Fractionation of wastewater characteristics for modelling of Firle Sewage Treatment Works, Harare, Zimbabwe

Accepted Manuscript Fractionation Of Wastewater Characteristics For Modelling Of Firle Sewage Treatment Works, Harare, Zimbabwe Simon Takawira Muserere, Zvikomborero Hoko, Innocent Nhapi PII: DOI: Reference:

S1474-7065(14)00090-4 http://dx.doi.org/10.1016/j.pce.2014.11.005 JPCE 2318

To appear in:

Physics and Chemistry of the Earth

Received Date: Revised Date: Accepted Date:

1 March 2014 21 October 2014 10 November 2014

Please cite this article as: Muserere, S.T., Hoko, Z., Nhapi, I., Fractionation Of Wastewater Characteristics For Modelling Of Firle Sewage Treatment Works, Harare, Zimbabwe, Physics and Chemistry of the Earth (2015), doi: http://dx.doi.org/10.1016/j.pce.2014.11.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

FRACTIONATION OF WASTEWATER CHARACTERISTICS FOR MODELLING OF FIRLE SEWAGE TREATMENT WORKS, HARARE, ZIMBABWE Simon Takawira Musererea,b, Zvikomborero Hokoa, Innocent Nhapic a

Civil Engineering Department, University of Zimbabwe, P.O Box MP167 Mt Pleasant Harare [email protected], Cell +263-772-338899 b Harare Water Department, City of Harare, 2nd Floor Old Mutual House Corner Sam Nujoma and Speke Avenue, Harare, Zimbabwe, e-mail [email protected], Cell +263-773-142217 c Department of Environmental Engineering, Chinhoyi University of Technology. P. Bag 7724, Chinhoyi, Zimbabwe, Cell +263-733-414529, [email protected]

ABSTRACT Varying conditions are required for different species of microorganisms for the complex biological processes taking place within the activated sludge treatment system. It is against the requirement to manage this complex dynamic system that computer simulators were developed to aid in optimising activated sludge treatment processes. These computer simulators require calibration with quality data input that include wastewater fractionation among others. Thus, this research fractionated raw sewage, at Firle Sewage Treatment Works (STW), for calibration of the BioWin simulation model. Firle STW is a 3-stage activated sludge system. Wastewater characteristics of importance for activated sludge process design can be grouped into carbonaceous, nitrogenous and phosphorus compounds. Division of the substrates and compounds into their constituent fractions is called fractionation and is a valuable tool for process assessment. Fractionation can be carried out using bioassay methods or much simpler physico-chemical methods. The bioassay methods require considerable experience with experimental activated sludge systems and associated measurement techniques while the physico-chemical methods are straight forward. Plant raw wastewater fractionation was carried out through two 14-day campaign periods, the first being from 3 to 16 July 2013 and the second was from 1 to 14 October 2013. According to the Zimbabwean Environmental Management Act, and based on the sensitivity of its catchment, Firle STW effluent discharge regulatory standards in mg/L are COD (<60), TN (<10), Ammonia (<0.2), and TP (<1). On the other hand Firle STW Unit 4 effluent quality results based on City of Harare records in mg/L during the period of study were COD (90 ± 35), TN (9.0 ± 3.0), Ammonia (0.2 ± 0.4) and TP (3.0 ± 1.0). The raw sewage parameter concentrations measured during the study in mg/L and fractions for raw sewage respectively were as follows total COD (680 ± 37), slowly biodegradable COD (456 ± 23), (0.7), readily biodegradable COD (131 ± 11), (0.2), soluble unbiodegradable COD (40 ± 3), (0.06), particulate unbiodegradable COD (53 ± 3) (0.08), total TKN (40 ± 4) mg/L, Ammonia (28 ± 6), (0.68), organically bound Nitrogen (12 ± 2), (0.32), TP (15 ± 1.4), Orthophosphates (9.6 ± 1.4), (0.64), and organically bound TP (5.4 ± 1.4), (0.36), soluble unbiodegradable TP (0.4 ± 0), (0.03), particulate unbiodegradable TP (0.05 ± 0), (0.003). Thus, wastewater at Firle STW was found to be highly biodegradable suggesting optimisation of biological nutrient removal process will generally achieve effluent regulatory standards compliance. Thus, opportunities for plant optimisation do exist of which modelling with the use of a simulator is recommended to achieve recommended effluent standards in addition to reduction of operating costs.

Key words: Biodegradable, Biological Nutrient Removal; Firle Sewage Treatment Works; Fractionation; Model; Simulator. 1. Introduction Globally, close to two million tonnes of sewage wastes are discharged into the world’s waterways resulting in deaths of at least 1.8 million children who are below five years-old every year from water related diseases according to a joint statement by UNEP and UN-HABITAT (Corcoran et al., 2010). In the WHO/UNICEF joint monitoring report it was reported that wastewater management in developing countries throughout the world is in a state of crisis. The report estimated that 2.6 billion people worldwide are living without adequate sanitation WHO (2010). Thus, poor wastewater management is contributing to water borne diseases and environmental degradation including limiting use and complicating drinking water treatment. The uncontrolled wastewater discharge into water sources is against a background of declining per capita water worldwide with Africa being the world’s second-driest continent coming after Australia according to World Water Atlas (UNEP, 2010). Africa accounts for 15 per cent of the global population yet contains only 9 per cent of global renewable water resources reports the World Water Atlas. The situation is worsened by deteriorating water quality mainly due to poor wastewater management. Previous research carried out in 2002 on Zimbabwe’s capital city pointed out poor wastewater management as one of the major reasons for water scarcity in the city according to Nhapi and Gijzen (2002). The research findings pointed out that since Harare’s raw water sources are downstream of wastewater treatment plants strict adherence to recommended wastewater discharge standards is imperative. An assessment of 2000 to 2012 Harare Water Quality Laboratory wastewater effluent results indicates results are generally in the highly environmental hazard category as reported by Muserere et al., (2014). Researchers and scientists responded to these wastewater management challenges and came up with mathematical models in the early 1980s to assist in the operation and process control for sewage works. The International Water Association (IWA) became a key player in the development of wastewater modelling. In 1986 IWA established a task group on mathematical modelling for design and operation of biological wastewater treatment systems. This gave rise to the development of a family of Activated Sludge Models such as ASM1 (Henze et al., 1987), ASM2, ASM2d (Henze, 1999) and ASM3 (Gujer et al., 1999). The development of these families of activated sludge models was supported by powerful computers and detailed knowledge on the metabolism of related bacterial groups. Today models are an important tool in the operation of most sewage treatment plants in developed and some developing countries. Around 1995, the global trends led to the establishment of the IWA Specialist Group on Good Modelling Practice (Meijer and Brdjanovic, 2014). Computer based process simulations are useful tools for plant optimisation, design, analysis, trouble-shooting, operator decision making and operator training (van Loosdrecht et al., 2008). Thus, using a simulator makes linkages of various unit

operations according to the flow regimes of a particular plant easier. The simulator then mimics the plant performance facilitating optimisation based on operational and influent loading conditions (Water Environmental Research Foundation, 2003). Particular attention should be given to the configuration and calibration of the plant simulator to avoid erroneous results. In the use of a simulator when predictions do not reflect observed behaviour one of the potential reasons is unusual or poorly fractionated wastewater characteristics according to Water Environmental Research Foundation (2003). A good example is the unbiodegradable particulate COD fraction which is adjusted to match sludge production quantities with model predictions as suggested by van Loosdrecht et al., (2008). Fractionation is the division of the substrates and compounds into their constituent fractions which is a valuable tool for process assessment according to Henze (1992). Such process assessments include balancing the plant nitrification capacity and denitrification potential as suggested by Ekama and Wentzel (2008). The denitrification potential is a function of slowly biodegradable COD and the portion of readily biodegradable COD which is not utilised in anaerobic phosphate release as demonstrated by Ekama and Wentzel (2008) in their denitrification potential equations. According to their analysis nitrification capacity is the difference between influent TKN and the sum of treated effluent TKN and wasted sludge TKN. Another important Firle STW fractional assessment is organically bound TKN fraction since settled sewage is treated. The fraction is important based on previous research that has demonstrated that organically bound TKN is hydrolysed to ammonia beyond primary settlement according to Ekama et al., (1984). On the other hand phosphorus is divided into orthophosphates and organically bound phosphates which are key modelling parameters according to Ekama et al., (1984). While orthophosphates are obviously linked to phosphate uptake the organically bound fraction has an unbiodegradable fraction which passes through the plant unchanged. The organically bound biodegradable fraction is hydrolised within the system. Slowly biodegradable COD is not linked to anaerobic phosphate release even though it can be hydrolysed to readily biodegradable COD; therefore the influent readily biodegradable COD fraction is critical for optimum phosphate removal according to Ekama and Wentzel et al. (1999). Therefore, given the current state of wastewater management in Zimbabwe, adaptation/development of innovative approaches to wastewater management is an urgent issue. The modelling approach provides wide variants to plant optimisation in relation to the current traditional design and operations in practice. The ideal approach to optimise Firle STW with an existing treatment process is through adjusting the process operating conditions. The study was carried out in the period July to October 2013 at Firle Sewage Works in Harare, Zimbabwe. This is part of a plant optimisation study being carried out at Firle STW. The main objective of the study is to investigate opportunities for optimizing nutrient removal at Harare’s Firle STW through the use of a model. The study is broken down into four specific objectives namely (i) General characterisation of the raw sewage to check quality and treatability by activated sludge processes; work on this objective was completed and already published, (ii) Assess the current performance and efficiency of the system at Firle STW (COD and nutrient removal to identify problem areas) work on this objective is in progress, (iii) Fractionation of

wastewater characteristics and calibration of the model, this paper is addressing the fractionation part of this objective (iv) application of BioWin modelling tool to determine optimum treatment conditions for COD and nutrient removal (assess the treatment processes at each stage, its calibration and validation) this will be the final stage of the research. 2. Study area description 2.1 Location, population and socio-economic issues Zimbabwe is located in the Southern Africa Region between latitudes 160 and 21 south and longitudes 25 0 and 330 east. The country is bounded by Zambia to the north, Mozambique to the east, South Africa to the south and Botswana to the west. Today Harare is the capital of Zimbabwe and most populous city and economic hub of the country. 0

The total population of Zimbabwe was estimated at 12.97 million (Zimbabwe National Statistical Agency, 2012). Of this population, approximately 1.5 million are in Harare constituting approximately 11.5 % of the country’s population. Zimbabwe has two climatic regimes, a rainy season from November to March and a dry season from April to October with cold weather May to July. In 2013 Charles Robertson Renaissance Capital global chief economist, claimed the country’s sovereign debt is just over 100 % of the GDP. According to NEPAD website on 19 June 2013, “the municipality of Harare is going through financial challenges and frequently fails to supply water to its residents, with cholera epidemic in 2008 and water borne diseases on the increase”. Thus, it is critical to provide sustainable water and wastewater management systems to reduce the earnings and expenditure gap of Harare residents. 2.2 Wastewater management in Harare Sanitation systems consists of water borne reticulated systems collected to mainly central wastewater treatment systems and onsite sanitation systems. The formal onsite sanitation systems are mainly septic tanks and soakaways and these have been generally permissible to medium and low density areas. The sewer network of Harare consists mainly of reinforced concrete pipes, PVC pipes and steel pipes. The total length of sewer network is approximately 4,500 km with 75 % of Harare on sewer reticulation and the rest on septic tanks according to Harare Sewage Master Plan Report by SAFEGE, (2003). Following over a decade of economic meltdown, budget allocation for operation and maintenance, and expansion for sewage infrastructure has been very limited. As such the infrastructure is in a state needing urgent repairs and rehabilitation. Sewers have collapsed in a number of areas in Harare and raw sewage spills into the environment. Most of the sewage works are not operating fully and show clear signs of neglect and now have inadequate capacity. The sewage treatment plants in Harare are overloaded with total design capacity of 219,500 m3/d while current inflows average 287,000 m3/d. Firle STW is the largest sewage treatment plant in Harare with a design capacity of

144,000 m3/d with its catchment situated to the South West of downtown Harare (Fig. 1). The current state of affairs has lead to massive environmental degradation, including deterioration of raw water for drinking water treatment. The poor wastewater management has been linked to the cholera outbreak of 2008/9 which killed around 4,300 people in Zimbabwe as reported by UNICEF (2010).

Fig. 1: Location map of Firle Sewage Treatment Works 2.3 Firle STW Unit 4 Firle STW Unit 4, the specific study area in this research, is an activated sludge system specified to a Dry Weather Flow (DWF) capacity 18,000 m3/day and a Wet Weather Flows (WWF) capacity is 54,000 m3/day. Firle Unit 4 is part of Firle STW which is composed of 5 units of similar size and Harare’s largest sewage works. The Unit consists of preliminary treatment stage (screens and grit removal system), 4 Primary Settling Tanks (PSTs), a 3 stage activated sludge reactor basin and 3 clarifiers. The three main sections of the bioreactor are (i) anaerobic zones (a) and (b) each with a volume of 1,400 m3, (ii) combined anoxic/aerobic zone with a total volume of 10,300 m3, a single basin without any physical subdivision. The basin consists of two fermentation tanks followed by ten equal pockets, the two pockets closest to the fermentation zone are the anoxic zone and the remaining 8 pockets with 55 kW surface aerators make up the aerobic zone. The oxygen input can be varied by both automatically controlled switching to the aerators and by manually controlled weir which vary aerator immersion depth. Concrete baffle walls have been constructed and draught tubes installed beneath each aerator to prevent vortexing, reduce short-circuiting, improve mixing and reduce wave action. The clarifiers are an

integral part of the complete treatment process; they separate suspended solids in the mixed liquor from the final effluent. 3. MATERIALS AND METHODS 3.1 Location of sampling points Samples were taken from division box upstream of PSTs (S2) influent wastewater and downstream of clarifiers (S9) treated effluent (Fig 2). The sampling point S2 was upstream of PSTs for influent wastewater and downstream of intake works to exclude grit from the sample, sampling point S9 was after clarifiers for final effluent to measure soluble unbiodegradable matter. Sampling points S1, S3 to S8 are routine monitoring plant sampling points.

Fig. 2: Schematic diagram of Firle Unit 4 flow diagram and sampling points (S1...... S9 are sampling points) adopted from Muserere et al. (2014).

3.2 Sample collection and preservation The recommended two week period monitoring campaigns were used, and repeated under different operating conditions as recommended by Dold and Marais (1986). A total of 28 daily composite samples were collected and analysed using City of Harare and Zimlab facilities. The monitoring campaign involved collecting daily composite samples of influent wastewater and treated effluent then analyse for model state variables, combined variables and influent wastewater fractions of the parameters. Sampling was carried out in the periods 3 – 16 July 2013 (Zimlab

facilities used) and 1 – 14 October 2013 (City of Harare facilities used). Samples were stored below 4 0C in a cooler box on site and in a refrigerator in the lab respectively. 3.3 Selection of parameters The parameters selected for analysis were COD, TN and TP. The COD parameter was selected due to its consistence basis of describing activated sludge process according to Orhon and Cokgor (1997). The other parameters TN and TP are the macronutrients to be removed during treatment to minimum concentrations. According to Water Environmental Research Foundation (2003 the data requirements for plant modelling is case specific, hence for raw and final influent: average daily flows, COD, COD (GF filtrate), ffCOD (ff – flocculated and filtered through 0.45 µm filtrate), TKN, TKN (GF filtrate), Ammonia, TP, TP (GF filtrate), Ortho-Phosphates (PO43-- P), were considered. 3.4 Fractionation 3.4.1 COD The raw sewage Chemical Oxygen Demand was divided into biodegradable and unbiodegradable components (Fig. 3). The biodegradable organic material was further divided into readily biodegradable and slowly biodegradable fractions. The subdivision as suggested by Dold et al. (2003) was based on the dynamic response observed in activated sludge systems. The fractionation is related to bi-substrate modelling approach which identifies that there is an observed order of magnitude difference between the degradation rates of these two fractions, hence it is a biokinetic one according to Dold et al (2003). This division has a crucial role in process dynamic behaviour and plays a major role in the design of biological nutrient removal systems as suggested by Henze (1992). Although Fig. 3 illustrates subdivision beyond readily biodegradable and slowly biodegradable for clarity purposes, further fractionation into the complex, SCFASA, particulate and colloidal fractions were not necessary for the purposes of this research.

Fig. 3: Wastewater division of Total Influent COD into components: adopted from Metcalf and Eddy Inc (2003)

3.4.2

Total nitrogen fractionation

Raw sewage characterisation of the nitrogenous material was in terms of TKN (Fig. 4) as suggested by Water Environmental Research Foundation (2003. Fractionation was therefore carried out in two branches one for free and saline ammonia and the other for organically bound nitrogen (Fig. 4). The fractionation was for raw wastewater hence the assumption that nitrate and nitrite concentrations are negligible was reasonable according to Ekama and Marais (1984). Samples analysed were checked for nitrite and nitrate and was found to be absent which was confirmed by City of Harare laboratory historical records. Further division of organically bound nitrogen into biodegradable and unbiodegradable fractionation was theoretical. Thus, biodegradable and unbiodegradable organically bound fractions were calculated from literature since it is practically difficult to analyse these fractions according to Water Environmental Research Foundation (2003..

Fig. 4: Division of Nitrogen into constituent fractions: adopted from Metcalf and Eddy Inc (2003)

3.4.3 Total Phosphorus fractionation Fractionation of raw sewage was carried out on phosphorus. Most environmental regulations worldwide including the Environmental Management Act of Zimbabwe stipulate phosphorus levels below 1 mg/L. Thus, a rigorous phosphorus analysis is often required especially in light of the low effluent phosphorus concentrations requirements as suggested by Gauthier (2012). However, in this study the division of wastewater phosphorus into its fractions was carried out according to Fig. 5 which was adequate for the purposes of the model being

developed. Thus, assumptions based on Water Environmental Research Foundation (2003) were adopted to estimate phosphorus fractions at lower levels.

Fig. 5: Division of wastewater phosphorus into constituent fractions: (Metcalf and Eddy Inc., 2003) 3.5 Sampling method The six criteria for quality data collection were followed i.e. collecting representative samples, formulating the objective of the sampling program, proper handling and preservation of samples, tracking chain of custody, sample ID procedures, field quality assurance and proper analysis Tjandraatmadja et al. (2009). Flow-weighted composite samples were taken over 24 hrs on each sampling day. A calibrated beaker tied to a 2 meter long steel rod was used to collect samples at the mid-depth of each sampling point. The beaker was rinsed with acidified water first then three times with sample water before final sample collection, an effective way of eliminating sample contamination. The sampling containers were filled to capacity and tightly closed. To check quality control, duplicate samples were systematically send to the lab. Samples were taken to the lab within an hour after taking the last sample fraction and were preserved by keeping them in a freezer at temperatures below 4 o C in accordance with Standard Methods for the Examination of Water and Wastewater by American Public Health Association, American Water Works Association and Water Environment Federation (APHA-AWWA-WEF, 2005). 3.6 Testing methods

Sample analyses were in accordance with APHA-AWWA-WEF (2005). COD was analysed in accordance with APHA (5220), TKN in accordance with APHA (4500-N), while phosphates were determined according to APHA (4500-P). The experimental protocols required for bioassay methods calls for close attention. Thus, considerable experience with experimental activated sludge systems and associated measuring techniques are required. Because of the expertise required in bioassay methods, alternative physicochemical methods were employed to fractionate the parameters. Guidelines for wastewater characterisation developed by the Dutch Foundation for Applied Water Research (STOWA) were used for fractionation as recommended by Roeleveld and Loosdrecht (2002). These methods were used to analyse readily biodegradable COD, soluble unbiodegradable COD, soluble readily biodegradable, particulate unbiodegradable COD, slowly biodegradable COD, Nitrogen compounds and Phosphorus compounds. 4. RESULTS The mean concentrations, standard deviations and coefficients of variation were calculated for raw sewage and treated effluent (Table 1). Table 1: Statistical analysis of wastewater parameter concentrations for the periods 3 – 16 July and 1- 14 October 2013: Firle STW Unit 4 2013 Parameter

Raw sewage Total COD (unfiltered) 1 COD (Glass filtered) COD (0.45µm filtrateinfluent) Total TKN Ammonia Total Phosphorus Ortho-phosphates Treated effluent 2 COD (0.45µm filtrateeffluent)

Mean concentrations mg/L

Conc. Range mg/L

Standard Coefficient Deviation of Variation

N

680 456 171

594 - 710 436 - 490 156 - 188

37 23 11

0.05 0.05 0.06

13 10 8

40 28 15 9.6

37 - 43 22 - 35 13 – 17 7 – 11

4 6 1.4 1.4

0.1 0.2 0.09 0.15

11 12 12 12

40

38 - 45

3

0.08

9

1. Conc. Is concentration 2. N is the number of samples

Based on lab analysed results as reported in Table 1 the wastewater fractions for Firle Sewage Works were then calculated (Table 2).

1

The terms “readily” and “slowly” biodegradable are taken as synonymous with the terms “soluble” and “particulate”

2

The concentration of readily biodegradable COD in the effluent of a properly operating activated sludge is negligible

Table 2: Table showing wastewater mean concentrations and fractions: Results for the periods 3 – 16 July and 1- 14 October 2013 (adopted from Water Environmental Research Foundation (2003). Wastewater characteristic

Analysed Conc

Typical Municipal Conc

Conc. Units

680

250 – 700

Slowly biodegradable COD Readily biodegradable COD Soluble unbiodegradable COD 4 Particulate Unbiodegradable COD Nitrogenous Material Total TKN Free and saline ammonia Organically Bound TKN

456

5

Organic Material Total COD

Soluble biodegradable TKN 6 Particulate unbiodegradable TKN 7 Particulate biodegradable TKN

3

4

5

6

7

Analysed Fractions

Typical Fractions3

Fraction Units

g COD m-3

1

1

g COD /g of total COD

200 – 400

g COD m-3

0.7

0.4 – 0.80

g COD /g of total COD

131

25 – 125

g COD m-3

0.2

0.05 – 0.25

g COD /g of total COD

40

20 – 50

g COD m-3

0.06

0.04 – 0.16

g COD /g of total COD

53

35 – 110

g COD m-3

0.08

0.07 – 0.22

g COD /g of total COD

40 28

30 – 80 20 – 30

g N m-3 g N m-3

1 0.68

0.5 – 0.75

g N /g TKN g N /g TKN

12

25 – 70

g N m-3

0.32

0.25 – 0.5

g N /g TKN

0.3

0–5

g N m-3

0.007

0.4

g N /g TKN

4.3

-

g N m-3

0.1

-

10.4

-

g N m-3

0.24

-

g N/ g particulate unbiodegradable COD g N /g TKN

typical municipal concentrations and fractions adopted from Water Environmental Research Foundation (2003)

Particulate biodegradable TKN was calculated from other fractions Plant was assumed to be fully nitrifying and ammonia concentration in the final effluent was 0.2 mg/L Assume particulate unbiodegradable TKN is 10 % of influent TKN Particulate biodegradable TKN is calculated from the other fractions

Phosphorus Material TP Orthophosphate

Biodegradable organically bound TP Soluble Unbiodegradable TP Particulate Unbiodegradable TP

15 9.6

4 – 15

g P m-3

1

1

g P/ g TP

2 – 12

-3

0.64

0.5 – 0.85

g P/ g TP

-3

gPm

5.4

0 – 10

gPm

0.36

0 – 0.25

g P/ g TP

0.4

0-?

g P m-3

0.03

0 -?

g P/ g TP

0.05

1–4

g P m-3

0.003

0.02 – 0.03

g P/ g particulate unbiodegradable COD

The plant was always operated for 60 hours uninterrupted before sampling commence for steady state conditions. Steady state conditions are attained when plant runs for period equal to five times the plant Hydraulic Retention Time according to Vaiopoulou et al. (2011). 5. DISCUSSIONS 5.1 Total Chemical Oxygen Demand (COD) This section presents raw sewage results. The influent total COD concentration at Firle STW was 680 ± 37 mgCOD/L. Thus, the concentration was in the low to medium range according to Metcalf and Eddy Inc. (2003). The COD concentration range was within the same range as reported by Muserere et al. (2014) for the same plant. Firle STW catchment is predominantly from domestic developments and some industrial developments however; the industry is reported to be at very low production capacity due to economic situation in the country. In IWA based models, characterisation is in terms of COD analysis primarily due to its consistency in activated sludge description, quantifying sludge and oxygen demand according to Water Environmental Research Foundation (2003). However, the relationship of COD to BOD is of critical importance in activated sludge treatment processes. Furthermore, biodegradable fraction has a major influence on process dynamic behaviour especially oxygen demand and denitrification capacity as suggested by Dold et al., (1980). The COD to BOD ratio was 1.2 indicating sewage is predominantly domestic suggesting the wastewater at Firle STW was easily treatable by biological processes according to Ekama and Wentzel (2008). Where the ratio is above 3 it suggest the waste is toxic or acclimated microorganisms may be required in its stabilisation. However, the COD parameter was used against the generally used BOD. Thus, COD was fractionated for use in model calibration according to IWA based models. The results found in this study compared well with 527 mg/L found in the previous research by Muserere et al. (2014) already published. The results confirm raw sewage at Firle STW is generally within the low to medium range concentration based on Metcalf and Eddy Inc. (2003) classification.

5.2 Biodegradable COD The biodegradable COD was measured from raw sewage. The total biodegradable COD fraction was 587 ± 17 mg/L constituting 86% of total COD. This fraction represents the carbonaceous constituents and is a measure of wastewater biodegradability. The high biodegradable component at the plant results in larger aeration basin volume requirements, higher oxygen demand, and higher sludge production as suggested by Dold et al. (2003). According to Marais and Dold (1985) a portion of biodegradable COD is oxidised to carbon dioxide and water in the energy production process for maintaining existing cell mass homeostatic balance. The high biodegradable COD fraction translates to a high oxygen demand for complete oxidation to take place at the plant. This suggests why the plant was equipped with un-tapered 55 kW aerators. However, reference to typical design recommendations by Metcalf and Eddy Inc (2003) suggest 22.5 kW aerators could be adequate for Firle STW aeration basin. The difference between these aerator sizes suggest the plant could be operating at high oxygen concentrations. Thus modelling the aeration processes at the plant is critical for plant optimisation. 5.2.1 Slowly biodegradable COD The slowly biodegradable COD concentration was 456 ± 23 mg/L constituting a fraction of 70% of total COD. The slowly biodegradable COD is critical for denitrification in the anoxic zone, for cell growth and energy in the aeration zone. In the anoxic zone the slowly biodegradable COD fraction is responsible for the slower denitrification rate while the readily biodegradable COD fraction accounts for the more rapid denitrification rate according to van Haandel et al. (1981). In another research Clayton et al., (1991) concluded that the increased rate of denitrification in anoxic zone can result from an increased hydrolysis rate of slowly biodegradable COD albeit induced in the anaerobic zone. The dentrification potential is proportional to slowly biodegradable COD concentration according to Ekama and Wentzel (2008). The high slowly biodegradable COD fraction suggest a high denitrification potential which then explains why the plant complies with nitrogen effluent standards. 5.2.2 Readily biodegradable COD The readily biodegradable fraction was 131 ± 11 mg/L, accounting for 20 % of total COD. Quantifying the readily biodegradable fraction is important for phosphorus removal which is quickly assimilated by biomass in the anaerobic zone as suggested by Ekama and Wentzel (1999). The fraction has a direct impact on the activated sludge biological kinetics and plant performance according to Mamais et al. (1993). The readily biodegradable COD fraction is converted to SCFA by non-polyP organism mass; in this way the SCFA becomes available to the polyp organisms according to (Meganck et al., 1985; Brodisch, 1985; Wentzel et al, 1985). Higher fractions of readily biodegradable COD result in greater amount of biological phosphorus removal according to Ekama and Marais (1984). Another advantage of higher fraction is that it provides more COD for floc forming bacteria in the selector hence a greater impact in improving sludge volume index according to Metcalf and

Eddy Inc (2003). The greater the concentration of readily biodegradable COD the faster is the denitrification process. Thus, the 131 mg/L or 20 % readily biodegradable fraction at the plant was within the recommended range for phosphate release. The parametric model by Siebritz et al. (1982) suggests a concentration range of 70 to 220 mg/L or a fraction of 0.12 to 0.27 readily biodegradable is required for phosphate release. Thus, the plant none compliance with phosphorus effluent standards suggest process control problems. Thus, modelling the plant to establish optimal point is critical. 5.3 Unbiodegradable COD The unbiodegradable component was 93 ± 3 mg/L constituting 14 % fraction of the total COD. Marais and Dold (1985) hypothesised that the particulate unbiodegradable portion of COD is used for new cell mass production. 5.3.1 Soluble unbiodegradable COD The unbiodegradable soluble COD was measured from treated effluent. The soluble unbiodegradable COD fraction was 40 ± 3 mg/L accounting for 6% of total COD concentration. The soluble unbiodegradable COD was measured in the effluent. The soluble unbiodegradable COD has no impact on activated sludge treatment as it passes through unchanged. For domestic wastewater with small industrial contribution unbiodegradable soluble COD is equal to the effluent soluble COD concentration for an activated sludge process at Sludge Retention Time (SRT) greater than 4 days according to Metcalf and Eddy Inc (2003). Firle STW is reportedly operating at SRT of 14.8 days, thus the assumption that the effluent soluble unbiodegradable COD was equal to the influent soluble COD was reasonable. The recommended final effluent COD concentration is 60 mg/L suggesting why the plant effluent COD concentration was compliant. 5.3.2 Particulate unbiodegradable COD The unbiodegradable COD was measured from raw sewage. The particulate unbiodegradable COD concentration was 53 ± 3 mg/L constituting 8% of total COD. This fraction contributes to sludge production and Mixed Liquor Suspended Solids (MLSS) in the Biological Nutrient Removal plant reactors. Thus, this fraction in model simulations accounts for sludge quantities differences between measured and predicted results. The unbiodegradable particulate COD constitute part of the organic material which contribute to Volatile Suspended Solids (VSS) concentration of the wastewater and mixed liquor in the activated sludge. The concentration of particulate unbiodegradable COD was within the range for municipal sewers of 35 to 110 mg/L according to Water Environmental Research Foundation (2003).

5.4 Total Kjeldahl Nitrogen (TKN) The concentration of total TKN was 40 ± 4 mg/L. Total nitrogen is composed of organic nitrogen, ammonia, nitrate and nitrite. The TKN parameter measures the sum of organic nitrogen and ammonia according to van Loosdrecht (2008). The concentration of nitrate and nitrite at Firle STW were negligible since the sewage arrived at the plant fresh and any nitrate formed was possibly denitrified. The concentrations of nitrite and nitrate were measured and found to be absent in all samples analysed, furthermore, records of City of Harare plant routine tests confirmed this. Thus, TKN concentration was equal to total nitrogen in raw sewage. The 40mg/L TKN concentration was within the normal municipal sewer range which is 32 to 70 mg/L Water Environmental Research Foundation (2003). 5.4.1 Ammonia Free and saline Ammonia concentration was 28 ± 6 mg/L constituting 68 % of total TKN. Approximately 60 to 70 % of the influent TKN concentration is free and saline ammonia according to Metcalf and Eddy Inc (2003). They suggested that free and saline ammonia is readily available fraction for bacterial synthesis and nitrification. The plant was operated at sludge age of 15 days which was above 4 days the minimum nitrifying sludge age, hence with adequate dissolved oxygen nitrification should complete at the plant. 5.4.2 Organically bound Nitrogen The organically bound Nitrogen was 12 ± 2 mgN/L constituting 32 % of TKN. The organically bound fraction is available in soluble and particulate form with portions of each being either biodegradable or unbiodegradable. The particulate degradable organic nitrogen fraction is removed more slowly than the soluble degradable nitrogen fraction which requires undergoing hydrolysis reaction first Ekama and Marais (1984). The particulate unbiodegradable nitrogen is captured in the activated sludge floc and exits via waste sludge, while the soluble unbiodegradable will be found in final effluent Ekama and Wentzel (2008). Practically it is not possible to separate soluble biodegradable and unbiodegradable fractions Water Environmental Research Foundation (2003). The method used to separate them was based on Metcalf and Eddy Inc. (2003) who suggest that the unbiodegradable fraction in total nitrogen concentration in the effluent is a small fraction less than 3 %. The soluble unbiodegradable organic nitrogen concentration in domestic wastewater is within the range of 1 to 2 mg/L as N according to Perkin and McCarty, (1981). Firle STW has predominantly domestic sewage. Literature suggests that some soluble unbiodegradable organic nitrogen may be produced from endogenous respiration as suggested by Metcalf and Eddy Inc. (2003). Thus, with plant set for optimum nitrification capacity and a high denitrification potential from high biodegradable sewage the plant is expected to comply under optimum conditions.

5.5 Total Phosphorus (TP) The total phosphorus concentration was 15 ± 1.4 mg/L, Phosphorus is essential for algae growth and other biological organisms hence there is interest in controlling the amount of phosphorus discharged into the surface water (Metcalf and Eddy Inc. 2003). The medium to high concentration of phosphorus at the plant is attributed to less stringent conditions in the catchment, there is no phosphate ban. Domestic wastewater has high inorganic phosphorus concentration contributed by use of synthetic detergents and human waste according to Sawyer et al. (1994). Thus, a ban in phosphate detergents is required. The concentration is above design specification of 11 mg/L by the process designer hence plant optimisation in addition to enactment of more stringent regulations are required. 5.5.1 Orthophosphates The concentration of orthophosphates was 9.6 ± 1.4 mg/L accounting for 64% of total Phosphorus. The soluble portion of the influent total phosphorus especially orthophosphates, is important when modelling phosphorus removal in PSTs Water Environmental Research Foundation (2003). The foundation suggest that the division into orthophosphate and organically bound phosphorus is not important when modelling activated sludge system only since particulate organic phosphorus is hydrolised to orthophosphate in the process. 5.5.2 Organically bound TP The concentration of organically bound TP was 5.4 ± 1.4 mg/L constituting 36% of TP. This fractions determine sizing, number of anaerobic reactors needed, the need for primary sedimentation and achievable maximum phosphate removal Water Environmental Research Foundation (2003). A proportion of 36% organic phosphorus to 67% inorganic phosphorus was suggested by Metcalf and Eddy Inc. (2003) while Dold et al. (2013) suggest a ratio of 15 % organic phosphorus to 85 % orthophosphates is generally found in municipal wastewater. The fractions for organically bound fraction that are biodegradable and unbiodegradable fractions were calculated from literature (Table 2). Thus the 0.64 orthophosphates to 0.36 organically bound TP ratio at Firle STW is within municipal wastewater range. The 4.95 mg/L concentration of organically bound phosphorus suggests the final effluent TP-concentration may not be below the recommended 1 mg/L hence plant optimisation is important. 6. CONCLUSIONS The Firle STW raw sewage fractions were generally wastewater ranges found by other researchers. The wastewater highly biodegradable a condition favourable for biological nutrient was concluded that with highly biodegradable wastewater treated

within municipal was found to be removal. Thus, it at Firle STW the

treated effluent should generally comply with recommended standards for COD, TN and TP parameters.

7. RECOMMENDATIONS Calibration of the model using these fractions is recommended and subsequent use of the BioWin modelling tool to inform operations at the plant. The use of a model has proven to improve plant efficiencies in other activated sludge plants thus reduction in operational costs is anticipated. 8. ACKNOWLEDGEMENTS This paper presents part of the research results of an MPhil study by Simon Takawira Muserere at the University of Zimbabwe. The authors thank the City of Harare especially the Harare Water Department for allowing this study at Firle Sewage Works and for the assistance in sample collection and analyses. Finally the authors are grateful to the participants at the 14th WaterNet Symposium for the comments that added value to the paper. 8. REFERENCES APHA-AWWA-WEF, 2005. Standard Methods for the Examination of Water and Wastewater, twenty first ed. American Public Health Association, America Water Works Association, and Water Environmental Federation, New York. Brodisch, K.E.U., 1985. Interaction of different groups of microorganisms in biological phosphate removal. Water Science and Technology, 17(11-12), 139-146. Clayton, J.A., Ekama, G.A., Wentzel, M.C., Marais, G.V.R., 1991. Dentrification kinetics in biological nitrogen and phosphorus removal in activated sludge systems treating municipal wastewaters. Water Science and Technology, 23 (4/6), 1025-1035. Corcoran, E., Nellemann, C., Baker, E., Bos, R., Osborn, D., Savelli, H., 2010. Sick Water? The central role of wastewater management in sustainable development, A rapid response assessment. United Nations Environment Programme, UN-HABITAT, GRID-Arendal, Norway. Dold, P.L., Jones, R.M., Takacs, I., Melcer, H., 2003. Practical guidance for WWTP model calibration and associated data gathering requirements, Proceedings of the Water Environmental Federation WEFTEC 2003: Session 1 through Session 10. Water Environmental Federation, pp 202-224(23). http://www.ingentaconnect.com/content/wef/wefproc/2003/00002003/0000001 2/art00013?crawler=true accessed February 2014 Dold, P.L., Marais, G.V.R., 1986. Evaluation of the general activated sludge model proposed by the IAWPRC task group. Water Science and Technology. 18(6), 63-89

Ekama, G.A., Marais, G.v.R., 1984. Nature of municipal wastewaters. Theory, design, and operation of nutrient removal activated sludge processes. Water Research Commission, Pretoria. Ekama, G.A., Marais, G.v.R., Siebritz, I.P., 1984. Biological excess phosphorus removal, in: Wiechers, H.N.S., Ekama, G.A., Gerber, A. (Eds.), Theory, design and operation of nutrient removal activated sludge processes. Water Research Commission, Pretoria, pp 7.1-7.32. Ekama, G.A., Wentzel, M.C., 1999. Denitrification kinetics in biological N and P removal activated sludge systems treating municipal wastewaters. Water Science and Technology, 39 (6), 69-77. Ekama, G.A., Wentzel, M.C., 2008. Organic material removal, in: Hence, M., van Loosdrecht, M.C.M., Ekama, G.A., Brdjanovic, D. (Eds.), Biological Wastewater Treatment: Principles, Design and Modelling. IWA Publishing, London, pp 53-86. Gauthier, R., 2012. Phosphorus Removal from Wastewater: Phosphorus Chemistry http://www.thewaterplanetcompany.com/docs/10pdf/Phosphorus/20Chemistry .pdf accessed March 2013. Gujer, W., Henze, M., Mino, T., van Loosdrecht, M.C.M., 1999. Activated Sludge Model No. 3. IAWQ Task Group on Mathematical Modelling for Design and Operation of Biological Wastewater Treatment, Water Science and Technology, 39(1), 183-193. Henze, M., Grady, C.P.L.J., Gujer, W., Marais, G.v.R., Matsuo, T., 1987. Activated Sludge Model No. 1. IAWQ Scientific and Technical Report No. 1. International Association on Water Pollution Research and Control, London. Henze, M., 1992. Characterisation of wastewater for modelling of activated sludge processes. Water Science and Technology, 25 (6), 1-15. Henze, M., Gujer, W., Mino, T., Matsuo, T., Wentzel, M.C., Marais, G.v.R., van Loosdrecht, M.C.M., 1999. Activated Sludge Model No. 2d. Water Science and Technology, 39(1), 165-182. Mamais, D., Jenkins, D., Pitt, P., 1993. A rapid physical-chemical method for the determination of readily biodegradable soluble COD in municipal wastewater. Water Research, 27, pp. 195-197. Marais, G.v.R., Dold, P.L., 1985. Biological removal of Carbon, Nitrogen, and Phosphorus in single sludge systems. Advances in Biological Wastewater Treatment Seminar. Instituto di Ricerca Sulle Acque, Rome, pp. 207-233. https://fenix.tecnico.ulisboa.pt/downloadFile/3779571862497/modelacao%20 ETAR-teoria_V03.pdf accessed March 2014 Meganck, M., Malnou, D., Le Flohic, P., Faup, G.M., Rovel, J.M., 1985. The importance of acidogenic microflora in biological phosphorus removal. Water Science and Technology, 17(11-12), 199-212. Meijer, S.C.F., Brdjanovic, D., 2014. A Practical Guide for Activated Sludge Modelling. UNESCO-IHE Lecture notes. UNESCO-IHE Institute for Water Education. Delft. Metcalf and Eddy Inc., 2003. Wastewater Engineering: Treatment and Reuse, fourth ed. MacGraw-Hill Companies Inc., London. Muserere, S.T., Hoko, Z., Nhapi, I., 2014. Wastewater characterisation and Primary Settling Tanks Performance Assessment. Physics and Chemistry of the Earth, 67-69, 226-235. Nhapi, I., Gijzen, H., 2002. Wastewater management in Zimbabwe, Sustainable Environmental Sanitation and Wastewater Services, 28th WEDC Conference,

Calcutta, India. http://www.researchgate.net/profile/Innocent_Nhapi/publication/256943187_W astewater_management_in_Zimbabwe/links/02e7e52416ae7bf7df000000?ori gin=publication_detail. accessed February 2014 Orhon, D., Cokgor, E.U., 1997. COD fractionation in wastewater characterisation, the state of the art. Chemical Technology and Biotechnology, 66, 283-293. Perkin, G.F., McCarty, P.L., 1981. Sources of soluble organic nitrogen in activated sludge effluents. Water Pollution Control Federation, 53, 89-98. Roeleveld, P.J., van Loosdrecht, M.C.M., 2002. Guidelines for wastewater characterisation in the Netherlands. Water Science and Technology, 45(6), 77-87. SAFEGE, 2003. Harare Sewerage Master Plan Final Report, Unpublished Consultancy Report, Harare, Zimbabwe Sawyer, C.N., McCarty, P.L., Parkin, G.F., 1994. Chemistry for environmental engineering, fourth ed. McGraw-Hill Companies Inc., Singapore. Siebritz, I.P., Ekama, G.A., Marais, G.v.R., 1982. A parametric model for biological excess phosphorus removal. Water Science and Technology, 15 (3/4), 127152. Tjandraatmadja, G., Pollard, C., Gozukara, Y., Sheedy, C., 2009. Characterisation of priority contaminants in residential wastewater. Water for a Healthy Country National Research Flagship, CSIRO, Australia. UNEP, 2010. Africa Water Atlas. Division of Early Warning and Assessment, UNEP, Nairobi, Kenya. http://www.unep.org/pdf/Africa_Water_Atlas_Executive_Summary.pdf accessed February 2014 UNICEF, 2010. The 2008-2009 Cholera Epidemic in Zimbabwe, United Nations Childrens Emergency Fund, unpublished report, Harare, Zimbabwe. Vaiopoulou, E., Melidis, P., Aivasidis, A., 2011. Control of bulking sludge caused by Eikelboom type 021N and Thiothrix SPP in a BNR Plant. Deptartment of Environmental Engineering, Democritus University of Thrace, Greece. http://www.srcosmos.gr/srcosmos/generic_pagelet.aspx?pagelet=Author&aut hor_id=652&title=Aivasidis%20A accessed February 2014 van Haandel, A.C., Ekama, G.C., Marais, G.v.R., 1981. The activated sludge process part 3, Single sludge denitrification. Water Research 15, 1135-1152. van Loosdrecht, M.C.M., 2008. Innovative nitrogen removal, in: Henze, M., van Loosdrecht, M.C.M., Ekama, G.A., Brdjanovic, D. (Eds.), Biological Wastewater Treatment: Principles, Modelling and Design. IWA Publishing, London, pp. 139-154. van Loosdrecht, M.C.M., Ekama, G.A., Wentzel, M.C., Brdjanovic, D., Hooijmans, C.M., 2008. Modelling Activated Sludge Processes, in: Henze, M., van Loosdrecht, M.C.M., Ekama, G.A., Brdjanovic, D. (Eds.), Biological Wastewater Treatment: Principles, Modelling and Design. IWA Publishing, London, pp. 361-392. Water Environmental Research Foundation, 2003. Methods for wastewater characterisation in activated sludge modelling, IWA, Water Environmental Federation, New York. Wentzel, M.C., Dold, P.L., Ekama, G.A., Marais, G.v.R., 1985. Kinetics of biological phosphorus removal. Water Science and Technology 17(11-12), 57-71. WHO, 2010. Progress on Sanitation and Drinking Water-Fast facts. WHO/UNICEF

joint monitoring report 2010. WHO, http://www.who.int/water_sanitation_health/monitoring/fast_facts/en/ ccessed February, 2014 Zimbabwe National Statistical Agency 2012. Preliminary Census Report. Zimbabwe. Government Printers, Harare.

HIGHLIGHTS • • • •

We fractionate domestic with industrial contribution raw sewage Fractionation information will aid in plant optimisation Nutrient removal is dependent on readily biodegradable COD The calibrated simulator is meant to optimise nutrient removal