Calcium carbonate pellets are produced as a by-product in the pellet softening process. In the Netherlands, these pellets are applied as a raw material in several industrial and agricultural processes. The sand grain inside the pellet hinders the application in some high-potential market segments such as paper and glass. Substitution of the sand grain with a calcite grain (100% calcium carbonate) is in principle possible, and could significantly improve the pellet quality. In this study, the grinding and sieving of pellets, and the subsequent reuse as seeding material in pellet softening were tested with two pilot reactors in parallel. In one reactor, garnet sand was used as seeding material, in the other ground calcite. Garnet sand and ground calcite performed equally well. An economic comparison and a life-cycle assessment were made as well. The results show that the reuse of ground calcite as seeding material in pellet softening is technologically possible, reduces the operational costs by €38,000 (1%) and reduces the environmental impact by 5%. Therefore, at the drinking water facility, Weesperkarspel of Waternet, the transition from garnet sand to ground calcite will be made at full scale, based on this pilot plant research.

INTRODUCTION

Amsterdam has the ambition to develop itself as a competitive and sustainable European metropolis. The flows of energy, water, resources and waste streams within the urban environment of Amsterdam have a large potential to contribute to this ambition. The sustainability of cities can be increased using circular instead of linear management of urban flows. In a linear management of urban flows, materials and resources are used only once (‘take-make-waste’), while in a circular management, materials and resources are reused and not considered waste (Van der Hoek et al. 2013).

Waternet, the public utility responsible for the water management of Amsterdam and surroundings, has the ambition of a sustainable and environmental-friendly operation of all activities related to water management, like drinking water treatment and supply (Graveland et al. 1983; Van Dijk & Wilms 1991; Mohapatra et al. 2002; Barrios et al. 2008; Klaversma et al. 2013). Waternet treats all of the drinking water with the pellet softening process for reasons of public health (copper and lead solubility), costs, environmental benefits and customer comfort (Hofman et al. 2007). In the Netherlands, approximately 50% of the drinking water is treated with the pellet softening process. The reactor used by Waternet for the pellet softening process consists of a cylindrical up-flow reactor, which is partly filled with seeding material, such as river sand or garnet sand. Water is pumped in an upward direction through the reactor at a velocity varying between 60 and 100 m/h, maintaining the seeding material in a homogeneous fluidised bed condition. At the bottom of the reactor, a chemical base is dosed (e.g. caustic soda or lime). Upon mixing with the influent water, the pH of the water increases and the water becomes super-saturated with calcium carbonate. Consequently, crystallisation on the surface of the seeding material takes place, resulting in growth of the pellets. If the conditions in the bottom part of the reactor are not well controlled, besides crystal growth on the pellets, nucleation of calcite may occur in solution. At regular intervals, pellets at the bottom of the reactor are removed (Van Dijk & Wilms 1991). The pellets are applied as a (secondary) raw material in construction, agriculture and the mineral-resource sector. This is organised by the Reststoffenunie, the shared service centre that processes the residual products of all Dutch drinking water companies. The sand grain inside the pellet inhibits reuse as seeding material and application in high-potential market segments such as glass and paper. The garnet sand used by Waternet originates from Australia. The transportation to the Netherlands causes a negative environmental impact. Therefore, it is worthwhile to study the possibilities of the use and reuse of locally produced ground calcite as a seeding material.

Recently performed research showed that replacement of sand grains with commercially available calcite grains (nearly 100% calcium carbonate) obtained from limestone quarries from Italy or Austria is technologically possible (Palmen et al. 2012). To further reduce the transport costs and environmental impact and to enhance the reuse of the pellets, a closed loop is introduced in the pellet softening (Schetters 2013). About 10% of the removed fully grown calcite pellets would be needed as a seeding material, after grinding and sieving. In the closed loop, after a first start-up of the process with commercial calcite grains, 20% of the removed fully grown calcite pellets have to be ground, assuming a grinding efficiency of 50%. The remaining calcite pellets can be used for other industries. The single component composition of these calcite pellets increases the potential market value. Recycling of the pellets in the drinking water treatment, as well as increasing higher (local) market potential, contributes to the circular economy approach of the city of Amsterdam.

This paper presents the research into the possibilities of the use of ground calcite pellets as an alternative for the use of garnet sand as seeding material in the pellet softening process. At pilot plant scale, a comparative study was executed between ground calcite and garnet sand, extended by modelling of the process. To judge the feasibility in full-scale application, a life-cycle assessment (LCA), a cost evaluation and a risk analysis were performed at full scale.

METHODS

Pilot plant experiments

Pilot plant experiments were performed at the treatment plant Weesperkarspel (WPK) of Waternet, located in Amsterdam, The Netherlands. The raw water originates from seepage water from the Bethune polder. The water is pre-treated by coagulation and sedimentation, followed by approximately 90 days' retention in a lake reservoir. Subsequently, it is filtered through rapid sand filters. The treatment plant that follows contains ozonation, pellet softening, biological activated carbon filtration and slow sand filtration. At the WPK treatment plant, a pilot plant with two identical pellet softening reactors with a height of 6 m and a diameter of 0.3 m was operated. The reactors had the same height as the full-scale reactors. The influent of the pilot plant was withdrawn from the ozonation of the full-scale plant to mimic full-scale conditions. The first pellet reactor was filled with garnet sand, and the second pellet reactor was initially filled with commercial (Italian) calcite as a seeding material. After sufficient pellets were produced, the pellets were ground and used as seeding material in this second pellet reactor. At the bottom of the reactors caustic soda (50% NaOH W/W) was added as a base. By running two reactors simultaneously in parallel, a good comparison can be made between the performance of both seeding materials.

The saturation index (SI) of calcite is defined as the pH offset at which the actual calcium concentration is in equilibrium with the carbonate. The SI is a measure for the driving force of the crystallisation process. The theoretical calcium carbonate crystallisation potential (TCCP) is the amount of calcium in (mmol/L) that can theoretically crystallise from the solution (to obtain SI = 0) and is a measure for the amount of calcium carbonate that can be formed in consecutive process steps. The TCCP and SI were determined based on the calcium carbonic acid equilibrium with the use of calculation software Aquacalc (based ultimately on the work by Plummer & Busenbrug (1982)). The major input parameters were temperature, pH, conductivity, Ca2+ and concentration.

To determine the fluidised bed status, the reactors were equipped with online measurements of water flow, water temperature, bed height, pressure drop over the total fluidised bed and pressure drop between 20 and 60 cm from the bottom of the reactor (Van Schagen 2009). In line with the study of Van Schagen (2009), the turbidity, pH, total hardness (TH), m-alkalinity (where m stands for methyl orange) and the conductivity were automatically measured every 60–120 minutes using an online titration unit (Applikon ADI 2040). Measurements were taken in the influent and effluent of both reactors. Manual measurements of the caustic soda dosage were taken on a daily basis. Once per week laboratory measurements of the TH, pH, water temperature, alkalinity, turbidity and conductivity were taken.

Both reactors were operated identically, based on flow (60–80 m/h), caustic soda dosage (0.16–0.25 L/m3), bed height (3.5–5.0 m) and pellet discharge diameter (0.6–1.2 mm), resulting in a total effluent hardness in the range of 0.5–2.0 mmol/L.

A selection of the continuous online monitoring measurements of the water quality parameters was made. A series of measurements was selected when a stable operation was reached. The operation was considered stable when the pH, turbidity, head loss and calcium concentration were stable for at least 3 h. The mean value of the measurements during the stable operation at a specific setting was determined as one data point in Figures 1 and 2. The total data set contained 36 parallel determined settings from the garnet sand and ground calcite reactor, totalling 72 settings. The measurements were taken after inserting the first ground calcite.

Figure 1

Total hardness (TH), pH and turbidity (NTU) as a function of the date (dd-mm) of the effluent of the garnet (garnet) and ground calcite reactor (calcite).

Figure 1

Total hardness (TH), pH and turbidity (NTU) as a function of the date (dd-mm) of the effluent of the garnet (garnet) and ground calcite reactor (calcite).

Figure 2

TCCP, SI and temperature of the effluent of the garnet (garnet) and ground calcite (calcite) reactor as a function of the date (dd-mm).

Figure 2

TCCP, SI and temperature of the effluent of the garnet (garnet) and ground calcite (calcite) reactor as a function of the date (dd-mm).

Modelling of the process

The results of the pilot plant experiments were compared with modelling in the modelling environment Stimela (Rietveld 2005; Van Schagen 2009). This model is used for model-based control and optimisation of the full-scale pellet softening process at WPK. Density and diameter of the seeding materials are input parameters in Stimela. In the current model, garnet sand is used as a seeding material (Van Schagen et al. 2010). The used parameters of the garnet sand are: ρgarnet = 4,114 kg/m3, Øgarnet = 0.25 mm. The used parameters of the ground calcite grains are: ρcalcite = 2,670 kg/m3, Øcalcite = 0.5 mm.

Costs and life-cycle analyses

To determine the environmental and financial impact of the use of the garnet sand and the grinded calcite, the complete material flow including mining, handling, transport, use and reuse of the material was analysed using SimaPro (Klaversma et al. 2013). SimaPro uses the Ecoinvent database, which contains life-cycle information about the materials and processes. The impact of the reuse of the pellets is determined by the reduction of the environmental impact by replacing a raw material with the pellets. The results were compared with earlier research on the life-cycle impact of the entire full-scale treatment plant of WPK, in which garnet sand is used (Barrios et al. 2008).

FMECA

A Failure Mode, Effects and Criticality Analysis (FMECA) was carried out for the transition of garnet sand to ground calcite in the full-scale treatment plant WPK. All possible failures of the process elements in the water and the seeding material stream were identified and evaluated, in order to estimate their effect on water quantity, water quality, safety, environment and social image. The results were compared with the risk matrix of Waternet in order to identify the critical elements and to construct preventive measures. An inventory session with a multidisciplinary team of Waternet employees, including expertise on operation, maintenance, research and strategy, was held to validate the results of the FMECA.

RESULTS AND DISCUSSION

Reactor performance comparison for garnet sand and ground calcite

For both the pilot reactor with garnet sand and the reactor with ground calcite as seeding material, the TH, pH and turbidity are shown as a function of time (Figure 1). Furthermore, the TCCP, SI and temperature are shown (Figure 2). During the test period, the effluent temperature ranged from 1–4 °C (Figure 2). To analyse the performance of both reactors, the TH, pH, turbidity, TCCP and SI of the effluent were compared.

Total hardness

It can be observed (Figure 1) that at the start of the experiment, a TH of 1.5 mmol/L was reached for both the effluent of ground calcite and garnet sand reactor. During the experiment, the TH varied from 0.5 to 1.6 mmol/L. On average, the TH of the effluent of the calcite reactor was 0.05 mmol/L higher with a standard deviation of 0.06 mmol/L, while measurement accuracy was approximately 0.1 mmol/L. Both reactors were operated in a similar way; however, small deviations of the settings, e.g. flow and NaOH dosage, could not be prevented. As this difference is small, it may not be significant. Based on the TH, no difference could be observed for the garnet sand and calcite reactor.

pH

The pH of the effluent of both reactors showed similar values during the experiment, as can be seen in Figure 1. On average, the pH of the effluent of the calcite reactor was 0.2 units lower compared with the garnet sand reactor. As a consequence, the TH of the effluent of the calcite reactor was 0.05 mmol/L higher. This again indicates that there were small differences between both the reactors, either due to the seeding material or slightly different operation conditions.

Turbidity

Due to the smaller diameter of the garnet sand, the specific surface in the top section of the garnet sand reactor was higher than the specific surface in the top section of the calcite reactor. In both the reactors, nucleation occurred under the same operating conditions. Hence, the effect of the difference in specific surface could not be observed based on the turbidity.

The appearance of small white particles in the effluent of the garnet sand and ground calcite reactor could be visually observed when NaOH dosage resulted in a TH ≤0.8 mmol/L and for pellet discharge diameters >1.1 mm (which results in a low specific crystallisation surface). The turbidity measurements confirm the visual observation, showing values in the range of 2–10 nephelometric turbidity units (NTU). This indicates that nucleation occurred in these circumstances due to the combination of the high oversaturation and low crystallisation surface and temperature.

TCCP and SI

The TCCP is the amount of calcium (mmol/L) that would need to precipitate to obtain water in chemical equilibrium (SI = 0). From Figure 2, it can be seen that the TCCP as well as the SI of the effluent of the garnet and ground calcite reactors are similar. The deviations are mainly caused by small differences in the reactor settings, such as flow and NaOH dosing. Again, this indicates that there were small differences between both the reactors, either due to the seeding material or slightly different operation conditions.

At an effluent TH >0.8 mmol/L, no large difference between the TCCP of the garnet sand and calcite was observed. At an effluent TH below 0.8 mmol/L, the TCCP increased due to the fact that microcrystals, formed by nucleation, are not taken into account in the calcium concentration measurement. Therefore, the measured calcium concentration was higher than in reality, which resulted in a higher calculated TCCP. The TH-TCCP curve showed similar results as measured with garnet sand as seeding material by Stroom (1993) and modelled by Van Schagen (2009). These results indicate that the amount of calcium carbonate that can be formed in the consecutive process steps does not increase when ground calcite is used as seeding material.

Combination of the parameters

No large differences were observed between the TH, pH, turbidity, TCCP and SI in the effluent of the garnet sand and the ground calcite reactor. It can be concluded that the lower specific surface area of the calcite grains and pellets in the reactor is not the limiting parameter of the crystallisation process in this case, even at the very low water temperatures used.

Comparing pilot plant results with modelling results

The results of the modelling in the Stimela environment of the softening reactors with the garnet sand and with ground calcite were compared with the pilot plant results. The data in Figure 3 show comparable results between the model and the results of the pilot plant experiments. The average deviation of the calcium concentration of modelling and the results of the pilot plant was 0.01 mmol/L for both the garnet sand reactor and the calcite reactor. However, the standard deviation for garnet sand was 0.09 mmol/L and for calcite 0.02 mmol/L. The deviations are in the same order of magnitude as the measuring accuracy. The calcite grains have a similar density as the crystallised calcium carbonate (2,670 kg/m3). The uniform density of the calcite grains in the reactor may be an explanation for better model fit. Based on this data set, it could be concluded that the calculation of the calcium concentration by the model fits the measurements better with calcite than with garnet sand. In the garnet sand reactor, the density of the grains changes over the height, because it depends on the amount of calcium carbonate (2,671 kg/m3) that is crystallised on the garnet sand grain (4,114 kg/m3).

Figure 3

Results of the calcium concentration of the pilot installation (PI) and the modelling in Stimela, for garnet sand and ground calcite, pellet discharge diameter of 1 mm.

Figure 3

Results of the calcium concentration of the pilot installation (PI) and the modelling in Stimela, for garnet sand and ground calcite, pellet discharge diameter of 1 mm.

Introduction in the full-scale plant Weesperkarspel

Effect on sustainability

The results of the LCA of the use and reuse of the garnet sand, calcite and ground calcite are presented in Table 1. The environmental impact of the use of the ground calcite consists only of the grinding and sieving of the pellets on site, so no transport is required. Therefore, the impact is lower compared with the current garnet sand, which is transported by boat and truck from Australia to the Netherlands and the commercial calcite, which is transported by truck from Italy to the Netherlands. The impact of the reuse in other industries is determined by the replacement of raw material in these industries by the softening pellets. Pellets with a nucleus of garnet sand can replace the raw material sand in other industries. Pellets with a nucleus of calcite can replace the raw material calcite in other industries. In all cases, the reuse of the pellets in other industries results in a lower environmental impact. The replacement of calcite showed the highest environmental impact reduction. The impact reduction of commercial calcite is higher than ground calcite due to the fact that a part of the ground calcite will be used as seeding material and cannot be reused in other industries. If both the use and reuse are taken into account, on-site ground calcite showed the highest environmental impact reduction of 10,640 EcoPoints. Earlier research showed that the environmental impact of the entire treatment facility of WPK is 215,250 EcoPoints (Barrios et al. 2008). By using and reusing ground calcite, the environmental impact of the entire treatment facility of WPK could be reduced by 5%.

Table 1

Life-cycle assessment including the environmental impact of garnet sand, calcite and on-site ground calcite for the WPK treatment plant

 Garnet Calcite On-site ground calcite 
Seeding material (tonne/y) 150 750 750 
Pellets for reuse (tonne/y) 1,922 2,154 1,404 
Use (EcoPoints/y) 3,826 20,218 
Reuse (EcoPoints/y) −809 −11,698 −7,625 
Total (EcoPoints/y) 3,017 8,520 −7,622 
Netto (EcoPoints/y) 5,503 −10,640 
 Garnet Calcite On-site ground calcite 
Seeding material (tonne/y) 150 750 750 
Pellets for reuse (tonne/y) 1,922 2,154 1,404 
Use (EcoPoints/y) 3,826 20,218 
Reuse (EcoPoints/y) −809 −11,698 −7,625 
Total (EcoPoints/y) 3,017 8,520 −7,622 
Netto (EcoPoints/y) 5,503 −10,640 

Note. Annual drinking water capacity of WPK treatment plant is 25 Mm3/y.

Total hardness reduction: 0.75 mmol/L.

Effect on costs

According to Barrios et al. (2008), the operational costs of the treatment facility WPK, including the pre-treatment, are €0.1345/m3. At a capacity of 25 million m3/year, this amounts to €3,662,500. Using ground calcite can reduce the cost of the whole treatment process by €38,000 (1% of the total drinking water treatment operational costs). Table 2 presents the financial comparison for garnet sand, commercially purchased calcite and reused ground calcite. The larger amount of required calcite seeding material is caused by the lower density of calcite compared with garnet and its effect on the required grain size. The increased pellet turnover in the calcite scenario is caused by the improved pellet quality.

Table 2

Costs of seeding material for Weesperkarspel treatment plant

 Garnet Calcite Ground calcite 
Grain density (kg/L) 4.1 2.7 2.7 
Grain size, d50 (mm) 0.25 0.5 0.5 
Annual grain usage (tonne/y) 150 750 750 
Annual grain costs (k€/y) 55 113 38 
Annual pellets for industry (tonne/y) 1,900 2,200 1,400 
Annual turnover pellets (k€/y) 33 21 
Net (k€/y) −55 −80 −17 
Overall difference (k€/y) Reference −25 + 38 
 Garnet Calcite Ground calcite 
Grain density (kg/L) 4.1 2.7 2.7 
Grain size, d50 (mm) 0.25 0.5 0.5 
Annual grain usage (tonne/y) 150 750 750 
Annual grain costs (k€/y) 55 113 38 
Annual pellets for industry (tonne/y) 1,900 2,200 1,400 
Annual turnover pellets (k€/y) 33 21 
Net (k€/y) −55 −80 −17 
Overall difference (k€/y) Reference −25 + 38 

Annual drinking water capacity: 25 Mm3/y.

Total hardness reduction: 0.75 mmol/L.

Effect on risks

An FMECA was carried out for the transition of garnet sand to calcite and ground calcite in the treatment plant of WPK. The results from the FMECA showed that the hygiene of the seeding material is critical for the water quality. Measures to control the hygienic handling of the pellets, grinding, sieving and transportation are required to ensure high water quality, especially microbiological quality due to the fact that water is distributed without chlorine in the Netherlands. The most important measure is the control of the quality of the seeding material before it enters the reactor. After storage in a silo, the seeding material is washed in a washing installation to remove small particles and impurities. The abrasive character of calcite could cause material loss or problems in the washing installation. This should be closely monitored in the initial phase of the project. Backup plans need to be prepared in case the water cannot be softened due to a failure of one of the process elements. A possible solution is additional backup storage of garnet sand or commercial calcite. The FMECA showed no critical risks that could not be controlled by taking sufficient measures.

Implementation

The proposed transition has started in January 2014. The full-scale treatment plant started with commercially available calcite. The garnet seeding material in the reactors was gradually replaced with calcite. The performance of the full-scale reactor is in line with the pilot plant experiments. In June 2014, the first calcite pellets were ground at full scale. In the long term, the full-scale treatment plant at WPK will use ground calcite as seeding material.

CONCLUSIONS

The main goal of this research was to determine if ground calcite pellets could be used as alternative seeding material for garnet sand in pellet softening at low temperature (proof of principle). The pilot plant experiments showed that the softening performance of the ground calcite as a seeding material was similar to garnet sand as seeding material. In terms of water quality parameters calcite was comparable with the garnet sand. Results of the Stimela modelling showed comparable results to the pilot plant experiment.

To assess the feasibility of calcite as seeding material, an LCA, economic evaluation and risk analysis (FMECA) were performed. The LCA concerning the WPK plant and its pre-treatment showed a score of 215,250 EcoPoints. The scenario of ground calcite seeding material results in a decrease of about 3,824 EcoPoints (1.8%) due to the elimination of transportation of garnet sand. Usage of ground pellets in Dutch industries (instead of imported calcite) results in a total decrease of 10,640 EcoPoints (5%), mainly caused by the reduced transportation. This contributes to a circular material flow and a more sustainable drinking water production in Amsterdam. The economic evaluation showed that costs of the seeding material would be lower due to the reuse of the ground pellets as seeding material. Due to the improved pellet quality, the pellet turnover will increase. A cost reduction for the WPK plant of €38,000 (1%) can be obtained using ground calcite pellets. The FMECA showed no critical risks that could not be controlled by taking sufficient measures.

The results of this research show that using an internal and external circular material flow approach to innovate the drinking water treatment process can benefit both the drinking water company and the industries in the surroundings of Amsterdam. The optimisation of the quality of the residuals from the drinking water treatment process for use in other industries achieves both financial benefits and increased sustainability of the drinking water treatment process.

ACKNOWLEDGEMENTS

This research is conducted within the joint research programme of the drinking water companies Dunea, PWN and Waternet (DPW). The authors thank E. Baars (Waternet), W. Oorthuizen (Dunea) and B. Martijn (PWN) for their contributions.

REFERENCES

REFERENCES
Barrios
R.
Siebel
M.
van der Helm
A.
Bosklopper
K.
Gijzen
H.
2008
Environmental and financial life cycle impact assessment of drinking water production at Waternet
.
Journal of Cleaner Production
16
(
4
),
471
476
.
Graveland
A.
van Dijk
J. C.
de Moel
P. J.
Oomen
J. H. C. M.
1983
Developments in water softening by means of pellet reactors
.
Journal AWWA
75
(
12
),
619
625
.
Hofman
J.
Kramer
O. J. I.
van der Hoek
J. P.
Nederlof
M.
Groenendijk
M.
2007
Twenty years of experience with central softening in The Netherlands: water quality – environmental benefits – costs
.
Water21
9
(
1
),
21
24
.
Klaversma
E.
Van der Helm
A. W. C.
Kappelhof
J. W. M. N.
2013
The use of life cycle assessment for evaluating the sustainability of the Amsterdam water cycle
.
Journal of Water and Climate Change
4
(
2
),
103
109
.
Mohapatra
P.
Siebel
M.
Gijzen
H.
Van der Hoek
J.
Groot
C.
2002
Improving eco-efficiency of Amsterdam water supply: a LCA approach
.
Journal of Water Supply, Research and Technology: AQUA
51
(
4
),
217
227
.
Palmen
L.
Oorthuizen
W.
Koppers
H.
Hofs
B.
2012
Calciet als alternatief entmateriaal bij ontharding produceert hoogwaardige kalkkorrel (Calcite as alternative seeding material in softening produces high quality calcite grain)
.
H2O Magazine
45
(
10
),
35
37
.
Plummer
L. N.
Busenbrug
E.
1982
The solubilities of calcite, aragonite and vaterite in CO2–H2O solution between 0 and 90 °C
.
Geochimica et Cosmochimica Acta
46
,
1011
1040
.
Rietveld
L. C.
2005
Improving Operation of Drinking Water Treatment through Modeling. PhD Thesis, Faculty of Civil Engineering and Geosciences
,
Delft University of Technology
,
Delft, The Netherlands
.
Schetters
M. J. A.
2013
Grinded Dutch Calcite as Seeding Material in the Pellet Softening Process. Master Thesis, Faculty of Civil Engineering and Geosciences
,
Delft University of Technology
,
Delft, The Netherlands
.
Stroom
J.
1993
Minimalisatie van de uitstoot bij de hardheidsreductie (Minimisation of the Emission in the Softening Process)
.
Internal Research Report
,
Gemeente Waterleidingen Amsterdam
,
Amsterdam, The Netherlands
.
Van Dijk
J. C.
Wilms
D. A.
1991
Water treatment without waste material-fundamentals and state of the art of pellet softening
.
Journal Water Supply, Research and Technology: AQUA
40
(
5
),
263
280
.
Van der Hoek
J. P.
Struker
A.
Danschutter
J. E. M.
2013
Amsterdam as a sustainable European metropolis: integration of water, energy and materials flows
. In:
Proceedings of the IWW Conference ‘Integrated Water Solutions for a Green Economy’
,
Van der Hoek, J. P., Struker, A. & Danschutter, J. E. M. (eds)
,
International Water Week Amsterdam
,
The Netherlands
, pp.
1
9
.
Van Schagen
K. M.
2009
Model-Based Control of Drinking-Water Treatment Plant. PhD Thesis, Faculty of Civil Engineering and Geosciences
,
Delft University of Technology
,
Delft, The Netherlands
.
Van Schagen
K. M.
Rietveld
L. C.
Veersma
A.
Babuška
R.
2010
Control-design methodology for drinking-water treatment processes
.
Water Science & Technology: Water Supply
10
(
2
),
121
127
.