Effluent concentration vs plant loading

Table 23 presents system characteristics of the 54 plants for which effluent concentration measurements and loading estimations were available. The table excludes all plants where the wastewater was presumably diluted by storm water on the day of sampling. COD data invalid due to high salinity of the sample was also not considered. It furthermore excludes two plants of which the particularly high effluent concentration measurement results were suspected to be not representative (see Section 5.3.2).

Table 23: Number of plants depending on system type, pre-treatment, location and year of implementation presented in this section

System type BGD as pre-treatment

SSS CSC Mixed SBS Yes No

13 28 9 4 41 13

Province North Sumatra West Sumatra West Java &

Banten

Central Java &

Yogyakarta East Java Bali

1 0 6 26 16 5

Year of implementation

2005 2006 2007 2008 2009

2 10 12 13 17

Figure 56 relates the load estimations of the plants presented in Table 23 to measured effluent concentrations. The average design system load range was computed as defined in Section 5.2.6. The upper and lower limits of the design performance prediction were calculated as described under Section 5.2.7.

Figure 56: Effluent concentration values plotted against estimated plant loading expressed as number of connected people per m³ total reactor volume (n= 54). The curves “Design prediction upper/ lower limit”

delimit the confidence range of the design system performance predictions taking into account a per capita wastewater production of 20 to 130 l cap^-1 d^-1 and 20% uncertainty in the COD concentration measurement, the confidence range of the average design load is computed using the average load of 4.9 and the standard deviation of 1.6 cap m^-3 (see Section 5.2.6)

Most plants fall within or below the average design system load range with only four plants having a higher load. Under-loaded plants are mainly CSCs and Mixed systems. Most SSS fall within the average design system load range.

Surprisingly, the data indicates that reduced plant loading does not guarantee an improvement in effluent concentrations, for either of the system types. The data shows no clear correlation between loading and effluent concentration contrary to that generally expected from literature (see Section 2.3.5). Also the existing DEWATS design tool predicts an effluent concentration reduction with reduced plant load (see Figure 56). The effluent COD concentration measured at a large number of systems is however considerably higher than predicted by the design-tool, even exceeding national discharge standards (see Figure 56).

The lack of correlation between loading and effluent concentration is most apparent for CSCs of which low loaded systems show low, medium and high effluent CODt concentrations above 200 mg CODt l^-1. Low system loading would limit the establishment of a stable anaerobic MO community which would consequently lead to little or no COD removal (Bischofsberger et al., 2005; Shen et al., 2004). But even when considering the uncertainties in measurement and loading estimation, the spread of effluent concentration values is such that other factors than the loading must strongly be influencing the treatment of these plants.

Table 24 therefore compares the available information on potentially treatment-influencing factors for the systems depending on their effluent concentration and estimated load. The plants are divided into two groups: “Complying with design” for concentrations within or below design expectations and “Not complying with design” for concentrations exceeding the design expectations.

Average loading range

Design prediction upper limit

Design prediction lower limit

0 2 4 6 8 10 12 14

0 100 200 300

Measured effluent concentration (mg CODtl-1)

Number of connected people per m³ reactor volume

SSS SSS coastal CSC CSC coastal Mixed Mixed coastal SBS

The occurrence of flow surge signs, water scarcity and the various indicators for applied O&M practices does not seem to significantly differ between groups. It was evident however that most plants built in coastal areas had effluent concentrations above design range (also see the data points highlighted in Figure 56). It was shown in Section 5.3.4.2 that on average, wastewater treated by DEWATS in coastal areas has a significantly higher salinity than inland. Based on the available data it is therefore hypothesised that elevated raw-water salinities observed at plants built close to the coast, or the seasonal variation of salinity, may have negative effects on the treatment.

This however will have to be investigated further, also since the dataset does contain data points which contradict this general trend. Six systems for instance that had effluent concentrations comparable to design prediction are built in coastal towns.

Also a number of plants performing comparably poorly are inland and certainly not affected by elevated wastewater salinity. The reason for their poor performance remains unclear.

Table 24: Comparing potentially treatment-influencing factors of DEWATS with effluent concentrations within or above design effluent concentration range

Potentially treatment-influencing factors Complying with design

Not complying with design

General

Total number of systems 33 21

SSS 11 2

CSC 16 12

Mixed 3 6

SBS 3 1

BGD 22 19

Systems in a coastal town 3 10

Feed Signs of flow surge 16 11

Fresh water conductivity > 1 mS cm^-1 1 7

General water scarcity 1 1

Applied O&M practices

No CBO 1 1

No operator 5 0

No biogas usage 4 7

Older than 3 y and never desludged 17 14

No O&M training operator 8 7

No O&M training users 9 1

Figure 56 shows that most plants loaded within the average loading range and built inland produce in sixteen out of nineteen cases effluent concentrations within the range predicted by the design. Their effluent concentrations are mostly around or below 100 mg CODt l^-1.

The four plants loaded above the average design load are two CSCs and two SBS and their effluent concentrations were all within or below design predictions. Surprisingly two of these systems had very low effluent values below 100 mg CODt l^-1. In principle this supports the view that DEWATS could generally be designed smaller while still complying with discharge standards. The number of systems indicating this is however too small to be able to draw strong conclusions. Further investigations are needed in order to confirm this.

5.4. Conclusions

In this chapter treatment indicators of an unprecedentedly large number of DEWATS were compared.

Conclusions are however limited by factors typical for research on sanitation in developing countries and are based on intrinsically error-prone single effluent concentration measurements of plants with varying reactor configuration, each exposed to a unique combination of treatment-influencing circumstances. Statistically meaningful conclusions on factors influencing the system efficiency can therefore in most cases not be drawn. The data however enables a number of important observations:

Many systems were under-loaded. This is especially true for CSC and Mixed systems, less so for SSS and SBS. The effect of this on the effluent concentration was difficult to assess due to the surprisingly large spread of concentration data for low loaded plants.

The field-data and field-observations have shown that all DEWATS types including CSC were exposed to storm-water which possibly impeded the treatment processes. Chapter 6 further discusses this issue.

Water treated by DEWATS in coastal areas of Sumatra, Java and Bali tended to have an elevated level of electric conductivity, most probably due to sea water intrusion to over-exploited aquifers. A large proportion of DEWATS with effluent concentrations above design predictions is built in coastal areas suggesting a possible negative impact on the treatment because of elevated salinity or the variation of salinity inside reactor chambers due to the combined effect of salty ground-water and seasonal rain-water influence. Literature reporting good anaerobic treatment of high saline wasterain-water may not be directly comparable to the here presented situation because of the frequent low organic plant loading.

The dataset did not provide indications that any of the following potentially influencing factors had a statistically meaningful influence on the effluent concentration: location (province), system type, inclusion of BGD in the design, date of implementation, exposure to storm-water, general water scarcity at the site, existence of a CBO and operator, occurrence of desludging of systems older than 3 y, O&M training of the operator and users and use of biogas for systems including a BGD in their design.

It can however not be ruled out that single systems were influenced by these factors, especially since the reason for poor treatment could not be identified for a large number of investigated plants. Each project was exposed to a specific set of circumstances which creates a multi-dimensional space in which the effects of single factors are difficult to isolate.

This obviously also affects the confidence with which conclusions can be drawn on the relation between system loading and effluent concentration. Data on low loaded plants is erratic. However, most plants built inland with loads close to design assumptions appear to produce effluent concentrations within the range of design predictions. Most high loaded plants perform surprisingly well with low effluent concentrations which supports the view that DEWATS are robust towards high loads. Whether this robustness allows future systems to be designed significantly smaller could not be established within the survey presented in this chapter. Future research will need to address this important question by excluding external non-quantifiable influences that the plants discussed in this chapter have been exposed to.

The data indicates guaranteed maximum concentrations of 200 mg CODt l^-1 for the effluent of the anaerobic DEWATS treatment if the influence of saline water can be excluded. It is however important to realize that this value was deduced from systems that were hydraulically over-loaded for large parts of the year due to storm-water intrusion. It is hypothesised that their treatment would improve significantly if their maximum hydraulic load was actually close to their respective design-value. The currently observed treatment-efficiencies however imply the need for anaerobic DEWATS effluent to be further treated through a polishing step in order to comply to the comparably stringent effluent regulations of countries such as Vietnam, Cambodia and the Philippines.