ADDRESSING FINANCIAL AND TECHNICAL SUSTAINABILITY CONSIDERATIONS

DURING THE SELECTION OF SMALL WATER TREATMENT SYSTEMS

G.S. Mackintosh

Cape Water Programme, CSIR, P O Box 320, Stellenbosch, South Africa,

(E-mail: gmackint@csir.co.za)

1. BACKGROUND

When considering water treatment options and alternatives for a particular project it is crucial to consider all aspects

that input into the ultimate long term viability and sustainability of the project. The South African Water Services

Act of 1997, and the associated draft Regulations, oblige water service authorities to progressively ensure

“efficient, affordable, economical and sustainable access to water services”. This is easier said than done, and in

essence requires that the responsible project engineer consider technology choice with regards to technical

sustainability, financial sustainability, environmental sustainability and socio-political sustainability. If any one of

these “legs of sustainability” is defective, the overall project viability will become suspect.

Much has been written and discussed about the need for environmental and socio-political sensitivity for rural water

treatment and water supply projects in South Africa. However, what has received less attention is the technology

choice and the financial implications thereof. These decisions are usually left to technical experts, such as

engineers and hydrogeologists, whom it is presumed are suitably skilled. However, inappropriate decisions can

result in the project being financially or technically unsustainable thereby leading to the ultimate collapse of the

water treatment process. At this stage, it is often the community that are erroneously blamed for their inability to

maintain the system. This paper attempts to show, by virtue of two case studies, the importance of addressing

financial and technical sustainability considerations when deciding on small water treatment systems.

2. SUSTAINABILITY

When considering water treatment in small, rural communities sustainability needs to be considered from the

financial, technological, environmental and socio-political perspectives. It is common place in South Africa that

well designed, effective water treatment plants are non-functional as a result of either financial non-sustainability or

socio-political non-sustainability. However, it is also found that where financial and socio-political sustainability

occur, a poor technology has been chosen and with time the water treatment system has become non-sustainable.

2.1 Technological Sustainability

Traditionally in South Africa, small water treatment systems have made use of conventional processes such as

chemical pre-treatment, coagulation, settling and dual media filtration in a scaled-down manner. These unit

processes, although effective on large-scale applications, inevitably prove troublesome for small-user systems – the

result being production of sub-standard water, and non-sustainability with the plant eventually falling into disuse.

This issue of non-sustainability of water supply schemes resulting from imposition of non-appropriate technologies

on communities requires considered attention in South Africa. It is generally recognized that an effort should be

made to use so-called “appropriate technologies”, which are usually low cost, robust, low operator attention

systems (such as slow sand filters). Paradoxically, at times the use of semi-automated “high-tech” plants (for

example, incorporating membrane technology) will also be appropriate for rural water treatment. The critical

deciders here should include:

o The ability of a local community member to operate the water treatment system, with minimal outside

support.

o The ongoing and ready availability of necessary consumables (typically, chemicals).

o The availability, response time and affordability of technical backup where required (especially wrt higher

tech solutions, eg membranes).

2.2 Financial Sustainability

A significant influencing factor as to sustainability is community affordability in terms of the ongoing running costs

of water treatment. It is particularly important to consider this aspect in instances where the initial capital layout of

a technologically attractive system is considered to be too high in comparison to competing, yet technologically less

sustainable technologies. Examples of such instances include:

o Slow sand filters vs pressure filters;

o Electro-chlorinators vs gas chlorination;

o Membrane based processes vs conventional clarification processes.

In such instances, the higher capital outlay projects have potentially lower operating costs; hence this is

advantageous to the community that must guarantee payment thereof into the future. However, the funding

organisation that incurs the upfront capital costs is often discouraged by the significantly higher capital costs. It is

hence important that a clear and unambiguous financial comparison be carried out to enable informed decision-

making. In this paper the Net Present Value (NPV) based approach considers operating costs/cash flow projection

of a project over a ten-year period. This simple NPV assessment captures both the capital and operating costs for

alternative technologies and relates these as one financial sum in terms of today’s monetary value. This approach

clearly indicates which alternative is financially more viable, or the magnitude of any variation.

3. CASE STUDY ONE: MEMBRANE BASED WATER TREATMENT PLANT

3.1 Background

In this case study, the treatment of poor quality and problematic surface water at two neighbouring rural

communities in the Western-Cape, South Africa is contrasted. Both communities number about 2 000 people and

make use of the same raw water source; however, in one instance treatment is being considered via a membrane

based process and in the other instance treatment is via a conventional processes. A critical comparison will be

made between mentoring and training of community-based plant operators, operability and maintenance

requirements, plant performance, sustainability, community acceptance and affordability of both a conventional and

a membrane-based water treatment plant.

3.1 Technical Assessment of Membrane Based Water Treatment

The pilot scale membrane-based plant considered is a movable package water treatment plant designed to condition

surface water so as to comply with international water quality standards. The unit incorporates flocculation and pre-

filtration, membrane filtration, adsorption and chemical disinfection and treats approximately 2000 L/hr depending

on the raw water characteristics. Chemicals used to treat the raw water included those required for aiding

flocculation (polyaluminium chloride - PAC and polyelectrolyte), pH adjustment (soda ash) and disinfection

(calcium hypochlorite - HTH). In addition, calcium hypochlorite and citric acid were used as cleaning chemicals for

the membranes. The membrane-based water treatment plant is shown in Figure 1.

Figure 1: Membrane based water treatment plant

Figure 1: Membrane based pilot plant

The membrane-based plant was installed at the town of Suurbraak, Western-Cape; and a community member with

no previous water treatment training or experience was trained over 2 weeks in all aspects of plant operation.

Sample analysis to evaluate plant performance was based on SABS 241-2001: South African Standard for Drinking

Water (SABS, 2001), which is similar to international drinking-water quality standards.

Typical results obtained from the membrane based water treatment plant are shown in Table 1. The results obtained

during the initial trial period indicated that the plant performed well, consistently providing a high quality drinking

water (De Souza and Mackintosh, 2000). During the subsequent eight-month assessment period, the plant failed on

two occasions (Mackintosh and De Souza, 2001). On the first occasion the problem was found to be a faulty relay,

which was rapidly rectified with the assistance of a local electrician. In the second instance the plant malfunctioned

as a result of the failure of an air valve, which was subsequently replaced under specialist supervision. Hence, the

only problems that occurred during the eight-month period were trivial and easily rectified. However, in both cases

plant downtime was approximately 1 week, which highlighted the requirement for adequate back-up

service/support.

Table 1: Membrane based water treatment plant: typical physico-chemical and microbiological results

Determinant Raw Final

Calcium as Ca (mg/L) 0.5 1.6

Alkalinity as CaCO3 (mg/L) 0.0 4.0

Iron as Fe (mg/L) 0.5 0.05

Aluminium as Al (mg/L) 0.4 0.06

Electrical Conductivity (mS/m @ 25ºC) 4.7 4.7

PH 4.6 7.2

Total Dissolved Solids (mg/L) 30 30

Turbidity (NTU) 0.44 0.28

Colour (Unfiltered) (mg Pt/L) 250 < 10

Colour (Filtered) (mg Pt/L) 250 < 10

Determinant Raw Final

Heterotrophic Plate Count (per 1

mL @ 35ºC)

291

+10000

277

+10000

291

+100001 3 5

Total Coliform (per 100 mL) 1660 1613 1810 0 0 0

Faecal Coliform (per 100 mL) 193 177 162 0 0 0

3.3 Technical Assessment of Conventional Water Treatment Plant Performance

SABS Class 0 (Ideal) Comparable to international water quality standards

SABS Class I (Acceptable) Acceptable for lifetime consumption

SABS Class II (Max. allowable) Acceptable for short term consumption

Failure SABS Class II Unfit for human consumption

No SABS guideline

The performance of the membrane-based plant was compared to a nearby conventional plant, which treats

essentially the same source water. This existing water treatment system treats approximately 10 000 L/hr and

employs conventional water treatment principles of coagulation, flocculation, sedimentation, sand filtration and

disinfection (see Figure 2).

Typical results obtained from the conventional water treatment plant are shown in the following Table 2. The

results obtained during the assessment period showed that the conventional plant is highly vulnerable to passing on

contaminated treated water to the end-user when not operating optimally. Frequent episodes of treated water quality

failing SABS 241-2001 Maximum Allowable standards (i.e. not fit for human consumption) occurred. Both the

plant operator and the community confirmed that the plant did not continuously operate at an optimal level, and

often passed on sub-standard water.

Figure 2: Conventional water treatment plant

Table 2: Conventional water treatment plant: typical physico-chemical and microbiological results

Determinant Raw Final

Calcium as Ca (mg/L) 1.6 1.5

Alkalinity as CaCO3 (mg/L) 5.3 0.5

Iron as Fe (mg/L) 1.3 0.11

Aluminium as Al (mg/L) 0.62 0.57

Electrical Conductivity (mS/m @ 25ºC) 8.0 9.8

PH 6.1 4.7

Total Dissolved Solids (mg/L) 51 63

Turbidity (NTU) 2.5 1.5

Colour (Unfiltered) (mg Pt/L) 300 20

Colour (Filtered) (mg Pt/L) 250 10

Determinant Raw Final

Heterotrophic Plate Count (per 1 mL @ 35ºC) 3100 6550

Total Coliform (per 100 mL) 1575 90

Faecal Coliform (per 100 mL) 730 28

3.3 Comparative Financial Assessment – Membrane-Based Vs. Conventional Water Treatment Plant

The cost comparison was based on a water treatment plant capacity of 10 000 L/hr operating for 20 hours/day,

using April 2001 prices. The cost comparison shown in Table 3 shows that the total installed capital cost of the

membrane based plant is significantly more expensive (~ 1.9 times) than that of the conventional water treatment

plant. Furthermore, the cost comparison showed that the membrane based plant shows significant operating cost

savings over the conventional plant. This can mostly be attributed to lower labour and chemical requirements.

Table 3: Cost comparison input variables

Membrane ConventionalPLANT CAPACITY 10 000 L/hr 10 000 L/hr

TIC COST R600 000 R320 000

TOTAL OPERATING COSTS R1.54/kL R1.12/kL

Chemicals Chemical Dose Cost (R/kL) Chemical Dose Cost (R/kL)

PAC @ R8.96/kg 35 mg/L 0.315 60 mg/L 0.54

Polyelectrolyte @ R8.96/kg 0.1 mg/L 0.001-

-

HTH @ R14.00/kg 0.5 mg/L 0.007 2 mg/L 0.028

Soda Ash @ R2.40/kg 50 mg/L 0.120 50 mg/L 0.12

Citric acid (membrane cleaning) @ R30.00/kg 0.1 mg/L 0.003-

-

HTH (membrane cleaning) @ R14.00/kg 0.1 mg/L 0.001-

-

Chlorine gas @ R14.00/kg - - 2 mg/L 0.028

Chemical Wastage @ 5% - 0.022 - 0.0358

SABS Class 0 (Ideal) Comparable to international water quality standards

SABS Class I (Acceptable) Acceptable for lifetime consumption

SABS Class II (Max. allowable) Acceptable for short term consumption

Failure SABS Class II Unfit for human consumption

No SABS guideline

Electricity @ R0.2/kWh Power consumption Cost (R/kL)Power

consumptionCost (R/kL)

Plant power consumption 8.5 kW 0.150 10 kW 0.20

Labour @ R18.75/hr Time Cost (R/kL) Time Cost (R/kL)

Plant operation, maintenance, etc 1 hr a day 0.094 4 hrs a day 0.376

Maintenance @ 5% of capital cost Cost (R/kL) 0.411 Cost (R/kL) 0.216

Meaningful comparison of the capital and running cost figures given in Table 3 is difficult. A more useful manner

of comparing the two processes is to use a Net Present Value (NPV) based approach. The NPV approach relates

the cash flow projection of a project over a specific time period (in this case 10 years). The NPV assessment

captures both the capital and operating costs for the two alternative technologies and relates these as one financial

sum in terms of today’s money. An important aspect is the discount rate used. For this case study the total discount

rate included: inflation [@ 7% in South Africa], required real return [@ 0%, as no return on investment required by

government funders] and risk [@ 10% for the membrane based process, as membrane based processes are less

familiar for rural use in South Africa].

The NPV based cost comparison shown in Table 4 shows that the use of a membrane based plant (with higher

capital costs and higher risk but lower running costs), yields a nominally negative NPV of R33 500 (and an Internal

Rate of Return of 14%). This result shows that there is very little difference in financial performance between the

two technologies when compared over ten years. It is important to note that this observation is contrary to

conventional thinking in South Africa, where the initial significantly higher capital costs of membrane-based plants

are considered to make the use thereof a non-option.

Table 4: Project financial assessment summary – membrane vs. conventional (10 kL/hr)

Conventional Membrane

Capital cost (R) 320 000 600 000

Total Operating cost (R/kL) 1.54 1.12

Discount rate

Average inflation

Required real return

Estimated risk

7%

0%

0%

7%

0%

10%

Internal Rate of Return (IRR) 14%

Net Present Value (NPV) - R33 500

4. CASE STUDY TWO: LIMESTONE MEDIATED STABILISATION

4.1 Background

The majority of water sources in South Africa require stabilisation, to mitigate against corrosive and

aggressive attack of distribution systems, prior to discharge of treated water. Conventional stabilisation (via

the addition of lime and carbon dioxide, or sodium alkali’s and carbon dioxide) is well documented and

understood, and practised worldwide. However, control of the process is expensive and requires well-trained

staff and reliable equipment. Hence, in many cases only lime is dosed, such that pH is adjusted from low

levels to more desirable levels of, say, 8.0, thereby providing a partially stabilised water. Even so, for smaller

water treatment system stabilisation using lime remains notoriously problematic and difficult to control, and

the reality is that most lime based stabilisation processes are ineffective. In addition, the increasing limited

availability of high quality (white) lime locally is resulting in increased operating chemical costs at water

treatment facilities. An alternative approach is partial stabilisation using limestone. In this case study,

reference is made to Sedibeng Water’s proposed 110 ML/day Fika Patso Water Treatment Plant, Qwa-Qwa

(Mackintosh et al, 2000) .

Figure 3: Lime dosing system at Sedibeng Water’s 5.5 M:/day Makwane Water Treatment Plant, QwaQwa

•1 mFigure 4: Limestone contactors, Stellenbosch4.2 Comparative Technical Assessment of Lime and Limestone Mediated Stabilisation Partial stabilisation has been shown to be effective in preventing cement aggression, copper corrosion and greatly

reducing corrosion of any ferrous material in the water system (De Souza et al, 2002). On-site pilot plant tests at

Fika Patso were carried out. Chlorinated Fika Patso Dam water is soft, and aggressive and corrosive and would

benefit from stabilisation. The trials showed that the limestone stabilisation process was shown to be capable of

bringing about effective partial stabilisation within a retention time of about 10 minutes. CCDP was effectively

reduced to about 2 mg/L as CaCO3, and pH increased to desirable levels of about 8.2. Importantly, the limestone

system has the ability to handle fluctuations in water quality, as often recorded at the Fika Patso Dam.

Limestone mediated stabilization has been shown to have the following technical sustainability advantages over

lime:

o Very little operator input. (Approximately monthly flushing and chemical recharge, versus the round

the clock attention that lime dosing requires).

o Very little operator skill required. (pH is controlled naturally at desirable levels as the water

approaches chemical equilibrium, versus the carbonate chemistry skill required to maintain steady

lime based stabilisation).

o Very robust equipment, which requires little maintenance, versus lime dosing equipment, which is

generally problematic on small water treatment plants.

o No risk of alkali overdosing.

4.3 Comparative Financial Assessment of Lime and Limestone Mediated Stabilisation

The cost comparison was based on a water treatment plant capacity of 110 ML/day, using November 2000 prices.

Costs considered included Capital costs, Chemical costs, Routine labour costs, Preventative and line maintenance

equipment costs, and Preventative and line maintenance labour costs. The total installed cost of the limestone

contactor was estimated to be the same as that of a conventional lime dosing facility (R 7 700 000). However, as

seen in Figure 5, the running costs associated with limestone stabilisation are significantly lower.

Rozendal, 6 ML/day

• Height 5.1 m

• Diameter 4.5 m

Jonkershoek, 2.5 ML/day

• Height 4.1 m

• Diameter 3.9 m

Figure 5: Operational Cost Comparison: total stabilisation costs

5. CONCLUSION

This paper set out to illustrate the importance of addressing financial and technical sustainability considerations

during the selection of small water treatment systems. The paper also set out to illustrate that under some

circumstances a “low tech” solution will be the better solution, from both a financial and technical perspective; and

yet however, almost paradoxically, under other circumstances a “high tech” solution can be both technically and

financially preferable.

From analysis of both case studies, the following points should be noted:

Wrt financial sustainability:

o Considering that South African rural communities are required to guarantee the payment of the operational

costs of water treatment whilst the capital costs are usually covered by a funding organization, it is crucial

to consider the perspective of community financial sustainability ie ongoing operational costs.

o The NPV based assessment ensures that a balanced assessment of project finance is acquired; importantly,

capital costs are seen in respect of running costs. In the case of the higher capital cost of the membrane

based process, it was shown that there is very little difference in financial performance between the two

technologies when compared over ten years.

Wrt technical sustainability:

o Technological sustainability must be seen in the light of the ability of local community based plant

operators to provide a consistently desirable level of water quality. Where regular malfunction of the

conventional plant occurs, with time the technology will become unsustainable from a socio-political

perspective.

In conclusion, technical experts (such as engineers and hydrogeologists) must carefully consider technology choice

with regards to technical sustainability, financial sustainability, environmental sustainability and socio-political

sustainability. If any one of these “legs of sustainability” is defective, the overall project viability will become

suspect.

6. ACKNOWLEDGEMENTS

The author would like to acknowledge the generosity of Suurbraak Town Council and Sedibeng Water for allowing

publication of the operational findings.

7. REFERENCES

De Souza, P.F. and Mackintosh, G.S. (2000) A Critical Assessment of a Membrane-based Package Plant for Small-

User Systems Water Treatment. Conference proceedings, WISA 2000 Biennial Conference, Sun City, South Africa,

28 May – 1 June 2000.

De Souza, P.F; Manxodid, T and Mackintosh, G.S. (2002) Addressing the Efficiency of Aggression and Corrosion

Mitigation via Limestone Contactor Mediated Partial Stabilisation; Conference proceedings, WISA 2002 Biennial

Conference, Durban, South Africa, 20 May – 23 May 2002.

Government Gazette, South Africa (1997) Water Services Act (Act 108 of 1997).

Mackintosh, G.S. and De Souza, P.F. (2001) A Critical Assessment of the Suitability of a Membrane-based Package

Water Treatment Plant for Application at Suurbraak. CSIR Report no. ENV-S-C 2001-039.

Mackintosh, G; Du Plessis, G; and De Souza, P.F. (2000) On-Site Assessment Of The Suitability Of Limestone

Mediated Stabilisation For Application At Fika Patso Dam.. CSIR Report no. ENV-S-C 2000-134.

South African Bureau of Standards SABS 241:2001: South African Standard for Drinking Water (2001) Pretoria,

South Africa.