By D de Haas, G Appleby, G Charakosa and N Dinesh.
First published in Water e-Journal Vol 3 No 1 2018.
The results of the latest round of energy use benchmarking on wastewater treatment plants (WWTP) conducted for 2015-16 by the Water Services Association of Australia (WSAA) and Intelligent Water Network (IWN) are presented. In total, 245 WWTPs were covered, mostly in Australia, with two plants in Auckland (New Zealand) included. This represents a large increase (by 103) in the number of WWTPs, compared with the previous WSAA benchmarking round in 2013-14. Most of the WWTPs added in the latest round were extended aeration-type activated sludge plants and aerated lagoons. These types tended to be less energy efficient. Nevertheless, the results show significant improvements in energy efficiency for some plants.
Within the subset of plants that participated in both the latest and previous rounds, slightly more than half showed some improvement in energy efficiency, and just over one third showed an improvement of more than 10%. The primary benchmark applied was specific energy use per unit of equivalent population per year (based on raw influent loads). A secondary benchmark of energy self-supply was applied for plants with on-site renewable energy, predominantly co-generation from biogas. Alternative benchmarks of specific energy use per unit of pollutant load removed were also investigated, namely Total Nitrogen (N), Chemical or Biochemical Oxygen Demand (COD or BOD).
Energy use accounts for a large portion of the operating cost and greenhouse gas (GHG) emissions profile of a typical wastewater treatment plant (WWTP). Understanding specific energy use of a WWTP plant, and benchmarking it against other similar plants, is an important step to reducing energy-related costs and GHG emissions. However, energy use data reported in the literature for WWTPs is often fragmented or expressed in different specific units.
In 2013-14, the Water Services Association of Australia (WSAA) conducted a first round of energy benchmarking for 142 WWTPs across Australia. A second-round benchmarking exercise was performed by WSAA and the Intelligent Water Network (IWN) for the 2015-16 data period, which included a total of 245 WWTPs (243 plants in Australia and two plants in Auckland, New Zealand). The purpose of the exercise was to benchmark specific energy consumption of WWTPs, compare these results to those from the previous benchmarking round, establish a baseline energy performance result for newly participating WWTPs, and assess new energy-related benchmarks.
This paper summarises the findings of the WSAAIWN 2015-16 benchmarking round. It represents the consolidation of outputs from a large database of discrete and time-series data submitted by 31 water utilities that participated in the study (WSAA, 2017). Of these, 17 utilities had also participated in the previous (2013-14) benchmarking round. Nearly 50% of the WWTPs in the 2015-16 round reported here (121 out of 245 plants) were included in the first round as well.
The approach to WWTP energy benchmarking selected for this study (WSAA, 2017) referenced an approach developed and applied in several European countries over the last two decades, including Switzerland, Germany, and Austria. In particular, this study’s approach followed current best practice from published manuals, research reports and reviews from Germany (LfU, 1998; Baumann and Roth, 2008; Haberkern et al., 2008; DWA, 2012; Baumann et al., 2014; DWA, 2015). It is also in line with the recommendations of the Water Environment Research Foundation (USA) (Crawford, 2010), and application, for example, by SA Water in South Australia (Krampe, 2013; Steele et al ., 2013; Corena et al ., 2015; Malekizadeh, 2016).
The data collected from each WWTP covered a period between July 2015 and June 2016, with most utilities providing data covering this entire period. The data collected included:
- Details of plant type and processes used;
- Electrical energy (imported and exported);
- Raw influent flow and quality;
- Effluent flow and quality;
- Other imported waste not included in the raw influent, and whether or not this waste is anaerobically co-digested;
- On-site energy generated, including that from biogas or other sources, such as wind or solar;
- Details around influent and/or effluent pumping (flow, head and sub-metered energy use), where relevant and available; and
- Other energy sub-metering data, if available.
The methodology applied calculated WWTP energy use in terms of average annual electrical energy use (kWh) per equivalent person (EP), derived from influent wastewater loading and expressed in units of kWh per EP per year (kWh/[EP.y]) as the primary benchmark. The calculations of EP for each plant were based on influent raw wastewater load, taking into account both influent organics (COD or BOD) and nitrogen, where these data were available. The data submitted was filtered to exclude outliers, using a commonly applied statistical method (McGill et al., 1978; Frigge et al., 1989). The wastewater data (flow and concentration) as well as energy use were assumed to be log-normally distributed.
For the majority of plants included in this study (242 out of 245, by number), sufficient data was available to calculate adopted EP as an average from influent organic and nitrogen loads. For the remainder, the adopted average EP was calculated from raw influent organic load only (COD or BOD, whichever was available).
Due to common use elsewhere, average flow-based energy use (kWh per megalitre wastewater treated) was also calculated and reported.
For plants with on-site energy production, a secondary benchmark of percent electrical energy self-supply (%Ess) was applied.
Where possible, calculations of specific energy use were also performed for additional metrics, with a view to developing additional benchmarks in future, related to effluent quality and treatment performance, namely:
- Total Nitrogen (N) removed: kWh per kg N removed,
- COD removed: kWh per kg COD removed, and/or
- BOD removed: kWh per kg BOD removed.
Five plant types (Types 1 to 5) were relevant to the WWTPs involved in this study and categorised according to treatment processes, with some sub-types defined. A high-level summary of the relevant five types is given in Table 1. Refer to Supplementary Information for more information.
Plant size classes
Five plant-class sizes (SC) were set according to various ranges of EP. These definitions were based on similar definitions applied in German WWTP energy benchmarking manuals (inter alia DWA, 2012; Baumann et al., 2014), as shown in Table 2. Refer to Supplementary Information for more information.
Energy benchmarks were set as “Target” or “Guide” Values, based on German energy manuals (Haberkern et al., 2008; DWA, 2012; and Baumann et al., 2014). The Target and Guide Values are indicative of “best practice” and “typical” (or “average”) performance respectively. The values applied vary according to plant Type and Size Class (refer to Figure 1), and typically decrease with increasing plant size, due to economies of scale.
For specific energy-intensive processes, an “energy supplement” was provided in the benchmarks. This represents the expected additional energy use from these processes and is added to the Target or Guide Values for a given plant. For example, high pumping requirements (nominally >4 metres of head), specifically for influent wastewater or treated effluent, was present in many cases.
For such cases, pressure (head) data was additionally requested to enable appropriate energy supplements for pumping to be applied. However, the relevant pressure information was only submitted for a small number of plants, which meant that the energy supplements for pumping could only be applied to those plants.
Refer to Supplementary Information and Figure 1 below for more information.
Overall industry benchmarks
Table 3, Table 4 and Table 5 provide high-level summaries of data collected from this study. Supplementary Information for this paper provides graphical summaries of the data for all plants, ranked in order of increasing energy use according to the primary benchmark of kWh/(EPy), grouped by Plant Type and Size Class. The guide and target values applied were reviewed for this benchmarking round and revised in some cases, compared with the previous round.
The average Target and Guide Values listed in Table 3, Table 4 and Table 5 include the energy supplements (where applicable) for the relevant WWTPs represented within each Type. Therefore, it is important to note that the Guide and Target Values tabulated (Table 3, Table 4 and Table 5) are weighted averages, based on the number and range of plants falling into each category (i.e. a combination of Type, Size Class and Supplements, where applicable).
Refer to Supplementary Information (Table 7) for a summary of the change in performance on primary benchmark for plants that participated in both this benchmarking round (2015-16) and the previous round (2013-14), broken down by Type and Size Class.
Ranking all plants by specific energy use
Of the WWTPs in this study, the ‘top 100’ WWTPs were assessed by ranking in ascending order of primary benchmark result, i.e. average specific energy use (kWh/ [EP.year]). All Plant Types (1 to 5) and all Size Classes (SC1 to SC5) were included in this ranking.
It was found that the ‘top 100’ plants had an average specific energy use of 50 kWh/(EP.y) or less. Furthermore, below approximately 15 kWh/(EP.y), only Type 4 (trickling filter types) and Type 5 (lagoon types) featured. The majority of the Type 1 plants (19 out of 22 in number) were placed in the ‘top 100’ list and had an average specific energy use ranging 15 to 49 kWh/(EP.y). There were three plants classified as Type 1 by default that lack secondary treatment and discharge via ocean outfalls. These three plants were ranked in first, second and fourth-lowest, with an average specific energy use ranging 15 to 25 kWh/(EP.y), and might be considered to be a type of their own in future benchmarking rounds.
Similarly, the ‘middle 100’ WWTPs assessed as part of this study had an average specific energy use in the range approximately 50 to 110 kWh/(EP.y). Although there are exceptions, the predominant plant types in the ‘middle 100’ list were Types 2 and 3, particularly the latter, which are more dependent on aeration for treatment and hence tend to use more energy.
Finally, the ‘bottom 45’ WWTPs assessed were predominantly plant Types 3 and 5, again because both include examples of plants with relatively high energy use i.e. extended aeration activated sludge processes (Type 3) or aerated lagoons (Sub-type 5.1). Some small membrane bioreactor plants (MBR, Sub-type 3.4) with relatively inefficient aeration systems, were amongst the highest specific energy consumers in this group.
General comments on specific energy use
Many local or plant-related factors underlie energy use on a given WWTP and constrain its performance against industry benchmarks. These might include plant size; type of process; operating energy efficiency of the installed equipment; pumping and mixing requirements; and water quality drivers (e.g. nutrient removal, disinfection and water recycling).
Nevertheless, looking broadly at the results from this study, compared with the primary benchmarks for each Type and Size Class (Figure 1 and Supplementary Information), it is possible to make general comments about the energy efficiency performance of Australian WWTPs, as summarised in Table 6.
For Type 1 plants, percent electrical energy self-supply (%Ess) performance was calculated and the results are summarised in Figure 2. The Guide Value is 60% Ess, with a Target Value of 100%.
Of the 22 Type 1 plants in review for this benchmark, none of the plants achieved the Target Value of 100% Ess on average. One plant achieved an average >90% Ess, which is an excellent result. A further six plants achieved better than the Guide Value (i.e. >60 % Ess) on average.
One plant reported 0% Ess. That plant was being upgraded during the review period. In future, it will include energy self-supply, but these facilities were not commissioned during the 2015-16 reporting period.
This study examined three new benchmarks that might be used as alternatives to, or in conjunction with, the primary benchmark. These are based on specific energy use relative to pollutant load removed, namely:
- Total N load removed: kWh/kgN removed;
- COD removed: kWh/kg COD removed; and/or
- BOD removed: kWh/kg BOD removed.
Where available, data was collected for the relevant influent and effluent parameters (flow and concentration) to enable the pollutant load removed to be calculated for each of the WWTPs in this study. Most plants reported either COD or BOD measurements for influent and/or effluent, but few plants reported both parameters. Where matching (or paired) measurements of influent and effluent COD or BOD were not reported, it was not possible to calculate pollutant load removed and hence the alternative benchmark.
Similarly, not all plants reported influent and effluent Total N. Where possible, Total Kjeldahl Nitrogen (TKN) was taken as a close approximation of raw influent Total N. Similarly, where effluent Oxidised N (Nitrate plus Nitrite) and Ammonia was reported, the sum of these (Total Inorganic N) was taken as an approximation of effluent Total N, in the absence of other data.
Overall, out of the total number of 245 WWTPs in this review, the reported data enabled the following alternative benchmarks for specific energy use to be calculated:
- Per unit Total N load removed (kWh/kg N removed): for 242 WWTPs
- Per unit COD load removed (kWh/kg COD removed): for 70 WWTPs
- Per unit BOD load removed (kWh/kg BOD removed): for 171 WWTPs
The summary results are given in Table 7. Refer to Supplementary Information for more detail.
Results for alternative benchmarks
A more detailed analysis of the results for the alternative benchmarks (refer to Supplementary Information) showed that the alternative benchmarks based on removal of organic load (expressed as kWh/kgCODrem or kWh/kgBODrem) were strongly correlated with the primary benchmark (kWh/[EP.y]). This can be intuitively understood to be because most WWTPs (particularly the types with activated sludge processes in Types 1, 2, 3) achieve nearly complete removal of COD and/or BOD (typically >93% removal on average – refer to Table 7).
However, the alternative benchmark based on nitrogen removal (expressed as kWh/kgNrem) was less strongly correlated with the primary benchmark (kWh/[EP.y]).
The underlying reason is that nitrogen removal is more variable for Australian WWTPs across the types represented in this review. That is evident even from the average percent N removal across all plants within a given type (refer to Table 7). As might be expected, extended aeration activated sludge plants (Type 3), which were strongly represented by numbers in this review, are typically designed for nutrient removal, giving an average of 90% N removal (Table 7). Although represented by a smaller number of plants, the Type 2 plants achieved an average of 80% N removal, whereas a similar number of Type 1 plants showed an average N removal of only 57%, or 71% excluding the three nominal Type 1 plants that lack secondary (activated sludge) processes and achieve little or no N removal (i.e. discharge to the ocean).
Similarly, Type 4 and 5 plants are not typically designed for advanced nitrogen removal, although some N removal occurs through treatment processes associated with removal of organic material and/or suspended solids. Table 7 shows that these plant types achieved on average 56 to 76% N removal.
It is recommended for future benchmarking studies that an additional benchmark of specific energy use based on nitrogen removal (kWh/kgNrem) be included. Further inclusion of the alternative benchmarks based on COD or BOD removal (kWh/kgCODrem or kWh/kgBODrem) could be optional but are unlikely to add significant value to understanding or improving energy efficiency for WWTPs, compared with the existing primary benchmark (kWh/[EP.y]).
Proposed interim benchmark Guide and Target Values for alternative benchmarks
Based on available data from this review, Table 8 gives proposed interim benchmark Guide and Target Values for the alternative benchmarks based on removal of pollutants. Given the limitations of the underlying datasets, as far as possible, the Guide and Target Values have been based around the 50th and 10th percentiles calculated from available data.
The interim values in Table 8 may be revised in future benchmarking rounds, subject to matching datasets for the relevant influent and effluent quality parameters being available for a larger number of WWTPs. A further breakdown of guide and target values by size class within plant type may also be possible in future, subject to data availability and industry acceptance of the alternative benchmarks.
Pumping energy requirements
This study included data collection for plants that were nominated by water utilities as having high pumping requirements (defined as >4 metres head) for influent and/or effluent. In addition to flow, more detailed data (i.e. time-series data, or at least single average estimates) of system pumping head(s) were requested for the relevant pump(s), where available. The benchmarking methodology for pumping requires data (measured or estimated) for pumped flow and head (pressure) to be reported in order that the benchmark Target and Guide Values for a site receive the relevant energy supplements (refer to Methodology and Supplementary Information).
Of the total number of plants reviewed (245 no. of all Types), the pumping data received can be summarised as follows:
- 57 plants (23%) reported high influent pumping requirements. Of these plants, sufficient data were provided to apply the relevant benchmark energy supplement (S3.1) at only six sites;
- 77 plants (31%) reported high effluent pumping requirements. Of these plants, sufficient data were provided to apply the relevant benchmark energy supplement (S3.2) at only eight sites; and
- 20 plants (8%) reported both high influent and high effluent pumping requirements. Of these plants, sufficient data were provided to apply both benchmark energy supplements (S3.1 and S3.2) at only three sites.
Despite the relatively small subsets of plants (see above) that reported sufficient data to benchmark pumping energy efficiency, the contribution from influent and/or effluent pumping to total plant average specific energy use was significant (i.e. 12 to 54% of total plant energy) – refer to Table 9.
The study findings suggest that the minority of plants have high influent and/or effluent pumping requirements directly linked to total plant energy records.
Furthermore, very few utilities have pump pressure data (measured or estimated) available for their WWTPs. The above results contrast with the results from the previous benchmarking round (WSAA, 2013-14) in which the majority of plants (117 out of 142 no. total or 82%) reported high pumping features (>4 m head). However, in the previous benchmarking round, no distinction was made between influent, effluent and internal plant high pumping requirements.
In this round, influent and effluent pumping was separately identified for benchmarking purposes, using the energy supplements (see above). For this round, plant internal pumping was included in optional sub-metering data requested, but regarded as part of the treatment process for energy benchmarking purposes. Therefore, plant internal pumping was assumed to be incorporated in the base benchmark Target and Guide values nominated (Figure 1).
Taken as whole, the pumping data from this round suggests that utilities will need to focus more on collecting the relevant data records for plants that have high influent and/or effluent pumping requirements in order to improve the value of energy benchmarking and to enable more meaningful comparisons between plants.
The number of water utilities and WWTPs that participated in the latest (2015-16) WSAA energy benchmarking round was nearly double that of the previous round (2013-14). This reflects the growing emphasis placed on better understanding and improving energy efficiency in wastewater systems. The latest round had the largest increase in numbers for plants of the extended aeration activated sludge type (Type 3, various sub-types) and aerated lagoon plant type (Sub-type 5.1).
These types of plant tend to be in the small to medium size classes (i.e. <100,000 EP and typically <1000 to 10,000 EP) and also less energy efficient, due to both diminished economies of scale and a greater reliance on aeration for treatment. Therefore, it is not surprising that the average results across all plants for this round showed an increase in specific energy use, relative to the previous round and the benchmarks applied. It reflects the challenge faced by many water utilities to improve energy efficiency, given the existing decentralised wastewater collection and treatment systems serving many suburban and regional metropolitan areas across Australia.
Based on specific energy use amongst the WWTPs across all types in the largest size class (SC5, >100,000 EP), this study showed results that were on average 12% lower in this benchmarking round (40 plants) than the previous round (35 plants). Similarly, Type 1 and Type 2 plants, which include primary treatment and anaerobic digestion and hence are less reliant on aeration for treatment, showed an improvement in energy efficiency on average. Irrespective of energy self-supply, the specific energy use, was 10% lower in this round for Type 1 (22 plants vs. 18 in the previous round), and 2% lower for Type 2 (24 plants vs. 16 in the previous round). These types of plant tend to be more feasible at the larger scale.
This is reflected, for example, in this study where all the Type 1 plants (with anaerobic onsite cogeneration from biogas) were in the two largest size classes (SC4 and SC5) and the majority were in the largest class (i.e. >100,000 EP). It confirms that some of the best opportunities for improving energy efficiency lie in centralised WWTPs with plant configurations that include primary treatment and anaerobic processes. Typically, the trade-offs around energy efficiency are capital investment for centralised wastewater collection and treatment processes (e.g. aeration system, anaerobic digesters and their ancillary equipment), as well as effluent quality (e.g. nitrogen removal requirements).
This study investigated alternative benchmark metrics that express energy use per unit mass of pollutant (organic material or nitrogen) removed. It found that metrics based on removal of organics (kWh/ kg removed as COD or BOD) offered little or no advantage over the primary benchmark adopted here of kWh/(EP.y), where EP is derived from raw wastewater loading. A metric such as kWh/ kg BOD removed would have meaning to some practitioners. However, BOD (or COD) removal is nearly always virtually complete (>80% and typically around 90%) for most WWTPs, which results in the linear correlation with influent load population equivalents being relatively strong (R2 >0.9). A linear relationship can therefore be used with reasonable confidence to convert between kWh/(EP.y) and kWh/kg BOD (or COD) removal.
Retaining the primary benchmark of kWh/(EP.y) in Australia/ New Zealand has the advantage of aligning with similar studies in some European countries that extend back over one to two decades.
However, N removal is more variable (ranging approximately 20 to 98% in this study, but typically >50%). It is dependent on WWTP type, configuration, loading and effluent TN target. When comparing plants with different N removal requirements, applying alternative metrics for specific energy use that include kWh/kgN removed might be more useful. Nevertheless, it is worth noting that most WWTPs in Australia are required to nitrify (oxidise ammonia), and hence at least partial denitrification is typically applied in activated sludge plants in order to recover oxidised N (oxygen equivalents) and hence reduce aeration costs. Therefore, it is not surprising that the results of this study showed a reasonable linear correlation (R2 = 0.75) between the primary benchmark of kWh/(EP.y) and kWh/kgN removed.
Aeration is typically considered to be the single largest energy user on most WWTPs, which was confirmed in this study (sub-metering data is not discussed on this paper – refer to WSAA, 2017). Pumping energy is also a significant contributor to total plant energy use, as the limited data from this study for influent and/or effluent pumping has confirmed. More attention to collecting the necessary data (including pressure or head) to include pump efficiency in overall plant energy benchmarks is recommended. Alternatively, benchmarking influent or effluent pump energy efficiency could be segregated from that of the treatment process itself for a WWTP. This requires separate sub-metering of pump energy use, as well as flow and pressure (or head) data.
More than one hundred additional WWTPs were included in the latest (2015-16) energy efficiency benchmarking study, compared with the previous study (in 2013-14), bringing the latest round total number to 245. Around half the plants by number (121) that participated in the latest round, also participated in the previous round (all within Australia). Within that sub-set, slightly more than half showed some improvement in energy efficiency, and slightly over one-third showed an improvement of more than 10%.
These are significant improvements and reflect the efforts of water utilities in Australia to improve energy efficiency. It also demonstrates the value of benchmarking as a monitoring tool for continuous improvement. At the other end of the range, more than 120 WWTPs were considered to be ‘below average’ in performance and not energy efficient. For these plants, there remains significant opportunity to improve energy efficiency and/or lower energy use.
A number of areas for future improvement can be identified from this study. These include data collection; data handling and sharing for benchmarking purposes, particularly in the context of ‘big data’; data validation and security around shared information; and future legacy. For example, there is a need for continued cooperative efforts to agree on additional and/or future benchmarks, both in terms of metrics and Target and Guide Values, within the Australian/New Zealand wastewater treatment context. This study also highlighted the need particularly to better systematise the distinction between wastewater pumping vs. treatment, which presents challenges to the way in which energy and related parameters (e.g. flow, concentration, pressure) are measured or sub-metered, and reported for benchmarking purposes.
About the authors
Dr David de Haas is a Principal Professional at GHD Pty Ltd with over thirty years of experience in Wastewater Treatment, including research, planning, design and operations. He has specialised in biological nutrient removal systems and greenhouse gas emissions from wastewater collection and treatment systems. His recent experience in Australia includes energy efficiency and benchmarking wastewater treatment plants.
Greg Appleby, Greg is a Senior Resource Management Advisor in Sydney Water’s Resources and Environment Team. His role focusses on identifying and developing renewable energy and energy efficiency projects as well as incorporating energy cost reductions into Sydney Water’s planning and design processes. Greg also has extensive experience across wastewater treatment operations.
George Charakos, George Charakos is an Energy Advisor in Melbourne Water’s Operational Technology Team, with 10 years’ experience in energy management. His role includes aspects related to energy auditing, energy efficiency, energy performance analysis, industry benchmarking, renewable energy, energy trading, and, energy and greenhouse strategy development/implementation.
Nirmala Dinesh, Nirmala Dinesh is a Principal Process Engineer (Wastewater) at SA Water. She is responsible for strategic planning and development of treatment strategies for fit for purpose effluent quality, biosolids, energy use and odour management across 25 wastewater treatment plants. Over the last 16 years, she has worked on a number of strategic projects and treatment plant upgrades for SA Water. Her focus has been on projects relating to energy benchmarking, development of business cases, plant optimisation and implementation of energy efficient measures including advanced aeration control, selection of energy efficient devices and increasing self-sufficiency through co-digestion of high strength waste.
Baumann, P. and Roth, M. (2008) Senkung des Stromverbrauchs auf Kläranlagen, Leitfaden für das Betriebspersonal (Reduction of the Energy Consumption of Wastewater Treatment Plants – Manual for the Operation of WWTPs), Heft 4, 2. Auflage (Issue 4, 2nd Edition), DWA Landesverband Baden- Württemberg, Stuttgart (in German).
Baumann, P., Maurer, P and Roth, M. (2014) Senkung des Stromverbrauchs auf Kläranlagen, Leitfaden für das Betriebspersonal (Reduction of the Energy Consumption of Wastewater Treatment Plants – Manual for the Operation of WWTPs), Heft 4, 3. Auflage (Issue 4, 3rd Edition), DWA Landesverband Baden- Württemberg, Stuttgart (in German).
Corena, M., Bonini, D., Dinesh, N., and Macpherson, R. (2015) Energy Benchmarking of Wastewater Treatment Plants – SA Water’s Experience. Proc. Ozwater2015 Conference, 12-14 May 2015. Adelaide Convention Centre, Adelaide, Australia.
Crawford, G. (2010) Best Practices for Sustainable Wastewater Treatment Initial Case Study Incorporating European Experience and Evaluation Tool Concept. WERF Report No. OWSO4R07a, Water Environment Research Foundation, Alexandria, VA, USA.
DWA (2012). 25. Leistungsvergleich kommunaler Kläranlagen. Deutsche Vereinigung für Wasserwirtschaft, Abwasser und Abfall, Hennef, Germany. www.dwa.de
DWA (2015). Energiecheck und Energieanalyse – Instrumente zur Energieoptimierung von Abwasseranlagen. Arbeitsblatt DWA-A 216. Deutsche Vereinigung für Wasserwirtschaft, Abwasser und Abfall, Hennef, Germany, December 2015. www.dwa.de
Frigge, M; Hoaglin, DC; Iglewicz, B (1989). Some Implementations of the Boxplot. The American Statistician 43 (1): 50–54.
Haberkern, B., Maier, W. & Schneider, U. (2008) Steigerung der Energieeffizienz auf kommunalen Klaeranlagen (Increasing the Energy Efficiency of WWTPs). Umweltbundesamt, Dessau-Roßlau, Germany. (in German).
Krampe J. (2013) Energy benchmarking of South Australian WWTPs Water Sci. Technol. 67 (9), 2059-2066.
LfU (1998). Stromverbrauch auf kommunalen Kläranlage – Handbuch Wasser 4. Landesanstalt für Umweltzschutz Baden-Württemberg. Kahlsruhe, Germany (in German).
Malekizadeh, A. , Dayman, I., Beard, H., McGarry, E., Atkinson, P. , Regel, R., Soriano, GA, and Langlais, C. (2016) Reducing Energy Consumption of Christies Beach WWTP. Proc. Ozwater2016 Conference, 9-11 May 2016. Melbourne Convention Centre, Melbourne, Australia.
McGill, R., Tukey, JW, Larsen, WA. (1978). Variations of Box Plots. The American Statistician 32 (1): 12–16.
Steele, R., Krampe, J. and Dinesh, N. (2013) Process level energy benchmarking as a tool to improve the energy efficiency of wastewater treatment plants. AWA Water Journal, April 2013, 129-134.
WSAA (2017). WWTP Energy Benchmarking 2015-2016. Part 2 – Technical Report. Report prepared for Water Services Association of Australia by GHD, October 2017. WSAA, Melbourne, Australia (250 pages).
Plant types and size classes
For this study, six plant-types were categorised according to treatment processes, together with some sub-types defined. Five plant-class sizes were set according to various ranges of equivalent person (EP).
These definitions were based on similar definitions
Energy benchmark values adopted for this study were defined as follows:
- Guide Values: The Guide Values are intended to reflect average (or typical) performance and were translated from the German references mentioned above, where necessary by extrapolation or scaling. It is anticipated that in future benchmarking projects (as possible extensions of this project), the Guide Values may be amended to reflect the results for Australian plants.
- Target Values: The Target Values are intended to reflect ‘top’ performance or current best practice (top 10%), and are derived from the German references mentioned above, where necessary by extrapolation or scaling.
This study applied two forms of energy benchmark:
- The primary benchmark is specific energy use, expressed as kWh/(EP.year), based on whole-ofplant electrical energy use (average kWh/year) and EP is ‘Equivalent Persons’ calculated from raw influent wastewater load.
- The secondary benchmark is extent of electrical energy self-supply, expressed as percent of average electrical energy use supplied from onsite generation from renewable energy sources (i.e. co-generation from biogas; wind, solar; or other renewable types). This benchmark is only applied to Type 1 plants, although other plant types might report energy self-supply (e.g. from solar or wind power).
The Guide and Target values adopted for the primary and secondary benchmarks respectively in each Plant Type and Size Class category are summarised in Table 4, Table 5 and Table 6 below. Note that Energy Supplement values (Table 6) are added to the base values for the primary benchmark (Table 4), calculated on a plant-byplant basis, depending on whether the relevant processes (effluent treatment, sludge drying or high pumping requirement) are present or not.
Number of plants
Figure 1 gives a summary breakdown of the number of plants by Type for this this benchmarking round (201516) and the previous round (2013-14). Similarly, Figure 2 gives a breakdown by Size Class (all Types).
Change in performance (this round vs. previous round)
Table 7 summarises the change in performance on primary benchmark for plants that participated in both this benchmarking round and the previous round, broken down by Type and Size Class.
Primary benchmark – Probability distribution by Size Class
Figure 3 shows the probability plot of all plants (all Types) in this benchmarking round, broken up according to Size Class. Figure 4 shows the probability plot of plants (all Types) in this benchmarking round that reported high pumping features (nominally >4m head) for influent and/or effluent, broken up according to Size Class. Figure 5 shows the probability plot of plants (all Types) in this benchmarking round that did not report high pumping features for influent and/or effluent, broken up according to Size Class.
Primary benchmark – Probability distribution by Type and Size Class
The following figures shows probability plots for results from this benchmarking round according to Type, including a breakdown according to Size Class:
- Figure 6 for Type 1 plants
- Figure 7 for Type 2 plants
- Figure 8 for Type 3 plants, excluding MBRs (Sub-type 3.4)
- Figure 9 for Type 3 plants, Sub-type 3.4 (MBRs) only
- Figure 10 for Type 4 plants
- Figure 11 for Type 5, Sub-type 5.1 plants (Aerated Lagoons)
- Figure 12 for Type 5, Sub-type 5.2 plants (Unaerated Lagoons/ Wetlands).
Removal of pollutants
The average percent removals across the plant Types (with breakdown by Size Class) are plotted for BOD and/ or COD in Figure 13 and for BOD and/or Total Nitrogen removal in Figure 14.
Relationship to Primary benchmark
Figure 15 shows plots of the relationship between the Primary benchmark used in this study (kWh/[EP.y]) and specific energy use per unit removal of COD, BOD and Total N respectively.
Probability distribution by Type and Size Class for specific energy use per unit nitrogen removed
Figure 16 to Figure 21 show probability plots of kWh/ kgN removed, according to Plant Type and Size Class. Breakdown by Sub-types is not given, except for Type 3, Sub-type 3.4 (MBRs), for which the number of plants and energy use profile was sufficient to merit separate representation.