Decentralized Wastewater Treatment Systems (DEWATS) are increasingly being recognized by decision makers and city developers across the world as an option for service delivery in densely populated low-income areas. However, little knowledge and field-data are available on some key design-, operation- and monitoring-parameters. This paper addresses these gaps by presenting data on per capita wastewater production of communities connected to DEWATS, hydraulic peak flow factors, per capita BOD5 load estimates, biogas-production of DEWATS biogas digesters, DEWATS effluent characteristics and their fluctuation over time and settled reactor sludge characteristics concerning TS and VS concentrations, specific methanogenic activity (SMA), effect of sludge storage on SMA and accumulation rates. The investigations have been conducted over five years at 24 communal systems in Indonesia, India and South Africa.

## INTRODUCTION

Decentralized Wastewater Treatment Systems (DEWATS) are increasingly being recognized by decision makers and city developers across the world as an option for service delivery in densely populated low-income areas. However, little knowledge and field-data are available on some key design-, operation- and monitoring-parameters. This is due to the relatively recent history of DEWATS implementation, the limited number of researchers active in this field and the existing challenges encountered during field investigations in project areas.

Detailed design-procedures for DEWATS reactors are described in Sasse (1998). The main design-parameters for communal DEWATS are the estimated per capita wastewater production, peak flow and per capita organic load. However, very little literature is available for communities in developing countries, forcing designers to use unsubstantiated estimations for the sizing of the plants. Biogas-production is often a welcome by-product of the DEWATS treatment process, but yield measurements for biogas-digesters (BGD) fed with communal wastewater have not yet been published. National effluent standards often stipulate maximum concentrations expressed as mg BOD5 L−1. The comparative ease of conducting COD instead of BOD5 concentration measurements in DEWATS implementation areas causes the need to assess the BOD5 to COD ratio in DEWATS effluents. Because of the remoteness of many sites, regular effluent monitoring is often impossible. In order to interpret available concentration data from effluent grab-samples, it is therefore essential to understand the typical variations of DEWATS effluent concentrations. Information on effluent nutrient content is important concerning compliance with national discharge standards, its impact on receiving water-bodies and its reuse value for agriculture.

Settled sludge characteristics such as TS and VS concentrations, activity and accumulation rates across reactor chambers are of importance due to a variety of reasons. The desludging of reactors is the regular DEWATS maintenance activity which requires the largest amount of funds and the highest level of sophistication regarding logistics. It is therefore crucial for city planners to have a good understanding of the required desludging periods of such systems, which chambers to desludge and the sludge densities that have to be expected. The current estimate for desludging periods (two to three years) is largely based on experience with septic tanks and has not yet been validated by formal measurement campaigns. It is further important to understand the effect of sample storage time on sludge activity for obvious methodological reasons and since no literature is available on it in a DEWATS context.

This paper addresses these gaps and presents data on per capita wastewater production of communities connected to DEWATS, hydraulic peak flow factors, per capita BOD5 load estimates, biogas-production of DEWATS biogas digesters, DEWATS effluent characteristics and their fluctuation over time and sludge characteristics. The investigations have been conducted over five years at 24 communal systems in Indonesia, India and South Africa.

## METHODS

### The plants

All investigated DEWATS have been implemented through the local partner network of the German not for profit organisation ‘Bremen Overseas Research and Development Association’ (BORDA). DEWATS are modular systems each adapted to local requirements and constraints, leading to various system configurations and sizes. The configuration always consist of a settling unit (either a BGD and/or settler), followed by an Anaerobic Baffled Reactor (ABR) with varying number of compartments. Optional further anaerobic treatment is achieved by an Anaerobic Filter (AF). Polishing steps such as Horizontal Gravel Filters and ponds such as implemented in India and South Africa are not considered in this survey. Communal DEWATS are either connected to households by a ‘Shallow Sewer System’ (SSS) or to ‘Community Sanitation Centres’ (CSC) with toilets, showers and at times laundry areas. CSC-type DEWATS are also used in boarding schools (SBS). All systems operate under tropical or sub-tropical (NLM in South Africa) climate. Table 1 below lists the plants at which the investigations presented in this paper were performed.

Table 1

Plants from which the field data was used to investigate various design relevant and operation relevant parameters

Plant information

Design parameters

Effluent characteristics

Sludge characteristics

Plant code Country* Type Per cap ww prod. Per cap BOD5 prod. Biogas prod. BOD5/COD ratio Effluent COD variation NH4-N conc. PO4-P conc. TS/VS SMA Sludge accum.
AF IDN SBS
AH IDN SBS X+
BWC IND SSS
DW IDN CSC
FOC IND SSS
GB IDN SSS
GG IDN SSS
KM IDN CSC
KA IDN CSC
KW IDN CSC
KT IDN CSC
KG IDN SSS
MB IDN CSC
MG IDN SSS
MM IDN SSS
NLM ZAR SSS
PW IDN CSC
PY IDN SSS
PB IDN CSC
RN IND SSS
SH IDN SSS
SK IDN CSC
ST IDN SSS
WY IDN SSS
Plant information

Design parameters

Effluent characteristics

Sludge characteristics

Plant code Country* Type Per cap ww prod. Per cap BOD5 prod. Biogas prod. BOD5/COD ratio Effluent COD variation NH4-N conc. PO4-P conc. TS/VS SMA Sludge accum.
AF IDN SBS
AH IDN SBS X+
BWC IND SSS
DW IDN CSC
FOC IND SSS
GB IDN SSS
GG IDN SSS
KM IDN CSC
KA IDN CSC
KW IDN CSC
KT IDN CSC
KG IDN SSS
MB IDN CSC
MG IDN SSS
MM IDN SSS
NLM ZAR SSS
PW IDN CSC
PY IDN SSS
PB IDN CSC
RN IND SSS
SH IDN SSS
SK IDN CSC
ST IDN SSS
WY IDN SSS

*Country abbreviations: IDN–Indonesia, IND–India, ZAR–South Africa.

+X indicates data availability.

### Parameters

Wastewater flows were measured after the last anaerobic treatment step. Former investigations have shown that inflow-fluctuations are not altered throughout the system (Reynaud 2008). Storm-water, although unwanted, affects the flows to most plants (Reynaud 2015). The presented data-sets therefore only include measurements done on days without rain. The integrity of the piping systems and household-connections were assessed with food-dye prior to measurements. The peak-factor f was calculated with the formula Number of connected people and the average household income were estimated through personal communication with the heads of communities. The Hydraulic Retention Time (HRT) was calculated using the formula In the case of BGDs purely fed with black-water average daily flow had to be assumed based on a daily per-capita black-water production of 20 L cap−1 d−1.

Biogas-production was measured after completely venting the BGD and letting the biogas flow freely through a gas-meter. Meter readings were recorded hourly or daily over a period of several days and production rates were computed through linear regression. Calculated biogas-production is based on COD-removal rates for settlers as proposed by Sasse (1998) with connected user numbers based on field investigations, per capita organic load of 60 g BOD5 cap−1 d−1, the fraction of removed COD being converted to methane set to 0.9 and the methane content in the biogas to 0.6.

COD, BOD5, NH4-N and PO4-P concentrations were almost exclusively measured by BORDA project teams using spectrophotometer test kit methods and in few instances outsourced to accredited external laboratories following APHA standards.

Settled sludge-heights were measured with a specially devised core sampler. Homogenised aliquots of decanted settled sludge-samples with known dilution with wastewater were used for solid determinations and activity testing. Total Solids (TS) and Volatile Solids (VS) sludge measurements were done following APHA (1998). All measurements were performed in triplicate with a standard deviation of triplicates below 10%. Sludge activity is compared across reactors using the maximum Specific Methanogenic Activity (SMAmax) test (Soto et al. 1993), which investigates the acetoclastic methanogenic activity of an anaerobic sludge by measuring the amount of CH4 produced by a known amount of sludge (expressed as VS) under ideal substrate saturated conditions. It is expressed as ‘mL CH4 (as COD-equivalents) g VS−1 d−1’. Added substrate was acetate. Further methodological details can be found in Reynaud (2015). Settled sludge accumulation rates were calculated through linear regression of total settled sludge-volumes in ABR chambers over periods undisturbed by desludging events. The amount of sludge being washed out from the ABR into down-stream reactor chambers over the four years of investigation was found to be negligible.

## RESULTS AND DISCUSSION

### Per capita wastewater production in households connected to DEWATS

Per capita wastewater production in the investigated communities was significantly lower than reported for western countries and depended strongly on availability of freshwater.

Table 2 presents the outcomes of fifteen wastewater production measurement campaigns at twelve SSS and one SBS DEWATS. Two communities (RN and BWC) had very limited access to fresh water and particularly low average household incomes. Also, in the case of RN only black-water and grey-water from bathrooms were discharged to the DEWATS. This explains the low wastewater production values comparable to the values proposed by the WHO for arid regions (WHO/UNEP 1997).

Table 2

Wastewater production of connected communities, dates behind plant codes indicate years during which measurements were conducted at the same plant

Total flow+   n^ ww prod. Peak flow
Plant code Number of people m³ d−1 RSD$L cap−1d−1 m³ h−1 Peak flow factor Average income class* RN 608 15.9 21% 26 0.8 1.2 BWC 2012 575 16.5 3% 29 1.5 2.2 KG 480 16.9 31% 10 35 1.0 1.5 BWC 2010 605 23.5 4% 39 1.8 1.8 SH 168 10.3 1% 62 1.1 2.6 WY 271 20.1 6% 74 1.5 1.8 PY 213 16.1 23% 76 0.8 1.2 AH 478 36.8 77 3.1 2.0 ST 450 36.4 5% 81 2.7 1.8 GB 195 16.6 13% 85 1.5 2.2 NLM 420 35.9 17% 107 86 2.5 1.7 GG 103 9.1 16% 88 0.8 2.1 MG 125 11.0 10% 88 0.6 1.4 MM 251 22.9 5% 91 2.2 2.3 Total flow+ n^ ww prod. Peak flow Plant code Number of people m³ d−1 RSD$ L cap−1d−1 m³ h−1 Peak flow factor Average income class*
RN 608 15.9 21% 26 0.8 1.2
BWC 2012 575 16.5 3% 29 1.5 2.2
KG 480 16.9 31% 10 35 1.0 1.5
BWC 2010 605 23.5 4% 39 1.8 1.8
SH 168 10.3 1% 62 1.1 2.6
WY 271 20.1 6% 74 1.5 1.8
PY 213 16.1 23% 76 0.8 1.2
AH 478 36.8   77 3.1 2.0
ST 450 36.4 5% 81 2.7 1.8
GB 195 16.6 13% 85 1.5 2.2
NLM 420 35.9 17% 107 86 2.5 1.7
GG 103 9.1 16% 88 0.8 2.1
MG 125 11.0 10% 88 0.6 1.4
MM 251 22.9 5% 91 2.2 2.3

+ mean.

\$ relative standard deviation.

^ number of days on which flow measurements were performed.

*the following denotations are used to characterize average household income: A = <50 USD month−1; B = 50 USD month−1 to 100 USD month−1; C = >100 USD month−1; ww = wastewater.

The wastewater production measured in KG is surprisingly low, especially since this plant is located in a water-rich area with a connected community with comparably high average income. The data has therefore possibly been affected by undetected leaks or blockages in the sewer system.

Average per capita wastewater production rates of the remaining systems varied from 62 L cap−1 d−1 to 91 L cap−1 d−1 with an average value of 81 L cap−1 d−1. This is significantly lower than the flows generally expected in western countries of 170 L cap−1 d−1 to 340 L cap−1 d−1 (Tchobanoglous et al. 2003). It corresponds however very closely to the values proposed by the WHO for developing regions (WHO/UNEP 1997) as well as to measurements performed in rural areas in Thailand (Tsuzuki et al. 2010). Wastewater production in poor communities in Brazil has been reported to depend on the average household income (Campos & von Sperling 1996). This does not seem to be the case for the investigated Indonesian communities, living in areas with shallow well water freely available to all income groups: the available data shows no correlation between measured daily per capita wastewater production values and the average monthly household income.

The dataset available from NLM represents 111 d of continuous flow-measurements and is by far the most comprehensive and therefore most reliable dataset available for this study. The relative standard deviation (RSD) of the daily flow over that period was 17%. Based on this and the other available data the daily variation of typical DEWATS dry weather feed flow is estimated to be 20%.

Measured diurnal flow variations were typical for communal wastewater with two peaks, one in the morning and one in the evening. The morning-peak was generally the strongest as typical for household discharge (Haestad et al. 2004) and lasted for 2 h to 3 h. The average peak-flow factor across all plants listed in Table 2 was 1.9 ± 20%.

### Per capita BOD production in households connected to DEWATS

Primary treatment effluent concentration measurements indicate that per capita organic loads are significantly lower than the generally assumed design value of 60 g BOD5 cap−1 d−1 although the available data did not enable a direct quantification.

Table 3 presents available data on four DEWATS pretreatment steps. Feed concentration data are not available due to methodological difficulties associated with feed sampling. The given HRTs represent maximal values since they do not take reactor volume reduction through accumulated primary sludge into account. Sludge height data however indicates that actual HRTs are not below 50% of the given values. Estimated per capita COD production values are calculated with available effluent concentration data and an assumed primary treatment efficiency of 50% (Foxon 2009; Mang & Li 2010). Although these estimates do by no means represent accurate per capita values, they do indicate quite clearly that the generally reported design value of 60 g BOD5 cap−1 d−1 significantly overestimates real values in low income communities in Indonesia and India. Similar observations have been reported by Campos & von Sperling (1996) for Brazil.

Table 3

Estimating per capita COD production based on primary treatment step effluent concentrations

Plant code Number of people HRT Qmean Primary treatment effluent COD concentration

Estimated per cap COD production
m³ d−1 mg COD L−1+ RSD n g COD cap−1 d−1
BWC 575 73 16.5 513 15% 13 29
GB 195 27 16.6 393 24% 13 67
MM 251 10 22.9 436 41% 12 80
ST 450 13 36.4 349 18% 56
Plant code Number of people HRT Qmean Primary treatment effluent COD concentration

Estimated per cap COD production
m³ d−1 mg COD L−1+ RSD n g COD cap−1 d−1
BWC 575 73 16.5 513 15% 13 29
GB 195 27 16.6 393 24% 13 67
MM 251 10 22.9 436 41% 12 80
ST 450 13 36.4 349 18% 56

+mean.

### Biogas-production in communal DEWATS applications

Per capita biogas production was found to be approximately 20 L cap−1 d−1, was comparably stable across a varying range of HRTs and did not significantly depend on feed concentration.

Figure 1 compares the per capita biogas production to the average HRT of the BGDs and to calculated predictions, computed as explained under ‘Methods’. Each round data point represents one site each, each triangular data point represents the outcomes of several measurement campaigns performed at the same plant. The average biogas-production across the systems is 20 L cap−1 d−1 ± 36%. With the exception of two outliers, the field data shows a good fit to the prediction given by the design procedure. Surprisingly, BGDs fed with black- and grey-water show similar biogas-production as systems purely fed with black-water with similar HRTs. Per capita biogas-production did not significantly increase with HRTs above 2.5 d. Since the biogas-production correlates with COD removal the data indicates an optimal BGD design with an HRT of 2.5 d.
Figure 1

Per capita biogas production depending on the HRT of the pre-treatment.

Figure 1

Per capita biogas production depending on the HRT of the pre-treatment.

No directly comparable literature was found on biogas-digesters treating purely communal wastewater at such low retention times. Garuti et al. (1992) reported a biogas production of approximately 14 L biogas cap−1 d−1 at the communal wastewater fed ABR they investigated. Lohri et al. (2010) and Pipoli (2005) reported similar or lower rates than observed in this study. The comparably high value of 41 L cap−1 d−1 reported by Zurbrügg et al. (2011) is due to the addition of kitchen waste which is known to significantly increase biogas production (Lohri et al. 2010). Remarkably, the BGDs presented here had similar biogas production as reported in literature although they were operated at far shorter HRTs (1 to 9 d compared to 15 to 37 d).

### Characteristics of DEWATS effluent

#### BOD5/COD ratio

A total of eighty simultaneous COD and BOD5 measurements were performed on anaerobic treatment effluent samples from sixteen different DEWATS plants (see Figure 2) which show an average BOD5/COD ratio of 0.46 ± 38%. Kerstens et al. (2012) present a dataset with 32 measurements from eight DEWATS plants showing an average BOD5/COD ratio of 0.4 ± 16%. Rochmadi et al. (2010) report to have measured average communal DEWATS effluent concentrations of 22 mg BOD5 L−1 and 61 mg COD L−1 which corresponds to a BOD5/COD ratio of 0.36.
Figure 2

COD vs. BOD5 effluent concentrations.

Figure 2

COD vs. BOD5 effluent concentrations.

The average BOD5/COD ratio of the outflow from anaerobic DEWATS reactors thus is surprisingly high and only slightly lower than screened North-American raw wastewater, reported to have a ratio of 0.49 (Smith & Eilers 1969; Dixon et al. 1972). Wastewater after biological treatment is reported to have a significantly lower ratio of 0.1 to 0.25 (Mara & Horan 2003).

While this could indicate a comparably low content of nonbiodegradable COD in the wastewater treated by DEWATS, it could also mean that significant amounts of biodegradable organics leave the DEWATS after the last anaerobic treatment step untreated. The latter explanation is supported by the often high measured effluent BOD5 concentrations (see Figure 2).

#### Variation of effluent COD concentration over time

The RSD of effluent COD concentrations over extended measurement periods was 11% to 18% for plants with effluent concentrations above 100 mg COD L−1 (see Table 4). RSD were found to be considerably higher for plants with lower effluent concentrations, probably due to the higher methodological error in these concentration ranges.

Table 4

Long-term variation of COD effluent measurements at seven different systems

Plants

COD effluent variation

Plant Code Plant type Period of sampling (months) Mean (mg COD L−1SD (mg COD L−1RSD n
BWC SSS 41 320 59 18% 37
FOC SSS 98 82 30 36% 33
GB SSS 49 127 22 18% 68
MM SSS 67 77 26 39% 122
NLM SSS 406 55 13% 18
SK CSC 39 167 27 16%
ST SSS 25 108 12 11% 19
Plants

COD effluent variation

Plant Code Plant type Period of sampling (months) Mean (mg COD L−1SD (mg COD L−1RSD n
BWC SSS 41 320 59 18% 37
FOC SSS 98 82 30 36% 33
GB SSS 49 127 22 18% 68
MM SSS 67 77 26 39% 122
NLM SSS 406 55 13% 18
SK CSC 39 167 27 16%
ST SSS 25 108 12 11% 19

Figure 3 shows the average hourly AF effluent COD concentrations from hourly measurements taken on five consecutive days. No rain was recorded on any of these days. All values taken at the same time of day were found to be normally distributed using the Shapiro-Wilk normality test1. A one-way between subjects ANOVA test showed that no significant difference exists at the p < 0.05 level between the values measured at different times of the day [F(13, 54) = 1.32, Fcrit = 1.91]. Thus, the time of day at which effluent samples are drawn does not significantly influence the measurement outcome. Total average of all measurements (n = 68) is 54 mg COD L−1 with a standard deviation of 10 mg COD L−1. This corresponds to a RSD of 20% which is comparably high. As explained above, the RSD is expected to be lower for higher average effluent concentrations.
Figure 3

Average hourly effluent COD-concentrations from hourly measurements done on five consecutive days from the 19th to the 23rd of July, 2008 in MM, Indonesia, error-bars indicate the standard deviation of hourly measurements.

Figure 3

Average hourly effluent COD-concentrations from hourly measurements done on five consecutive days from the 19th to the 23rd of July, 2008 in MM, Indonesia, error-bars indicate the standard deviation of hourly measurements.

#### Nutrients

Table 5 shows the nutrient concentrations measured at the effluent of the last anaerobic treatment steps of six sites. Nutrient effluent concentrations of all investigated systems were comparably high. The per capita nutrient loads varied little (ammonium) and moderately (phosphorous) across investigated communities. Measured concentrations therefore depended primarily on dilution through wastewater. Measured concentration ranges are comparable to other publications (Garuti et al. 1992; Foxon 2009; Kerstens et al. 2012). Anaerobic treatment processes do not affect the nutrients, therefore enabling per capita nutrient load estimations through effluent concentration data. The resulting average per capita load values of 5.6 g NH4-N cap−1 d−1 and 0.8 g PO4-P cap−1 d−1 are comparable to values reported in literature (WHO/UNEP 1997; Tchobanoglous et al. 2003).

Table 5

Effluent ammonia and phosphorous concentrations of seven different SSS

Plants  Ammonia (NH4-N) effluent conc.

Phosphorous (PO4-P) effluent conc.

Plant Code Period of sampling* Mean # SD# RSD n Per cap+ Mean# SD# RSD n Per cap+
BWC 41 123 40 33% 27 5.0 18 17% 10 0.7
FOC 98      18 39% 10 1.6
GB 49 76 20 26% 51 6.5 21% 0.5
GG 78 4% 6.8 11 1% 1.0
MM 67 49 8% 5.5 11% 0.6
NLM 61 21 34% 5.2 12% 0.8
ST 25 50 17% 6.0 20% 0.5
Plants  Ammonia (NH4-N) effluent conc.

Phosphorous (PO4-P) effluent conc.

Plant Code Period of sampling* Mean # SD# RSD n Per cap+ Mean# SD# RSD n Per cap+
BWC 41 123 40 33% 27 5.0 18 17% 10 0.7
FOC 98      18 39% 10 1.6
GB 49 76 20 26% 51 6.5 21% 0.5
GG 78 4% 6.8 11 1% 1.0
MM 67 49 8% 5.5 11% 0.6
NLM 61 21 34% 5.2 12% 0.8
ST 25 50 17% 6.0 20% 0.5

*in months.

#in mg L−1.

+in g cap−1 d−1.

### DEWATS sludge characteristics

Sludge investigations provide a coherent picture across investigated systems with highest sludge densities in settlers and highest SMA in the front ABR chambers.

All measurements indicate higher sludge (and especially TS) concentrations in the first chamber and approximately constant concentrations in all following reactor chambers (see Figures 4 and 5).
Figure 4

Settled sludge TS concentrations across reactor chamber at four plants; error bars indicate standard deviation of multiple measurements; numbers in brackets indicate number of respective measurements.

Figure 4

Settled sludge TS concentrations across reactor chamber at four plants; error bars indicate standard deviation of multiple measurements; numbers in brackets indicate number of respective measurements.

Figure 5

Settled sludge VS concentrations across reactor chamber at four plants; error bars indicate standard deviation of multiple measurements; numbers in brackets indicate number of respective measurements.

Figure 5

Settled sludge VS concentrations across reactor chamber at four plants; error bars indicate standard deviation of multiple measurements; numbers in brackets indicate number of respective measurements.

In all four plants settled settler sludge had higher TS concentrations than settled ABR sludge. In three plants (BWC, MM, ST) the highest ABR-TS concentrations were measured in the first ABR chambers. In the case of BWC and MM, the highest VS concentrations were observed in the settlers and first ABR chamber. In the other plants the VS concentrations were approximately constant throughout the ABRs. Average TS concentrations of ABR sludge varied across the systems from about 50 g TS L−1 to 95 g TS L−1. The sludges from all four ABRs had a similar average VS concentration of about 30 g VS L−1. Mtembu (2005) and Foxon (2009) reported much lower settled sludge densities on their pilot ABR of 12 g TS L−1 to 34 g TS L−1 and 7 g VS L−1 to 19 g VS L−1. Koottatep (2014) reported TS concentrations of thickened bottom sludge in onsite sanitation systems treating raw sewage of 40 to 220 g TS L−1. VS content of this sludge was 60 to 70%. The significantly lower VS content of the sludge observed in the four case studies (52%, 42%, 55% and 39% in BWC, GB, MM and ST respectively) may be due to better stabilisation. Standard deviations across multiple measurement-campaigns of the same sludges indicate moderate variation across measurements (see error-bars in Figures 4 and 5).

The effect of sludge storage time (at 2 °C to 6 °C) on SMAmax was tested by performing SMA tests on four different sludges after 1 d to 6 d and 37 d of storage. Significant reduction of sludge activity was noticed in all four cases and was more pronounced for sludges with little initial activity (see Figure 6). Current results therefore suggest that DEWATS sludges should be processed as soon as possible after sampling, certainly within the period of one week.
Figure 6

SMAmax values of sludges after varying sludge storage times.

Figure 6

SMAmax values of sludges after varying sludge storage times.

Settler sludge yielded very low SMAmax values indicating low fractions of active methanogenic microorganisms in all cases (see Figure 7). All sludges sampled from rear ABR chambers and especially AF chambers yielded low SMAmax values whereas sludges from the first three ABR chambers generally yielded the highest SMAmax values. All samples were processed within one week after sampling.
Figure 7

SMAmax values measured across reactor chambers of three plants.

Figure 7

SMAmax values measured across reactor chambers of three plants.

Average settled sludge accumulation rate of six ABRs was 5.5 L cap−1 y−1 ± 40% (see Figure 8). Foxon (2009) reports data corresponding to a much higher sludge accumulation of about 63 L cap−1 y−1 possibly due to be the nature of the feed which, although being screened, was not pretreated. No clear correlation exists between sludge accumulation rates and the HRT of the primary treatment or its type (see Figure 8). Figure 9 indicates that about 50% of all sludge accumulation inside the ABR occurred beyond the second chamber (except in GG, where almost all accumulation took place in the first two chambers). The minimum and maximum measured accumulation rates imply an approximate sludge accumulation of 1.5 L cap−1 y−1 to 4.2 L cap−1 y−1 in the rear compartments. With a typical area of 2.5 m × 0.7 m, a total of five ABR chambers and 300 connected users, this leads to a 90 mm to 240 mm sludge height increase per year in the three rear ABR chambers.
Figure 8

Per capita settled ABR sludge accumulation depending on the HRT of the pre-treatment.

Figure 8

Per capita settled ABR sludge accumulation depending on the HRT of the pre-treatment.

Figure 9

Fraction of total settled ABR sludge accumulation inside chamber as measured in 6 plants.

Figure 9

Fraction of total settled ABR sludge accumulation inside chamber as measured in 6 plants.

DEWATS ABR O&M manuals state that an ABR needs to be desludged after two to three years when sludge-blankets have reached a height of about 1 m (personal communication with BORDA). The available measurements however suggest an ABR desludging frequency of the last three ABR chambers of at least four years. Since highest specific sludge activity was found to establish in the first reactor chambers it is proposed to never desludge these.

## CONCLUSIONS

Average per capita wastewater production of ten investigated communities living in non-water scarce areas was 81 L cap−1 d−1 with long-term fluctuations of about 20%. It was not found to correlate with average household income probably due to freely available ground-water. Per capita wastewater production in poor and water scarce areas in Bangalore/ India was about 30 L cap−1 d−1. The average diurnal peak-flow factor across investigated communities was 1.9 and the strongest peak generally occurring in the morning for a duration of 2 h to 3 h. Primary treatment effluent concentration measurements indicate per capita organic loads significantly lower than the generally assumed design value of 60 g BOD5 cap−1 d−1. A direct quantification however was not possible with the available data.

Average per capita biogas-production measured at eight communal DEWATS BGDs was 20 L cap−1 d−1. It did not increase with HRTs above 2.5 d and it is proposed to use this value for the dimensioning of BGDs operating under DEWATS-typical circumstances. Field data compares reasonable well to the biogas production estimation by Sasse (1998).

The average BOD5/COD ratio of 16 investigated DEWATS effluents was 0.46 indicating incomplete degradation of organic material.

Five plant-effluents with average COD concentrations above 100 mg COD L−1 had a RSD of 11% to 18% over several years. Effluent COD concentrations below 100 mg COD L−1 had RSD above 35%. Diurnal variations of effluent COD concentrations were found to be statistically negligible.

Nutrient concentrations in the effluent of anaerobic DEWATS treatment steps were high and exceeded at some plants 100 mg NH4-N L−1 and 15 mg PO4-P L−1. This is attributed to the comparatively low per capita wastewater production in certain project areas since the per capita nutrient loads were similar across all seven investigated sites and comparable to literature.

Settled sludge TS and VS densities were generally highest in front reactor chambers and similar across downstream reactor chambers. Sludges from front ABR chambers had higher SMAs than sludges from rear ABR chambers. Sludges in settlers and AFs had very low SMAs. Sludge activity was negatively affected by storage time. Based on available data a maximum storage time of one week is proposed for communal DEWATS SMA investigations. Average settled sludge accumulation rates measured in 6 ABRs were 5.5 L cap−1 y−1. Approximately 50% of the total sludge accumulation occurred downstream of the first two compartments. Based on the available data, previously estimated desludging intervals of two to three years could be extended to at least four years. It is proposed to never desludge the first ABR chambers since they were the chambers where most specific sludge activity developed.

## ACKNOWLEDGEMENTS

The field- and laboratory investigations performed on plants by the Consortium of DEWATS Dissemination (CDD), Bangalore/ India and by the BORDA R&D teams Yogyakarta/ Indonesia are acknowledged. Research conducted in Durban was funded by the Water Research Commission (Research Project KS–2002).

1

The data subsets representing 10:00 and 19:00 could not be tested since they included only four data points. All other data subsets fulfilled the minimum requirement of five data points to conduct a Shapiro-Wilk normality test.

## REFERENCES

REFERENCES
APHA
1998
Standard methods for the examination of water and wastewater
.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington DC
,
USA
.
Campos
H. M.
von Sperling
M.
1996
.
Water Science and Technology
34
(
3–4
),
71
77
.
Dixon
H.
Bell
B.
Axtell
R. J.
1972
Design and operation of the works of the Basingstoke Department of Water Pollution Control
.
Water Pollution Control
71
(
2
):
167
175
.
Foxon
K. M.
2009
Analysis of a pilot-scale anaerobic baffled reactor treating domestic wastewater
.
PhD Thesis
,
University of KwaZulu-Natal
,
Durban
.
Garuti
G.
Dohanyos
M.
Tilche
A.
1992
Anaerobic-aerobic combined process for the treatment of sewage with nutrient removal – the ANANOX(R) process
.
Water Science and Technology
,
25
(
7
),
383
394
.
M.
Walski
T. M.
Merritt
L. B.
Walker
N.
Whitman
B. E.
2004
Wastewater collection system modelling and design
.
,
USA
.
Kerstens
S. M.
Legowo
H. B.
Gupta
I. B. H.
2012
.
Journal of Water Sanitation and Hygiene for Development
2
(
4
):
254
265
.
Koottatep
T.
Surinkul
N.
Panuvatvanich
A.
2014
Accumulation rates of thickened-bottom sludge and its characteristics from water-based onsite sanitation systems in Thailand, http://www.susana.org/docs_ccbk/susana_download/2-1624-koottatep.pdf, (Accessed 21st February 2014)
.
Lohri
C.
Voegeli
Y.
Oppliger
A.
Giusti
A.
Zurbrügg
C.
2010
Evaluation of biogas sanitation systems in Nepalese prisons
. In
IWA-DEWATS Conference 2010, Decentralized Wastewater Treatment Solutions in Developing Countries Conference and Exhibition
,
Surabaya
,
Indonesia
.
Mang
H.-P.
Li
Z.
2010
Technology review of biogas sanitation–Biogas sanitation for blackwater, brown water or for excreta and organic household waste treatment and reuse in developing countries, Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ)
,
Eschborn
,
Germany
.
Mara
D.
Horan
N.
2003
Handbook of Water and Wastewater Microbiology
.
Elsevier
,
San Diego, USA
.
Mtembu
D. Z.
2005
The Anaerobic Baffled Reactor for Sanitation in Dense Peri-Urban Settlements
.
M.Sc. Thesis
,
University of KwaZulu-Natal
,
Durban
.
Pipoli
T.
2005
Feasibility of biomass-based fuel cells for manned space exploration
. In:
Seventh European Space Power Conference
,
Stresa
,
Italy
.
Reynaud
N.
2008
Classifying and Monitoring DEWATS-plants in Java
.
M.Sc. Thesis
,
Technical University of Dresden
.
Reynaud
N.
2015
Operation of Decentralised Wastewater Treatment Systems (DEWATS) Under Tropical Field Conditions
.
PhD Thesis
,
Technical University Dresden
.
R.
Ciptaraharja
I.
T.
2010
.
Water Pratice & Technology
5
(
4
),
doi:10.2166/wpt.2010.091.
Sasse
L.
1998
DEWATS – Decentralised wastewater treatment in developing countries
.
BORDA
,
Bremen
.
Smith
R.
Eilers
R. G.
1969
A generalized computer model for steady state performance of the activated sludge process, Div. of Research, Cincinnati, Ohio, Advanced Water Treatment Branch
.
Soto
M.
Mendez
R.
Lema
J. M.
1993
.
Water Research
27
(
8
),
1361
1376
.
Tchobanoglous
G.
Burton
F. L.
Stensel
H. D.
2003
Wastewater Engineering, Treatment and Reuse
.
McGraw-Hill
,
New York, USA
.
Tsuzuki
Y.
Koottatep
T.
Jiawkok
S.
Saengpeng
S.
2010
.
Water Science and Technology
62
(
2
),
231
244
.
WHO/UNEP
,
1997
Water pollution control – a guide to the use of water quality management principles WHO/UNEP
, ).
Zurbrügg
C.
Vögeli
Y.
Estoppey
N.
2011
Digesting faeces at household level – experience from a ‘Model Tourism Village’ in South India. Sustainable Sanitation Practice (9)
.