Figure 20 Foreseen environmental impacts for enhanced coagulation
Related Figures (71)
Figure 1 presents the proposed framework in TRUST to follow during the optimization. This framework includes environmental and economic aspects that were not assessed in optimization roadmap developed within the FP7 TECHNEAU Project which considered only technical dimension. The present framework has been elaborated based on TRUST Master Framework (Deliverable 31.1) leaving aside social, governance, and assets dimensions. Figure 1 Optimization Framework Figure 2 Life cycle assessment framework according to the /LCD Handbook (2010) Coagulants (granulated aluminium sulphate) are then added before the flow is split in two lines consisting of flotation and settling. Both are followed by a typical media filtration. A UV disinfection system has been installed to provide a more robust treatment to give two hygienic barriers. Sodium hypochlorite and sodium hydroxide are then added to provide respectively a residual disinfectant (only for the time spent in the header tank, there is no residual chlorine in the distribution system) and corrosion control respectively. Figure 4 Process scheme - Stangaasen WTP Figure 5 Pictures from the flotation step (left), sedimentation (middle) and filtration (right) Table 2: Typical water quality data from Stangaasen WTP (Nov 2011) Figure 6 NOM fraction concentrations in raw and treated water samples from Stangaasen WTP As part of the treatment performance assessment, rapid NOM fractionation (Chow et al. 2004) and column-based BDOC analyses (Eikebrokk et al. 2007) were performed. The results are presented in Figure 6 to Figure 8 below. Figure 7 NOM fraction removal efficiencies at Stangaasen WTP Figure 8 BDOC concentration at Stangaasen WTP Figure 7 shows that the VHA and CHA fractions are effectively removed by coagulation (65- 80 %). The SHA fraction is also removed effectively (50-80%). In contrast, the NEU fraction concentration remains similar after coagulation treatment indicating poor removal of this particular NOM fraction. Thus, the percentage of NEU fraction to the total DOC increases from 5-8 % in the raw water to 17-20 % in treated water. In order to control NOM at acceptable levels supplementary treatment steps may thus be required when coagulating raw waters with significant NEU fraction concentrations. 3.2.3. Enhanced coagulation optimization Full-scale enhanced coagulation optimisation trials have been performed at Stangaasen WTP. The trials follow the optimisation procedures outlined in the EU-project TECHNEAU (www.techneau.eu) (Eikebrokk, 2009), and are based on a systematic variation of coagulation pH for a number of coagulant dose levels (Figure 9). In order to perform the optimisation trials, the WTP had to upgrade their equipment for coagulant dosing and pH- control. In addition, a UV disinfection system was installed in order to provide the two hygienic barriers required by Norwegian drinking water regulations. Figure9 Scheme of the full-scale enhanced coagulation optimisation trials at Stangaasen WTP aimed at identifying optimum coagulation conditions and the effects of filtration rate on performance (Eikebrokk et al, 2007) Figure 10 Illustration of the balance between safety and sustainable operation of enhanced coagulation facities In terms of water supply safety and sustainability, Figure 10 illustrates a major goal of the Table 4: Variation range of raw water quality and filter outlet water during optimization experiments Figure 11 to Figure 14 depict the behaviour of selected water quality parameters during the optimization experiments. Figures 11-14 show that the barrier indicator requirements for colour were met under all tested combinations of coagulant dose and pH. The residual Al requirement was also met except for the lowest pH values (pH 5.8 and 5.9) with lowest coagulant doses (40 and 47 mc ALG/L). Regarding the turbidity parameter the interpretation of the results is more comple» due to variations in raw water quality. The barrier indicator requirement of 0.2 NTU is met for several combinations of coagulant dose and pH, but here it seems that the combinatior of low coagulant doses and low pH gives values that exceed the barrier indicator requirement. Figures 11-14 show that the barrier indicator requirements for colour were met under all Figure 11 Turbidity variation during the optimization at Stangasen WTP 2011-2014 (Raw water turbidity varied between 1-3 NTU) Figure 13 TOC variation during the optimization at Stangasen WTP 2011-2014 (Raw water TOC varied between TOC 6-8.5 mg/L) Figure 14 Residual Al during the optimization at Stangdsen WTP 2011-2014 (Raw water Al< 0.3 mg/L) able 6: Water treatment efficiencies of the alternatives considered produced (1 m*) that meet the legal requirements on water quality for human consumption. Since previous studies have shown that construction and dismantlement phases of a DWTP are negligible with respect to the environmental impact (Raluy et al., 2005; Stokes and Horvath, 2006; Vince et al., 2008) and because the DWTP at Oppegard already exists, this LCA study focuses only on the production phase of the waterworks life cycle. The boundaries (Figure 15) of the system are selected to include all the activities that are an essential part of the current treatment processes. The functional unit is one cubic meter of drinking water produced (1 m?) that meet the legal requirements on water quality for human consumption. Life Cycle Inventory Regarding the carbon foot print (Figure 18), electricity has the largest contribution followed by the sodium hydroxide, the coagulant, transport, sludge and sodium hypochlorite. The type of energy conversion process may have a large effect on the overall assessment. Indeed, depending on the energy conversion process, the production of electricity may indeed use fossil fuels. Their combustion generates CO. emissions, thus leading to high potential impacts on climate change. Between electricity supply from renewable sources, nuclear energy or coal power plants, electricity production impacts can be multiplied by as type of energy conversion process may have a large effect on the overall assessment. Figure 16 Environmental profile-Contribution of WTP operation to the environmental impacts Figure 18 Contribution to the carbon foot-print from processes (as kg CO2 eg/nr, ) As illustrated in Figure 19, the local electricity production mix and transport have a tremendous influence on the impacts of water treatment processes. For the same treatment plant, by using the Norwegian electricity mix (99.06% Hydropower) the impact of energy production could be reduced by a factor of 2.5 and 3 compared to the Scottish electricity mix (30.1% hard coal, 29.9% nuclear, 18% natural gas, 2% oil, 20% renewable energy) and the Dutch electricity mix (59.4% natural gas, 24.9% hard coal, 4.1% nuclear, 11.6% other) respectively. Therefore, for the same treatment plant the effort involved in optimization of Table 9: Relative sustainability measures for the six alternatives As seen from Table 9, alternative B2 -iron chloride sulphate - turns out to be the best option with respect to holistic sustainability (9.7% overall improvement) — costs rise by 3.4%, environmental impacts decrease by 23.3% and water quality improves by 11.7%, with respect to the status quo situation. If the same coagulant that is being used currently is to be retained, in order to comply with the barrier indicator values mentioned earlier in the report, a minor optimization potential was identified. These results are, however, contrary to the findings from similar optimization experiments at most enhanced coagulation facilities in Norway, where potential reductions ranging from 20% - 40% in coagulant dose levels - and in the corresponding sludge production rates - were identified, thus allowing a more sustainable operation while still maintaining safety and barrier efficiency and stability (Eikebrokk, 2009). The process configuration of the Loenderveen pretreatment and the Weesperkarspel treatment facility is presented in Figure 23. The pretreatment comprises the intake of Bethune polder water (or Amsterdam-Rhine canal water as back-up facility), coagulation, lake water reservoir and rapid sand filtration. The ‘test reservoir’, shown in the picture below, is bypassed since 2009. The bypass led to a tenfold reduction of bacteria occurring by the feces of residing birds on the reservoir. The Weesperkarspel treatment comprises ozonation, fluidized bed softening, biological activated carbon filtration and slow sand filtration. Figure 23 Process configuration of Loenderveen pretreatment and Weesperkarspel treatment facilities [able 10: Typical water quality data from Weesperkarspel WTP (average 2011) Data on the consumption of energy during 2010, 2011 and 2012 show that the total energy consumption for water pumping and treatment is 0.406 kWh/m? of water produced. 3.3.3. Procedures to increase sustainability assessment (LCA) studies for investment decision support. In the period 2010-2013 Waternet performed about ten LCA studies, both on wastewater projects and drinking wate! projects. Prior to 2010, LCA studies were performed as well. Some of the LCA results were published by Barrios (2008), Tapia (2008) and Klaversma (2013). Barrios et al. (2008) founc that the impact of the operational stages of the pretreatment of Loenderveen was 2.6: mPt/m?, and the impact of the final treatment of Weesperkarspel was 5.98 mPt/m*. The total environmental impact is found to be 8.61 mPt/m?, Table 11 presents the contributior of each process unit to the total environmental impact (Barrios, 2008). The annua production of Weesperkarspel is approximately 26 Mm?/y, hence the total environmenta impact is almost 225000 Pt/year. This section provides a summary of the research projects and implemented investment projects with an impact on sustainability. The summary is presented in Table 12. This table provides a summary of the quantified contribution of several investments to a more sustainable water supply. A more complete description and analysis of this research and subsequent implementations was provided in TRUST Deliverable 41.1. The research and implementation was conducted by Waternet, and for this purpose mostly related to the Weesperkarspel facility and the pre-treatment at Loenderveen. In many cases, the research results are accompanied with LCA analyses and feasibility analyses in order to justify implementation or investment. Table 12: Quantification of research and implementation of improved sustainable water supply projects at Weesperkarspel and Loenderveen Waternet tested fluidized ion exchange (FIX) and magnetic ion exchange MIEX) in the pilot plant at Weesperkarspel, at various positions in the process configuration. Tests were performed with IEX prior to ozonation and IEX after slow sand fi distribution). Water quality analysis results and process operation [tration (prior to parameters were integrated in a conceptual design, and the addition of cost information has led to a business case calculation. The expected variation of DOC across the treatment process is summarized in Table 14, these data were used as an input for the cost evaluations ca 2013). Lculations (Grefte, Figure 25 Process scheme - Tullich WTWs Of interest, the water source has a high charge associated with the organic compounds in the sample. This is identified by the very negative zeta potential in the raw water. The charge is largely removed by the oxidation in the pre-ozone stage. However, after GAC the organic material is highly charged again, indicating that through biological transformation or release of biological DOC, the organic matter becomes negatively functionalised. The secondary ozone is able to partially remove this charge, but not to any great extent. The HPSEC traces (Figure 26) show the change in MW distribution of organic compounds ir the water that are able to absorb UV at 254 nm. On these figures, short elution times (lef hand side) represent high MW organic compounds and long elution times (right hand side are representative of low MW compounds. The higher the spike of the curve is, the highe the concentration of the organics for a specific MW. Because a large portion of the UV254 absorbing compounds are removed after the first ozone stage but a significant portion o DOC remains in the water, care must be taken in not over-interpreting the results for Tullict WTWs. The traces show that the bulk of the high MW organics are degraded or remover after the first ozonation stage, leaving a residual of medium to low MW DOC in the sampl that is not removed to any degree by the subsequent treatment processes. Figure 27 Fractionation on a proportion basis at Tullich WTWS, July 2012 Figure 28 Fractionation on a mass basis at Tullich WTWs, July 2012 The reduction in THM-FP on a ug.L” basis broadly agrees with the removal of DOC across the first ozone stage. This indicates that there is no change in the reactivity of the DOC as it was predominantly transformed from HPO to HPI material, which was a surprising result. This was confirmed by looking at the similar normalised THM values. This observation is worthy of further investigation because it indicates that a highly reactive HPI fraction of DOC is formed after ozonation. Across the GAC treatment stage, DOC reactivity is reduced and this stays reduced after the post-ozone process. The DBP-FP of the water shows the absolute concentration of DBPs that will form by over- dosing with chlorine and leaving the sample for seven days. Here we have measured the tri- halomethane formation potential (THM-FP) on a normalised tig THMs.mg DOC” basis and then calculated a yg.L™ concentration based on the DOC of the sample. The results show that the DOC in the water has high THM-FP indicating highly reactive organic compounds in the water: The WTWS treats a typical flow of 75 MLD, serving a population of 200 000 inhabitants. The flow sheet for the WTWs is shown in Figure 29. The site receives water from two separate locations: Derclach Loch and Bradan Loch. The site is a typical coagulation/flocculation, clarification, filtration site with the added addition of an ozonation stage before filtration. The site uses alum as the primary coagulant. The primary disinfectant for the treated water is chlorine due to the regulatory requirement of residual disinfection for the water in distribution. Results from Bradan WTWs able 18: Water quality data for Bradan WTWs, August 2012 The charge density for the Bradan raw water was 4.8 meq.gDOC™. For Derclach raw water, the charge density was 6.4 meg.gDOC”*. The results indicate that these sources were higher than typical with respect to their charge, indicating that the coagulant dose might not be high enough to match the charge demand. Further increases in the coagulant dose may help reduce the residual DOC given that the current zeta potential of coagulated water was <-10 mV, which is on the cusp of acceptable charge neutralisation. The balance of organic material in the raw waters was dominated by transphilic and hydrophilic NOM, accounting for >80% in Bradan and Derclach water sources (Table 19). The combined water appeared to have a much greater fraction of transphilic material in it indicating either a transformation of organic matter in the combined sample or a mistake in the analysis. Regardless of this, it was no surprise to see that coagulation resulted in a high residual of DOC after coagulation due to the high concentration of DOC that is more difficult to coagulate (/e the TPI and HPI in Figure 31 and Figure 32). However, it was surprising to see that the HPO component of the treated water did not change after coagulation indicating that further optimisation work may be necessary in order to reduce the treated water DOC further. Raw (Bradon) Raw (Derclach) Raw (Combined) Primary filter Secondary filter Table 21: Water quality data for Amlaird WTWs, August 2012 residual of low MW weight UV254 absorbing material was left after coagulation treatment. The charge density for Amlaird raw water was 4.49 meq.gDOC™. The results indicate that the source water was higher than typical with respect to its charge, but coagulation and subsequent treatment was effective at this site. The analysis of the MW distribution of the raw water showed two large peaks at 11 and 15 minutes elution time (Figure 34). A small residual of low MW weight UV254 absorbing material was left after coagulation treatment. Figure 35 Fractionation on a proportional basis at Amlaird WTWs, August 2012 Figure 36 Fractionation on a proportional basis at Amlaird WTWsS, August 2012 Table 23: Water quality data for Amlaird WTWs, August 2012 Filter performance status The organic spikes were dosed into Amlaird source water, which, as shown previously, is a relatively high DOC source with predominantly hydrophobic organic compounds present in the water. Preliminary tests on raw water identified the coagulation conditions required in order to give optimum removal of both UV2s4 absorbance and DOC from water that did not contain any spiked compounds (Table 25). The coagulant dose required was 20 mg/L as Fe. The MW distribution of UV absorbing organic compounds around this dose range can be seen in Figure 38, showing that coagulation was highly effective in removing these organic materials. Further increases in coagulant dose did not improve NOM removal. Figure 43 Change in THM-FP for treated water spiked oxalic acid and tannic acid Raw water with 5 mg/L oxalic acid spike The results from the floc profiling work shows that firstly flocs formed in the presence of compounds similar to aspartic acid are likely to be smaller and break to smaller sizes on exposure to high shear rates. The presence of tannic acid on the other hand forms flocs that are larger and hence could be better removed from clarification units than for other types of organic compound that may dominate a water source. Secondly, the presence of organic materials similar to oxalic acid will slow the growth of flocs such that if inadequate flocculation time is given in these types of water a poor floc might be formed. Raw water with 5 mg/L tannic acid spike Raw water with 5 mg/L tannic acid spike Figure 46 Floc breakage profiles for raw water with organic spikes coagulated using 25 mg/L as Fe coagulant Figure 45 Floc strength profiles for raw water with organic spikes coagulated using 20 mg/L as Fe coagulant Figure 47 Floc strength profiles for raw water with organic spikes coagulated using 25 mg/L as Fe coagulant Floc breakage profiles for spikes Figure 48 Floc strength profiles for organic spikes at 15 ana 20 mg/L DOC respectively, both using 20 mg/L as Fe coagulant. Note that oxalic acid did not produce a floc, so was not included in the figure
