrous sulphate in the sand filters. Sludge stabilization is performed in two mesophilic digesters. Produced biogas is upgraded to vehicle fuel quality and used for public transport in the Stockholm area. Final dewatering is made with a chemical conditioning with sulphuric acid and hydrogen peroxide followed by mechanical dewatering with cylinder filter presses (Thunberg, 2010).

3 Figure 1 Flow diagram of Käppala WWTP. Plant weaknesses The treatment efficiency is high seen over the year and current discharge limits have been kept since the last expansion. The plant does however encounter the same problem as most activated-sludge (AS) plants around the world; poor settleability of the mixed liquor (Metcalf & Eddy, 2003). The reasons for poor settleability can be many but among the most common ones is bulking sludge caused by filamentous bacteria or viscous bulking caused by an excessive amount of extracellular polymers (EPS). In a bulking sludge condition the MLSS floc does not compact or settle well and flock particles can be discharged in the clarifier effluent. Different process parameters e.g. food to biomass ratio (F/M), dissolved oxygen (DO) or complete mix operation all effect the development of filamentous bacteria or the production of extracellular polymers. It is not uncommon with a recurring period each year with poor sludge properties. Poor settleability can however also occur despite low levels of filamentous bacteria and EPS. If denitrification takes place in the secondary clarifier nitrogen gas can become trapped inside the sludge and make it buoyant. The effect of a rising sludge blanket will be the same as with bulking sludge, i.e. loss of solids from the clarifiers and a risk of clogging the sand filters in the tertiary treatment. At the Käppala WWTP poor settleability often occur during winter when influent water temperatures are low. This is also the time when the solids retention time (SRT) of the activated-sludge must be kept high in order to achieve a sufficient nitrification. During winter the MLSS is normally kept at kg MLSS/m 3 in the biological tanks giving a solids loading rate in the final clarifiers of approximately 2.5 kg TSS/h m 2. The design maximum solids loading rate is 7 kg TSS/h m 2 but still loss of solids often occur at these conditions. Winter in Sweden is normally dry and long periods with small influent flows are common. The design surface loading rate in the final clarifiers is 0.8 m/h. During winter the actual surface loading rate is rarely higher than 0.5 m/h which will lead to an accumulation of sludge with poor settleability. The situation with poor sludge qualities caused by filamentous bacteria or excessive growth of EPS in combination with low surface loading rates for longer periods often leads to heavy loss of solids at wet weather conditions. Since spring normally mean spring flood in Sweden, the dry winter is often followed by peak flows and heavy loss of solids from the secondary clarifiers with clogged sand filters as a result (i.e. exceeded sludge storage capacity). When this happen the hydraulic capacity of the whole plant is reduced and overflow of primarily treated water becomes necessary. In Figure 2 the influent flow and the

4 two major overflow situations between are shown. During the last overflow in 2010 approximately 1.2 Mm 3 was by-passed of a total yearly amount of 52 Mm 3. This corresponds to 2.3 % which was enough to cause the quarterly average of effluent phosphorous to exceed the discharge limit of 0.3 mg/l. If stricter discharge limits are imposed it is obvious that similar situations must be avoided. Figure 2 Influent flow and overflow situations between Future process configuration Since the Käppala WWTP is situated under ground and there is no room to expand the plant sideways new unit operations or processes must fit inside existing volumes. To achieve the necessary increase of treatment a large number of methods are possible to utilize. Calculations show that the existing volumes are enough to handle a load of p e with stricter discharge limits (scenario 4, Table 2). Some unit processes would however increase the electrical and chemical consumption dramatically. The degasification technology on the other hand presents a way of increasing the capacity with a low energy and chemical consumption and minor changes to the existing process configuration. The second most cost effective method is to add carriers to the biological tanks to support biofilm growth in a moving-bed biofilm reactor (MBBR) which is a fixed-film activated-sludge process. Both of these methods have been investigated in detail to find the costs of an increased load and stricter discharge limits at the Käppala WWTP. Besides the new biological processes other general changes would also be required as the load increases, e.g. increased capacity of digesters and sludge dewatering. The peak flow situations that normally occur during spring must also be addressed separately. New unit operation storm water treatment With a load of p e peak flows of 9 m 3 /s are expected, Table 1. The design capacity is only 6 m 3 /s and occasionally overflow occurs at smaller influent flows. With stricter discharge limits, mainly

5 regarding phosphorous and BOD 7, overflows cannot be allowed and the future process configuration must be strengthened hydraulically. Before making any decisions on what the hydraulic capacity should be for the future configuration the occurrence and duration of the peak flows must be considered. In Figure 3 the current and future ( p e) influent flows are shown as one day mean values. The graph show that flows above the design capacity will occur but at less than 1% of the time. This indicate that it would be unnecessary to increase the hydraulic capacity of the main treatment lines as they would then have a larger than needed capacity at 99% of the time. Instead an effective and compact storm water treatment unit is a more cost-effective choice. There are three decomissioned gritchambers at the Käppala WWTP and in one of these a compact high-rate clarification unit could be placed. This technology is well known and has been implemented in many plants world-wide. With ballasted flocculation and lamella plate clarification a capture rate of 90% of total phosphorous, 70% of BOD 7 and 10% of total nitrogen will be achieved at flow rates up to 3 m 3 /s. The surface loading rate to the clarifier is approximately 80 times greater compared to conventional direct precipitation. With this unit operation wet weather flows can be handled over several days without risking effluent phosphorous of 0.1 mg/l as a one year mean value. Figure 3 Current and future influent flows at the Käppala WWTP. New unit processes degasification of MLSS The existing volumes at the Käppala WWTP are enough to handle a load of p e and stricter discharge limits. In theory the organic capacity could easily be increased sufficiently just by increasing the concentration of the MLSS and possibly start adding external carbon (e.g. methanol or ethanol). In practice this would however exaggerate the problem with poor settleability of the mixed liquor and lead to heavy loss of solids at wet weather flows. The settling properties must first be improved. The degasification technology is a method of doing this. At the end of the biological reactors the water is saturated with gases, e.g. nitrogen gas, oxygen and carbon dioxide. Because of the saturation the sludge flocs contain micro-bubbles of gas which makes sludge settling difficult or even cause it to float (Maciejewski et al 2009). The buoyancy is further aggravated by filamentous bacteria or excessive growth of EPS. In the secondary clarifiers even more gas can be produced if denitrification occurs. By subjecting the water and sludge to a very low pressure dissolved gases will be released making it unsaturated which removes the gas bubbles and prevent the formation of new bubbles. The process is based on Henry s law which states that the amount of a given gas dissolved in a liquid is directly proportional to the partial pressure of that gas. Lowering the pressure of any gas in equilibrium with a liquid will cause the gas to escape from the liquid. For example the solubility of nitrogen gas in water is approximately 17 mg/l at 15 ºC and atmospheric pressure. At 0.05 bar, which is the pressure in the degasification unit, the solubility is only 1 mg/l. Such water will be well under the saturation level for nitrogen gas as it enters the secondary clarifiers. Due to a limited residence time in the degasification unit

6 this theoretical steady state value may not be achieved. An approximation is that only 50% of the potential gas amounts are removed (Maciejewski et al 2009). This is still enough for the degassed biomass to have a potential to absorb 5-10 mg/l of nitrogen gas in the secondary clarifiers before the liquid once again becomes saturated. The degasification is performed in a vacuum tower located between the last cell of the biological tanks and the secondary clarifiers, Figure 4. The top of the vacuum tower is placed 9.5 m above the water level in the biological tank to achieve a pressure drop of 0.95 bar. By connecting a vacuum pump to the top of the tower a low pressure can be maintained. The water is transported through the tower with a siphon effect making the relative energy consumption low, kwh/m 3. To overcome the internal hydraulic losses of the tower small openings are drilled at the base of the tower which functions as an airlift pump. Besides lowering the concentration of gases the process also changes the floc structure. The treatment is divided into three phases. In phase 1 the MLSS flows from the biological reactor to the top of the tower. As the pressure drops gas bubbles expand and the floc structure is destroyed. In phase 2 the removal of gases occurs at the top of the tower. In phase 3 the MLSS flows down the tower to the secondary clarifier and new flocs are formed as the pressure is restored to atmospheric conditions Figure 4 Flow diagram of the degasification technology. By improving the settleability of the sludge the concentration of MLSS in the biological tanks can be increased. The degasification technology is now installed in approximately 40 plants world-wide. Most of these plants have remarkably high concentrations of MLSS in the biological tanks, 6-8 kg/m 3, and with solid loading rates of 8-10 kg TSS/h m 2 in the secondary clarifiers. The Käppala Association has performed laboratory test of the technology with very good results and will install a full-scale degasification plant in 2012 to evaluate the technology on-site. By increasing the MLSS to 6 kg/m 3 the organic capacity of the plant will be sufficient to handle p e and stricter discharge limits with no or only little use of external carbon source. New unit processes MBBR Different types of fixed-film activated-sludge processes are becoming more common in biological nutrient removal (BNR) plants. At the Käppala WWTP one way of increasing the capacity is to implement a MBBR process in some of the treatment lines. Calculations show that it is sufficient if 6 of

7 the 11 treatment lines are modified for MBBR in order to handle p e and stricter discharge limits. By adding carriers to the biological tanks growth of biofilm is induced. Since the microorganisms are growing on carriers instead of being kept in suspension shorter hydraulic retention times (HRT) and higher concentrations of sludge are possible (i.e. higher concentration of micro-organisms per tank volume). The MBBR does not require any return activated sludge (RAS) flow and the secondary clarifiers would only be used to settle sloughed solids. The solids loading rate on the clarifiers would then be greatly reduced and hence the risk of solids wash-out during wet weather flows. The negative impact of filamentous bacteria would also be less pronounced as these would grow in the biofilm as well. The main drawback with this unit process is the thickness of the biofilm which requires significantly higher concentrations of dissolved oxygen (DO) and readily biodegradable carbon compared to the conventional AS process. The concentration of fines in the effluent would also be increased and a post-precipitation with ferric iron would be necessary. Compared to the deagasification technology larger modifications of the existing plant would have to be made and the operational costs would increase. New unit processes phosphorous removal Much focus has been placed on finding cost effective and compact technologies to increase the organic and hydraulic capacity of the Käppala WWTP. However, with discharge limits of 0.1 mg/l of phosphorous the existing phosphorous removal process must also be overseen. The existing EPBR would unlikely be sufficient to guarantee effluent concentrations even as one year mean values. The main reason is the internal loading of phosphorous from the digesters that often occur with this type of process. The EBPR is also heavily dependent on the presence of volatile fatty acids (VFA) which commonly is scarce at wet weather conditions. Today the EBPR is compensated by post-precipitation with ferrous sulphate in the sand filters and if necessary simultaneous precipitation. With effluent phosphorous of 0.1 mg/l the dose of precipitant would have to be increased with the risk of clogging the sensitive sand filters as iron hydroxides are formed. If 0.1 mg/l of effluent phosphorous becomes a reality the future process configuration would be a combination of pre-, simultaneous- and post- precipitation. The simultaneous precipitation would be the backbone of the phosphorous removal process and remove approximately 90 % of the phosphorous. On-line measurements of effluent phosphate would control the dose of precipitant on the sand filters. If the dose is high for a prolonged period of time a pre-precipitation with ferric chloride would commence. The use of simultaneous precipitation with ferrous sulphate in all of the eleven treatment lines would increase the concentration of MLSS and solids loading rate to the clarifiers. With the decasification technology this is not believed to be a problem. If an MBBR process has been implemented to 6 of the treatment lines, the simultaneous precipitation would only be used in the remaining 5 lines. Financial aspects The degasification technology presents a cost-effective way of increasing the capacity of the plant with a low environmental impact. The current AS process would be left unchanged and easily expanded with additional degasification units as the need arises. With the expected performance of 6 kg/m 3 of MLSS external carbon source may not be needed, or only during some periods of the year. The MBBR process is much less cost-effective and has a larger carbon footprint because of the high energy and chemical consumption. Both technologies have been compared financially for the different scenarios presented in Table 2. In Figure 5 the total additional costs (capital and operational) for scenario 4 are shown for the two technologies. The x-axis shows the load to the plant since the time of implementation for each task in the graph depends on how fast the load increases. Stricter discharge limits have been assumed to be imposed simultaneously as p e is reached. A number of general modifications have been included which have to be performed regardless of which technology is used. These are storm water treatment, a new digester, decommissioning of EBPR and new dewatering equipment. The graph shows that the MBBR process is approximately 20% more expensive compared to the degasification technology. The main reason to this is the need for external carbon and high operational costs with MBBR but also larger capital costs as bigger changes of existing tanks would be required.

8 Figure 5 Additional capital and operational costs for both methods with scenario 4. In figure 6 the same data is shown but in relation to the number of p e connected. In the graph the future cost per inhabitant relative the current cost per inhabitant is shown ( future pe -1 / current pe -1 ) as the load increases. I.e. if the graph is horizontal the cost per inhabitant is unchanged. The graph shows that despite increasing total costs, Fig. 5, the costs for the inhabitants of the connected municipalities would decrease. This is the positive effect of the investments already made in a large WWTP as Käppala. It would not be possible to build a new WWTP with the same capacity to a cost this low. In Figure 7 the relative treatment costs are described as /kg OCP. OCP stand for oxygen consumption potential and is a way of describing the total oxygen consumption of nitrogen, organic carbon and phosphorous. The graph show that the relative costs decreases as the load increases for the same reason as in Fig. 6. The marginal costs due to stricter discharge limits are also shown for both technologies. This is the additional relative treatment cost of removing the extra amounts of nitrogen, BOD 7 and phosphorous required with stricter limits. It is apparent how these extra kilos would be twice as expensive as the average treatment cost. The marginal costs become the greatest for the degasification technology. The reason is that external carbon only becomes necessary for the degasification technology as stricter discharge limits are imposed. The MBBR technology has a high chemical consumption also without stricter limits and the difference hence becomes less pronounced.

9 Figure 6 Costs per p e. Future costs are divided by current costs. Figure 7 Relative costs per kg OCP removed. Marginal costs are the relative costs for additional amounts nutrients removed with stricter discharge limits. Conclusions To meet an increasing load and possibly stricter discharge limits the Käppala Association has found two unit processes available by which the current process configuration can be modified without the need of an expansion. The degasification technology presents a way of improving the settleability of the MLSS sufficiently to achieve 6 kg/m 3 in biological tanks without risking solids washout from the secondary clarifiers. This method has proven to be the most cost effective way of increasing the organic capacity of

10 the plant. A less cost-effective method would be to instead implement an MBBR technology. Both technologies require that a compact high-rate clarifier is installed to handle peak flows, especially if 0.1 mg/l of effluent phosphorous is imposed. Regardless of which technology is used the total operational and capital costs will increase greatly as the load increases and stricter discharge limits are imposed.. However, the cost per p e will decrease as the number of inhabitants increases as an effect of large investments already having been made. For the same reason the relative treatment costs will also decrease as the load increases. The relative treatment costs for the additional amounts of nutrients removed with stricter discharge limits will however be twice as expensive as the average relative treatment cost. This stresses the importance of carefully evaluating the environmental benefits of stricter discharge limits before they are legislated. References Bengtsson, B. & Palmgren T. (1999). Käppala 2001 Expansion of the Käppala Wastewater Treatment Plant, Project Management. In proceedings of the LWWTP conference Budapest Maciejewski, J. Oleszkiewicz, J.A. Golcz, A. & Nazar, A. (2009). Degasification of mixed liquor improves settling and biological nutrient removal. In proceedings of the 2 nd IWA specialized conference - Nutrien Management in Wastewater Treatment Processes 2009, Krakow, Poland Manhem, P. & Palmgren, T. (2003). Upgrading and Expansion of the Käppala Wastewater Treatment Plant; Operational experiences and results. In proceedings of the LWWTP conference Prague. Metcalf & Eddy (2003). Wastewater Engineering. McGraw-Hill, Boston MA, 1818 p. Thunberg, A. (2010). Optimizing Sludge Dewatering by Using the KemiCond Process with the Bucher Hydraulic Filter Press Full Scale Experiences at Käppala WWTP. In proceedings of the WEF RBC conference 2010, Savannah, US.