Back to: Projects

Phosphate Recovery

Adapting Strategies for avoiding struvite build-up in plants and transforming them into strategies for recovery

 

Steve Williams
Thames Water Utilities Ltd.
Spencer House, Manor Farm Road, Reading, RG2 0JN.
 

Abstract

Phosphorus removal as part of wastewater treatment is becoming more widespread in the United Kingdom as the Urban Wastewater Treatment Directive is implemented. Phosphorus removal can be achieved by either chemical or biological means or by a combination of both methods. When the phosphorus is removed biologically the phosphorus stored in the waste sludge may be released during anaerobic digestion. As digested sludge contains a high concentration of ammoniacal nitrogen it is likely that struvite may be formed if the magnesium concentration is sufficiently high.

At Slough wastewater treatment plant struvite precipitation has been a problem for a number of years. A recent extension to the plant built as a 3-stage Bardenpho plant has reduced the phosphorus concentration in the effluent and lead to an increase in struvite precipitation in the digested sludge pipelines. These problems have been overcome in the short term by dissolving the struvite in 10% sulphuric acid. It is hoped to provide longer term protection against struvite precipitation by installing magnetic water softening devices at key locations. This form of protection is already providing some benefit at one point on the site.

The phosphorus concentration and loads at a number of points on the site has been investigated. By combining two streams, the centrate from the sludge dewatering plant and the liquor from thickening the waste activated sludge, a relatively clean stream containing 160 mgl-1 of soluble phosphate, a load of 136 kgd-1. In order to make best use of this stream it is important to prevent struvite formation at other points in the plant where the precipitation can cause operational problems. If magnetic water softeners prove to be successful they will provide an important part of the phosphorus recovery from wastewaters.
 

Introduction

The need for nutrient removal as part of the wastewater treatment process in the United Kingdom has arisen from the Urban Waste Water Treatment Directive which was adopted by the EC on March 18, 1991 (1). The Directive sets discharge limits for some of the established sanitary determinands e.g. BOD and suspended solids, and the nutrients nitrogen and phosphorus. The limits for nutrients are shown in Table 1 and are annual average values based on 24 hour composite samples.

Table 1. Limits for nitrogen and phosphorus as required by the Urban Waste Water Treatment Directive.
 

Parameter Population 10,000 to 100,000 Population  > 100,000 Minimum Reduction
Total Nitrogen * 15 mgl-1 10 mgl-1 70 % - 80 %
Total Phosphorus 2 mgl-1 1 mgl-1  80 %
*Alternatively the daily average total nitrogen concentration must not exceed 20 mgl-1, and the standard only applies when the wastewater temperature is greater than 12oC.

These standards become effective on or before December 31, 1998 for wastewater treatment works which discharge to watercourses that are designated as "sensitive." The DoE has designated much of the non-tidal Thames and some tributaries as sensitive. This designation may require, in the future, phosphorus removal at as many as 65 wastewater treatment works.

Phosphorus removal can be achieved chemically by precipitating the phosphorus with a metal salt e.g. Iron II, Iron III, Aluminium and calcium. Biological removal requires the bacteria to be exposed to firstly anaerobic and then, secondly anaerobic conditions. This sequence helps to select groups of bacteria that are able to concentrate the phosphorus in their cells as poly- phosphate chains (2) . However when these bacteria are exposed to anaerobic conditions, as in an anaerobic digester, some of the phosphorus is released as soluble phosphate. In order to retain the phosphorus in the sludge it may be necessary to precipitate the phosphorus as a metal phosphate or as struvite.
 

Slough Wastewater Treatment Plant

Slough wastewater treatment plant is situtated to the south of Slough town and is bordered on one side by the M4 motorway. The plant treats the wastewater from a domestic population of approximately 150,000 and an industrial population equivalent of 150,000. Until 1996 the wastewater was treated primarily by a two stage biological process following preliminary and primary treatment. The settled wastewater was treated firstly in a non-nitrifying activated sludge plant which removed most of the biodegradable organic load and then in nitrifying trickling filters. The surplus activated sludge and the humus sludge were co-settled in the primary tanks with the incoming wastewater and the sludge produced was digested and dewatered in a centrifuge before being recycled to agricultural land as a soil conditioner. The centrate was returned to the inlet works for treatment.

Slough wastewater treatment plant has had a struvite problem for many years. The centrifugal pumps which return the centrate had to be removed and the struvite chipped from the pump casing every two weeks. This task was accomplished using a hammer and chisel and was extremely labour intensive. In an attempt to alleviate the problem a magnetic water softening device supplied by Scale Watcher was fitted to the pipework before the centrifuge. It was hoped that this would prevent the struvite from precipitating in the pump and reduce the manpower cost of operating the plant. At the time this solution did not appear to have any beneficial effect despite a number of attempts to overcome the problem.

In 1994 Thames Water agreed with the Environment Agency to treat a larger volume of wastewater before discharging to the storm overflow. This lead to a doubling of the flow to full treatment and it was clear that the existing combination of activated sludge and trickling filter would not be able to meet the consent. During the design of the plant it was foreseen that the high concentration of soluble Chemical Oxygen Demand (COD) in the settled wastewater would give rise to a very high oxygen demand at the beginning of the activated sludge plant which would be difficult to meet (3). Experience from other Thames Water treatment plants suggested it was likely that the resulting low levels of dissolved oxygen would lead to "bulking", i.e. poor settling sludge, and the plant would therefore be difficult to operate.

Although the consent for Slough does not require phosphorus removal it was decided to build the extension to the works as a biological nutrient removal plant. In this way the soluble COD would be utilised in the un-aerated parts of the plant before it reached the first aerobic section, so reducing the risk of bulking.

Biological Nutrient Removal Plant at Slough

The new extension at Slough is a three stage Bardenpho plant designed to treat the wastewater from a population equivalent of 165,000. The remaining flow is treated in part of the existing two-stage activated sludge, trickling filter plant. The overall plant is shown schematically in figure 1 and one aeration lane is shown in figure 2.
 

Figure 1. Schematic representation of Slough Wastewater Treatment Plant.
 

 
 

Figure 2. Schematic representation of the 3-stage Bardenpho plant at Slough.




The 3-stage configuration consists of an anaerobic/aerobic sequence to facilitate phosphorus uptake and an anoxic/aerobic recycle sequence to nitrify the ammoniacal nitrogen to nitrate and then for the nitrate to be reduced to nitrogen gas. The effectiveness of the plant to remove nitrogen and phosphorus can be seen is table 1.

Table 1. Performance of the 3-stage Bardenpho plant at Slough,
 

Parameter Settled Wastewater Final effluent
Suspended Solids mgl-1 150 8
BOD mgl-1 225 4
COD mgl-1 500 40
Ammoniacal N mgl-1 34 0.1
Total Oxidised Nitrogen mgl-1 - 5.6
Total Phosphorus mgl-1 10.3 0.7

From table one it is clear that the plant is producing a very good quality effluent in terms of organic pollutants and that nutrient removal is effective. The effluent contains only 17/% of the incoming nitrogen and 7% of the incoming phosphorus. The nitrogen that is removed in the aeration lane is mostly lost to the atmosphere as nitrogen gas as a result of denitrification in the anoxic zone. The phosphorus, in contrast, is not lost from the system but accumulated in the waste activated sludge to the extent that the mixed liquor may contain up to 5% phosphorus (4). When the sludge is exposed to an anaerobic environment, for example in anaerobic digestion, much of the phosphorus is released.

The fate of the released phosphorus is important. If the digested sludge is dewatered the released phosphorus will be returned to the incoming wastewater making good phosphorus removal more difficult. Also important is the fate of the phosphorus once it has been recycled to agricultural land. If the phosphorus is not absorbed it may be lost from the soil and could be leached into the groundwater and become a diffuse source of phosphorus. It would be advantageous if the phosphorus could be "fixed" in some way and removed from the wastewater treatment system or maintained in the sludge as a beneficial fertilizer.
 

Struvite Formation

Since the commissioning of the new activated sludge plant at Slough the problem of struvite formation has been exacerbated. Eight months after the new activated sludge plant was commissioned the pipework between the digesters and the digested sludge holding tank became blocked with an accumulation of small struvite crystals in a matrix of digested sludge solids. The pipework formed an inverted syphon which contained flowing sludge only when the digesters were being fed with raw sludge. The blockage occurred in the lowest horizontal and final vertical sections of the pipe which was filled with stationary sludge when the digesters were not being fed with raw sludge. The pipeline has been re-routed so that it can drain freely in the hope that this would reduce the opportunity for the struvite crystals to form. This modification appears to have been successful and the pipeline has been free from struvite since the changes were made.

Approximately one year after commissioning the new plant it became apparent that the piepline from the digested sludge holding tank to the centrifuge had become restricted and it was not possible to transfer sludge for dewatering. In the short term the problem was overcome by laying 800 m of temporary pipework to the centrifuges whilst the extent of the problem was investigated. Dismantling the pipework around the transfer pumps revealed a significant deposit of struvite. The 100 mm diameter pipework was reduced in bore to approximately 50 mm, which accounted for the sludge pumping problem. It was not easy to determine how much of the pipeline had a struvite deposit as much of the 800 m length was underground.

During laboratory scale tests it was discovered that the struvite could be dissolved in 10% sulphuric acid. The pipeline was then "cleaned" by filling it with acid and leaving it overnight. The following day the pipeline was flushed with effluent and the spent acid was allowed to flow into the large sludge lagoons. The pipeline around the pumps was found to be clean and the pipeline was recommissioned successfully. In order to prevent further stuvite deposition in the pipeline a number of options were considered before two magnetic "softening" devices, supplied by Lifescience Products Ltd. were fitted to the pipes before and after the pumps. The effectiveness of this measure is now being evaluated.

Confidence in this type of solution has been increased due to recent changes to the magnetic softener installed near to the centrate pumps. The latest modifications to this installation have reduced the incidence of pump removal and cleaning from once every two weeks to five weeks. If this kind of increaed time between cleaning could be obtained for the digested sludge pipeline the acid cleaning would be needed only every two or three years which may be acceptable to operational staff.
 

Struvite Recovery

It is clear that generating the conditions for the precipitation of struvite is not very difficult. The greater difficulty is ensuring that the struvite formation occurs in a location where it can be recovered economically. In order to establish the best location for struvite recovery a number of streams at Slough have been monitored for a short period. These streams are:-

  • The settled wastewater
  • The waste activated sludge
  • The waste activated sludge liquor from the belt presses
  • The digested sludge
  • The centrate from the digested sludge centrifuge

Samples from each stream were analysed for suspended solids, soluble phosphorus, total phosphorus and the flow has been measured, or estimated if accurate flow data was not available. These results are shown in table 2.

Table 2. Results of the analysis of potential phosphorus recovery sites at Slough.
 

Stream Suspended solids mgl-1  Soluble phosphorus  mgl-1 Total  phosphorus mgl-1 Ammonia mgl-1 Flow m3d-1  Phosphorus load kgd-1
Settled Wastewater 125 8.7 11.7 34 59,000 690
Waste activated sludge 9,000 190 350 50 700 245
Waste activated sludge liquor 250 190 200 50 600 120
Digested sludge 28,500 70 800 750 300 240
Centrate 270 80 110 750 250 28

The ideal location for the recovery of struvite requires that the flow should have a high concentration of soluble phosphorus and ammoniacal nitrogen, a low concentration of suspended solids and a relatively high phosphorus load.

The results in table 2 show that none of the locations sampled is ideal for the recovery of struvtite. The raw wastewater has low concentrations of phosphorus and ammonia, the waste activated sludge and the digested sludge have high suspended solids concentrations. The waste activated sludge liquor has a low ammoniacal nitrogen concentration and the centrate load is quite small. The high soluble phosphorus concentration of the waste activated sludge arises because the waste sludge is stored in a tank for about one day before being thickened.

It may be possible to combine two of the flows, the waste activated sludge liquor and the centrate to produce a more acceptable stream. The combined stream at Slough would have the composition and flow as shown in table 3.

Table 3. Flow and composition of a potential combination of the centrate with the waste activated sludge liquor.
 

Parameter Value
Flow 850 m3d-1
Suspended Solids 255 mgl-1
Soluble Phosphorus 160 mgl-1
Total Phosphorus 175 mgl-1
Ammoniacal nitrogen 255 mgl-1
Total Phosphorus Load 148 kgd-1

As the sampling programme was relatively short and some of the flows have had to be estimated an extended monitoring programme to verify these findings is required.
 

Conclusions

Phosphorus removal from the effluent stream at wastewater treatment plants is becoming more widespread. The options for removal lead to the phosphorus becoming concentrated in the sludge stream. It is important that the phosphorus in the sludge is held in such a way that it is not readily released and lost from the soil but is available to plants as a fertilizer. An alternative is to remove the phosphorus from the wastewater treatment plant in a managed way and for farmers to add phosphorus fertilizers to their land when it is required.

It is obvious from Thames Waterís experience at Slough that struvite is one phosphorus containing compound that is very easy to precipitate. The challenge is to maintain the phosphorus in solution until it reaches a suitable point for recovery. One such suitable point at Slough would be a combined stream formed from the centrate and the thickened waste activated sludge liquor. The second challenge is to design a simple struvite precipitation process which enables the phosphorus to be successfully recovered.

References

1 Council of the European Communities. Directive concerning the collection, treatment and discharge or urban wastewater and the discharge of waste water from certain industrial sectors. (91/271/EEC). Official Journal L 135/40, 1991.

2 Levin, G.V. and Shapiro, J. Metabolic uptake of phosphorus by wastewater organisms. J. Wat. Pollut. Control Fed. 1965, 37, p800.

3 Williams, S.C. and Beresford, J. The effect of anaerobic zone mixing on the performance of a three-stage Bardenpho plant. Paper accepted for the IAWQ Conference in Vancouver. June 1998.

4 Jardin, N and Popel, H.J. Consequences of phosphorus elimination for sludge production - a comparison between physical-chemical and enhanced biological phosphorus removal. In Chemical Water and Wastewater Treatment IV. Eds Hahn, H.H. Hoffmann, E and Odegaard, H. 1996.