Abstract

The problems that arise from hydrogen sulfide formation in sewer rising mains are often dealt with by adding chemicals. The most commonly used chemicals are sodium hydroxide and magnesium hydroxide, ferric (or ferrous) chloride, calcium nitrate and oxygen. As there is a lack of tools to evaluate and compare dosing strategies, it is hard to identify the most appropriate dosing chemical in terms of investment and operating costs. As a consequence, chemical selection is often done ad-hoc. In this study, the average dosing costs (operating and capital) of these chemicals is assessed by simulations with a dynamic mathematical model, describing the transformations of sulfide, carbon and nitrogen in a sewer rising main. Based on the results of the simulations, and other parameters, a systematic approach to selecting dosing products for sewer rising mains is outlined.

INTRODUCTION

The formation and emission of hydrogen sulfide (H2S) in sewers is associated with odor nuisance and corrosion of concrete sewer infrastructure. The gas often forms in sewer structures with long retention times, such as rising mains or inverted siphons, where anaerobic conditions prevail. Like many other wastewater operators, Aquafin is frequently confronted with H2S-induced problems in its infrastructure, especially downstream of rising mains. In many cases, simple and cheap solutions, such as odor-tight covering or ventilation are not effective. The addition of chemicals to the wastewater on the other hand, is known to be very effective.

A large number of chemicals is used for sulfide control. The most common are sodium and magnesium hydroxide, ferric (or ferrous) chlorides, nitrate, and air or pure oxygen (Apgar et al. 2007; Ganigué et al. 2011). As it is unclear which chemical species is most (cost) efficient, selection is usually an ad-hoc process. The aim of this study was to develop a more systematic approach.

METHODS

Dosing chemicals

Dosing with ferric chloride, calcium nitrate, magnesium hydroxide and sodium hydroxide was considered for sulfide control in rising mains. In Australia, pure oxygen injection is frequently used (Ganigué et al. 2011), but was not considered here because of safety issues. Air injection has not yet proved effective in practical implementations by Aquafin and is only considered if no other options are available.

The optimum dosing product depends on many site-specific conditions including the desired level of sulfide control and dosing locations. However, based on Aquafin's experience, two parameters are key to the choice of chemical, namely the dosing cost (capital and operating costs) and the chemical's hazard classification. The latter determines whether an environmental permit (or a message to the city/municipality for small quantities) is necessary.

Simulations

To evaluate the average dosing quantity of these chemicals, simulations were performed with an in-house sewer process model (Donckels et al. 2014) called Aqua3S. It is basically an implementation of the WATS model developed by Hvitved-Jacobsen et al. (1998), but the stoichiometric and kinetic equations for sulfide formation and oxidation of the SeweX model (Sharma et al. 2008) were used because it was assumed in the original WATS model that no organic substrate is consumed when hydrogen sulfide is formed. Another approach was also followed for incorporating endogenous respiration (similar to ASM3, Henze et al. (2000)) because this simplified extension of the model to include nitrate consuming processes. The Aqua3S and WATS models follow the same methodology as the activated sludge models for the fractionation of the organic matter.

For reasons of simplicity, it was assumed that the COD fractions are constant in the simulations. In practice, however, they are known to vary considerably with time (Vollertsen et al. 2005). The default values used in the simulations for the COD fractions are shown in Table 1. They are reported in the literature (Hvitved-Jacobsen et al. 1998; Vollertsen et al. 2005) and characteristic. The total COD concentration was set at 300 mg/L, corresponding with the average COD concentrations measured in Flanders (Belgium).

Table 1

Default values of variables in the Aqua3S model, used for simulation

Component Description Default value (mg/L) 
SF Fermentable substrate 0.05*COD 
SA Fermentation products (i.e. VFAs) 0.05*COD 
XS1 Hydrolyzable substrate, biodegradable rapidly 0.10*COD 
XS2 Hydrolyzable substrate, biodegradable slowly 0.75*COD 
XBW Heterotrophic active biomass in the water phase 0.05*COD 
SSO4 Sulfate concentration 33 
SH2S Sulfide concentration 
SO Oxygen concentration 
Component Description Default value (mg/L) 
SF Fermentable substrate 0.05*COD 
SA Fermentation products (i.e. VFAs) 0.05*COD 
XS1 Hydrolyzable substrate, biodegradable rapidly 0.10*COD 
XS2 Hydrolyzable substrate, biodegradable slowly 0.75*COD 
XBW Heterotrophic active biomass in the water phase 0.05*COD 
SSO4 Sulfate concentration 33 
SH2S Sulfide concentration 
SO Oxygen concentration 

Only dry-weather conditions were considered in the simulations, the wastewater temperature was assumed to be 18°C, typical of the summer months. Simulations were performed for different rising main dimensions and hydraulic retention times (HRTs). The wastewater flow was assumed constant, for simplicity. The simulated dosing quantities were used to calculate the dosing costs, based on unit prices from suppliers, shown in Table 2.

Table 2

Unit cost (€/liter) of dosing product for sulfide control

Dosing product Description Supplier Unit cost (€/liter) 
Ferric chloride PIX111 Kemira 0.16 
Calcium nitrate CAN51 SNF 0.40 
Sodium hydroxide 29% solution Brenntag 0.70 
Magnesium hydroxide 53% suspension ImChemia 1.94 
Dosing product Description Supplier Unit cost (€/liter) 
Ferric chloride PIX111 Kemira 0.16 
Calcium nitrate CAN51 SNF 0.40 
Sodium hydroxide 29% solution Brenntag 0.70 
Magnesium hydroxide 53% suspension ImChemia 1.94 

Operating costs

Ferric chloride

Iron salts react with dissolved sulfides in the wastewater, forming highly insoluble metallic sulfide precipitates. The dosing cost of ferric chloride was calculated from the precipitating sulfide concentration and assuming a precipitation efficiency of 0.7 mol-Fe/mol-S (Apgar et al. 2007; Firer et al. 2008; Zhang et al. 2009). The precipitating sulfide concentration was calculated from the simulated sulfide concentration at the rising main outlet, diminished by the sulfide concentration that does not precipitate. For the latter, a value of 0.2 mg-S/L was used, as it is reported that ferric chloride cannot affect the sulfide concentration below 0.2 to 0.5 mg-S/L (Apgar et al. 2007). Other (realistic) concentrations were also tried but did not influence the order of dosing product cost prices.

Nitrates

Nitrates have frequently been added to rising mains to prevent anaerobic conditions, thereby controlling the formation of sulfide. However, Jiang et al. (2009) and Mohanakrishnan et al. (2009) demonstrated that it is the oxidation of sulfides by nitrate-reducing, sulfide-oxidizing bacteria that lowers the sulfide concentration. To assess the nitrate dosing cost, nitrate consuming processes were implemented in the Aqua3S model. Both denitrification and anoxic sulfide oxidation were implemented as two-step processes (formation of respectively intermediate nitrite and elemental sulfur). Simulations were performed for different nitrate concentrations in the rising main influent. The minimum nitrate concentration necessary to decrease the sulfide concentration at the outlet of the main below 0.2 mg-S/L, was used to compute the dosing cost, using the derivative-free optimization method. Lower sulfide concentrations can probably be achieved with nitrate dosing, but it was decided to use the same concentrations as for ferric chloride to enable direct cost price comparison.

Ganigué et al. (2011) point out that it is better (in terms of cost and performance) to inject nitrate downstream of the pumping station, close to the rising main outlet, but allowing adequate wastewater retention time downstream of dosing for complete sulfide oxidation by the nitrates. Because of this, simulations were also performed for different downstream nitrate dosing locations. For each scenario, ten equidistant dosing locations were simulated. The location of the optimum (i.e. cheapest) dosing point is a function of both the diameter and the HRT. If the rising main diameter decreases, the optimum dosing location shifts in the direction of its outlet, as the anoxic sulfide oxidation mainly occurs in the biofilm and is more efficient in smaller diameter pipes.

For this study it was assumed that nitrate could be dosed anywhere along the rising main and the minimum cost (i.e. the dosing cost at the optimum dosing location) was determined. However, in practical applications, the optimum dosing location will not always be acceptable for chemical storage (e.g. due to the proximity of houses).

Sodium hydroxide

Shock dosing of sodium hydroxide (for 20 to 30 minutes to increase the wastewater pH to 12 to 13) has been shown to de-activate sulfate-reducing bacteria (SRBs), which are responsible for sulfide formation in sewers, for between a few days and two weeks (United States Environmental Protection Agency 1974). The dosing rate depends on the wastewater's buffering capacity, which is strongly site-specific and time dependent. In a field test in Los Angeles, an average dose of 0.23 g-NaOH/L of wastewater was needed to obtain odor control (Apgar et al. 2007). Aquafin found that an average of 0.63 g-NaOH/L was needed to increase the pH of a WWTP's influent in Tielt, Belgium, to 12. Both of these dosing quantities were used to assess the dosing cost.

Weekly dosing was assumed, as this is the average recovery time reported in literature for the SRBs in the biofilm. It was assumed that the whole volume of the rising main needs to be filled with sodium hydroxide, which is probably unnecessary, due to plug flow.

Magnesium hydroxide

Magnesium hydroxide is added to wastewater to increase the pH to 8.5 to 9. Its solubility is limited and so larger pH shifts cannot be achieved. This limited pH increase is insufficient to inactivate the SRBs completely and hence the product must be dosed continuously, unlike sodium hydroxide, but it prevents sulfide release to the gas phase. Again, the dosing quantity depends on the wastewater's buffering capacity and should be determined on-site. The average dosing quantity was determined in a pilot-scale rising main (50 mm diameter and 160 m long), fed with the WWTP influent at Aartselaar, Belgium. The average dose necessary for sulfide control was 0.8 L/m3 (of a 53% suspension of magnesium hydroxide in water). This is much higher than the reported values of 0.01 to 0.1 L/m3 in the literature (Apgar et al. 2007; De Haas et al. 2008; Ganigué et al. 2016), so simulations were performed using both the value reported here and those from the literature.

Capital cost

The capital costs for dosing chemicals include those of the dosing tank and piping, electromechanical equipment (dosing pump, level meter, flow meter), and the concrete foundation plate for the dosing tank and fence. The unit prices for these are the same for all dosing products apart from the dosing tank: For magnesium hydroxide, a relatively cheap, single-walled tank can be used, as it is non-hazardous. For ferric chloride and sodium hydroxide, a double-walled or embedded single-walled HDPE tank is used as standard. For calcium nitrate, a more expensive, isolated (double-walled or embedded single-walled), fiber-reinforced, plastic tank is needed, as the chemical crystallizes easily in winter. Space limitations often require that a (more expensive) double-walled tank is used, so the costs of this were assumed in this study. Different suppliers were contacted for unit prices and these were used to compute the total capital cost and net present value (NPV) for each dosing product.

RESULTS AND DISCUSSION

The average annual dosing cost increases linearly with the length of the rising main. The cost relationships with diameter and HRT are shown in Figure 1. For an HRT of two hours, the average annual dosing costs of ferric chloride and calcium nitrate approach zero with increasing rising main diameters, because sulfides form mainly in the biofilm. Hence, the sulfide formation rate decreases with increasing diameter and the concentration at the main outlet approaches 0.2 mg-S/L, which is assumed to be the minimum concentration achievable with these chemical species. For longer HRTs, the dosing costs of ferric chloride and calcium nitrate in the rising main increase almost linearly with diameter, whereas that of calcium nitrate in the pumping station is second-order dependent on the diameter. This arises because denitrification occurs in both the biofilm and the water phase, whereas sulfide formation is thought to take place solely in the biofilm.

Figure 1

Top: Average annual dosing cost of different chemicals for sulfide abatement in a 2 km sewer rising main, as a function of diameter and HRT. Bottom: NPV of the chemicals after 75 years.

Figure 1

Top: Average annual dosing cost of different chemicals for sulfide abatement in a 2 km sewer rising main, as a function of diameter and HRT. Bottom: NPV of the chemicals after 75 years.

The dosing cost of sodium hydroxide depends on the wastewater's buffering capacity. For an average dose of 0.23 g-NaOH/L of wastewater, reported by Apgar et al. (2007), sodium hydroxide is the cheapest treatment product in rising mains of diameter below 200 mm (Figure 1). When an average dose of 0.63 g-NaOH/L (found in lab tests for this study) is assumed, ferric chloride becomes the cheapest dosing product, for all rising main diameters and HRTs (data not shown). However, the dosing quantity is thought to have been overestimated as it is unnecessary to flush the rising main completely, so it is assumed that sodium hydroxide is the cheapest dosing product for rising main diameters below 200 mm. As the aim of sodium hydroxide addition is to increase the wastewater pH to 12, it can only be used if it is diluted enough with fresh sewage before reaching a combined sewage overflow structure or wastewater treatment plant (WWTP).

Ferric chloride dosing is cheaper than nitrate dosing, even when the latter is added at the optimum dosing location downstream of the pumping station. Magnesium hydroxide is the most expensive dosing product. The average annual dosing cost shown in Figure 1 is based on the average dose of 0.05 L/m3, reported in the literature. For the dosing rates found in the present study, even higher dosing costs should be expected, so magnesium hydroxide should only be applied if nothing better is available.

Although ferric chloride is cheap, it is probably less efficient than other chemical species in solving odor problems, because it reacts mainly with sulfides and not with other odorous wastewater components. Similarly, sodium hydroxide may be relatively less effective than other species for odor control as it is dosed intermittently. Nitrate dosing in the pumping station ensures continuous anoxic conditions along the whole pipe, which may be necessary when dealing with odor problems as sulfides or other odorous compounds cannot be formed under these conditions (only under anaerobic conditions), but gives rise to high cost. To ensure that a dosing product can be used for odor control, field tests are necessary. It is good practice to test the cheapest dosing product first.

Figure 1 also shows the NPV of both operating and capital costs after 75 years (the minimum sewer-pipe life time required by Aquafin). Although the single-walled tanks for storage of magnesium hydroxide are much cheaper than the double-walled tanks, the NPV is still much higher than that of the other dosing products.

Sodium hydroxide, ferric chloride and nitrate compounds are all classified as hazardous chemicals. Flemish environmental regulations require that the storage of between 200 and 2,000 kg of hazardous chemicals (or 200 and 20,000 kg in industrial areas) is reported to the city/municipality. For larger quantities environmental permits are necessary. The application procedure is complex and time-consuming, and may not be achievable in urgent situations. Assuming monthly chemical deliveries, the required storage quantity (and hence the need for an environmental permit) can be assessed from the simulations. Curve fitting yielded the approximate relationships shown in Table 3.

Table 3

Approximate relationships between monthly dosing quantity (m3), and rising main diameter D (m), length L (m) and volume V (m3)

Hydraulic residence time (hr) Ferric chloride Nitrate dosing in rising main Nitrate dosing in pumping station 
   
10    
20    
Hydraulic residence time (hr) Ferric chloride Nitrate dosing in rising main Nitrate dosing in pumping station 
   
10    
20    

From these relationships rules-of-thumb can be derived to estimate for which rising mains an environmental permit is necessary. The rules-of-thumb shown in Table 4 are those for an HRT of twenty hours, which covers most situations. In some cities/municipalities, the storage of hazardous chemicals may not even be an option (e.g. for political reasons).

Table 4

Workflow for the selection of dosing products

1. Is the storage of dangerous chemicals allowed? yes 
  1. 1.

    Diameter ≤ 200: Sodium hydroxide (dilution should be checked)

  2. 2.

    Go to question 2.

 
no Go to question 4. 
2. Is an environmental permit possible? yes 
  1. 1.

    Ferric chloride

  2. 2.

    Go to question 3

  3. 3.

    Nitrate dosing in pumping station

  4. 4.

    Go to question 4

 
no 
  1. 1.

    D·L < 1,500 m2: Ferric chloride

  2. 2.

    D·L < 2,300 m2: Go to question 3

  3. 3.

    V < 300 m3: Nitrate dosing in pumping station.

 
3. Is the direct dosing in the rising main possible? yes Nitrate dosing in rising main 
no – 
4. Is the rising main continuously rising? yes Air injection 
No Magnesium hydroxide 
1. Is the storage of dangerous chemicals allowed? yes 
  1. 1.

    Diameter ≤ 200: Sodium hydroxide (dilution should be checked)

  2. 2.

    Go to question 2.

 
no Go to question 4. 
2. Is an environmental permit possible? yes 
  1. 1.

    Ferric chloride

  2. 2.

    Go to question 3

  3. 3.

    Nitrate dosing in pumping station

  4. 4.

    Go to question 4

 
no 
  1. 1.

    D·L < 1,500 m2: Ferric chloride

  2. 2.

    D·L < 2,300 m2: Go to question 3

  3. 3.

    V < 300 m3: Nitrate dosing in pumping station.

 
3. Is the direct dosing in the rising main possible? yes Nitrate dosing in rising main 
no – 
4. Is the rising main continuously rising? yes Air injection 
No Magnesium hydroxide 

The questions in column 1 should be answered in order. D is the diameter of the rising main, L its length and V its volume.

Among the commonly used methods and chemicals for sulfide control, only air injection and magnesium hydroxide are classified as non-hazardous. However, air injection has not yet proved effective in practical implementation in Aquafin's sewerage systems, probably because of the flat or curved longitudinal profile of the rising mains tested. If oxygen is to be transferred effectively from the gas- to the water- phase, the rising main should rise continuously (United States Environmental Protection Agency 1974). If this is not the case, magnesium hydroxide can be used, although at high cost.

Based on the considerations above, the optimum dosing product can be selected – see Table 4 for the workflow. The questions in column 1 should be evaluated. Based on the answers, the possible dosing products are shown in column 3, ranked by increasing cost.

The workflow presented in Table 4 was implemented in a Matlab graphical user interface (GUI), which also computed the cost of the selected dosing products for all rising mains in the sewer network. In addition, it could compute the cost of renovating sewerage infrastructure corroded by sulfide formed in the rising main(s). Thus it could serve as a decision-making tool for selecting the optimum measures not only for odor problems but also for corroded sewer infrastructure.

CONCLUSIONS

The optimum dosing product for sulfide control in sewer rising mains is mainly determined by the cost, chemicals hazard classifications and necessary level of sulfide control. The potential for storing chemicals on site (hazardous vs non-hazardous) and the rising main's dimensions enable preliminary selection of chemicals (as outlined in Table 4).

The necessary level of odor control is complex and depends on many factors, including the wastewater composition, nature of the odorous compounds, distance to houses, turbulence downstream of the rising main, sewer hydraulics, etc. Furthermore, each dosing product has a different level of effectiveness (towards different odorous compounds), so that field tests are necessary to determine which chemicals (if any) can achieve the desired level of control. Field tests should also be conducted with the remaining dosing products (that can be stored on-site), in increasing order of cost.

The simulations performed in this work demonstrate that sodium hydroxide is the cheapest dosing product for rising main diameters below 200 mm and wastewaters with normal buffering capacities. For larger diameters, ferric chloride is the cheapest, followed by nitrate dosing in the rising main at the optimum location, nitrate dosing in the pumping station and magnesium hydroxide.

The workflow outlined in this paper enables selection of the optimum dosing product for a rising main, which should significantly reduce operating costs relative to the outcomes of ad-hoc selections used until now.

REFERENCES

REFERENCES
Apgar
,
D.
,
Witherspoon
,
J.
&
Easter
,
C.
2007
Minimization of Odors and Corrosion in Collection Systems. Chapter 7: Liquid Phase Odor and Corrosion Control
.
Water Environment Research Foundation (WERF)
,
Virginia
,
USA
.
de Haas
,
D. W.
,
Sharma
,
K.
,
Corrie
,
S.
,
O'Halloran
,
K.
,
Keller
,
J.
&
Yuan
,
Z.
2008
Odor control by chemical dosing: a case study
.
Water – Journal of the Australian Water Association
35
(
2
),
138
143
.
Donckels
,
B. M. R.
,
Kroll
,
S.
,
Van Dorpe
,
M.
&
Weemaes
,
M.
2014
Global sensitivity analysis of an in-sewer process model for the study of sulfide-induced corrosion of concrete
.
Water Science and Technology
69
(
3
),
647
655
.
Ganigué
,
R.
,
Gutierrez
,
O.
,
Roosey
,
R.
&
Yuan
,
Z.
2011
Chemical dosing for sulfide control in Australia
.
Water Research
45
(
19
),
6564
6574
.
Ganigué
,
R.
,
Chen
,
J.
,
Vuong
,
L.
&
Yuan
,
Z.
2016
On-line control of magnesium hydroxide dosing for sulfide mitigation in sewers
.
Journal of Environmental Engineering
142
(
12
).
doi:10.1061/(ASCE)EE.1943-7870.0001151
.
Henze
,
M.
,
Gujer
,
W.
,
Mino
,
T.
&
van Loosdrecht
,
M.
2000
Activated Sludge Models ASM1, ASM2, ASM2d and ASM3 – Scientific and Technical Report No 9
.
IWA Publishing
,
London
,
UK
.
Hvitved-Jacobsen
,
T.
,
Vollertsen
,
J.
&
Tanaka
,
N.
1998
A process and model concept for microbial wastewater transformations in gravity sewers
.
Water Science and Technology
37
(
1
),
233
241
.
Jiang
,
G.
,
Sharma
,
K. R.
,
Guisasola
,
A.
,
Keller
,
J.
&
Yuan
,
Z.
2009
Sulfur transformation in rising main sewers receiving nitrate dosage
.
Water Research
43
(
17
),
4430
4440
.
Mohanakrishnan
,
J.
,
Gutierrez
,
O.
,
Sharma
,
K. R.
,
Guisasola
,
A.
,
Werner
,
U.
,
Meyer
,
R. L.
,
Keller
,
J.
&
Yuan
,
Z.
2009
Impact of nitrate addition on biofilm properties and activities in rising main sewers
.
Water Research
43
(
17
),
4225
4237
.
Sharma
,
K. R.
,
Yuan
,
Z.
,
de Haas
,
D.
,
Hamilton
,
G.
,
Corrie
,
S.
&
Keller
,
J.
2008
Dynamics and dynamic modelling of H2S production in sewer systems
.
Water Research
42
(
11–12
),
2527
2538
.
United States Environmental Protection Agency
1974
Process Design Manual for Sulfide Control in Sanitary Sewerage Systems
.
Washington DC
,
USA
.
Vollertsen
,
J.
,
Hvitved-Jacobsen
,
T.
&
Nielsen
,
A. H.
2005
Stochastic modeling of chemical oxygen demand transformations in gravity sewers
.
Water Environment Research
77
(
4
),
331
339
.