Elsevier

Water Research

Volume 147, 15 December 2018, Pages 25-32
Water Research

Selection of media for the design of ballasted flocculation processes

https://doi.org/10.1016/j.watres.2018.09.041Get rights and content

Highlights

  • The BM surface concentration was identified as a key criterion.

  • Lighter BM may be economical owing to its greater surface per unit mass.

  • BM PZC and shape were not identified as key criteria for ballasted flocculation.

  • Under low superficial velocity, the BM SG and size were less important.

Abstract

Conventional clarification processes imply specific facility footprints that translate into important capital costs. Ballasted flocculation, consisting of injecting ballast medium to increase floc specific gravity and size, is being increasingly used in the water industry owing to its potential for design with very high superficial velocities. However, no systematic approach has yet been proposed to compare and select an appropriate ballast medium with respect to its specific gravity and size. In order to facilitate this procedure, this research project explores the hypothesis that flocculation performance is controlled by the surface area of the medium available for ballasted flocculation. This hypothesis was tested at laboratory scale by evaluating five ballast media with differing specific gravity and size: granular activated carbon, anthracite, silica sand, ilmenite, and magnetite sand having specific gravities of 1.24, 1.45, 2.62, 3.70, and 5.08, respectively. Flocculation kinetics were monitored by measuring floc size through microscopy and with a camera installed directly on the jar-test beaker. Settling performance was monitored using turbidity measurements. This study shows that all ballast media, when expressed as total surface available during flocculation, required similar surface concentrations to achieve settled water turbidity near 1 NTU and lower. In addition, the effects from the ballast media size and specific gravity were lowered for settling time longer than 3 min. Inversely, for settling time of 12 s, larger and denser media produced lower settled water turbidity. For certain applications, lighter ballast media may be more economical because they offer more available surface area for a given mass concentration, hence reducing the amount of ballast media required in the flocculation tank. Finally, the ballast media point of zero charge and shape were not identified as key criteria for ballasted flocculation.

Introduction

Clarification processes require specific footprints that translate into important capital costs, especially in the Nordic climate where settling basins must be located inside heated buildings. Ballasted flocculation, consisting of injecting micron-sized granular media to increase the specific gravity (SG) and size of flocs, is being used increasingly in the water industry owing to its potential for achieving very high superficial design velocities (as high as 85 m/h; MDDEP (2009)). This advantage offers a more compact process (i.e., smaller footprint), stable suspended solids removal performance, faster start-up, and important savings compared to conventional flocculation (Desjardins et al., 2002; Guibelin et al., 1994). In earlier studies, we showed that the floc SG and size were two important factors for predicting settling performance (Lapointe et al., 2017). The final SG of flocs depends on the SG of the medium: a denser ballast medium (BM) leading to denser flocs. However, a denser medium requires a higher velocity gradient (i.e., G value in s−1) to maintain it in suspension. This leads to lower average floc diameters owing to floc break-up, and to higher fraction of un-ballasted flocs. Therefore, with reference to the design criteria (minimizing footprint or minimizing settled water turbidity); an optimal BM can be selected for a given application.

The concentration of BM inside the flocculation tank (typically referred to as the maturation tank) is very high, in the range of 4–10 g silica sand (SS)/L for industrial applications. The BM concentration depends on the particle concentration in the flocculated waters (which is linked to the coagulant dosage and the source-water particle concentration). For example, in jar testing, a higher BM concentration was necessary for a wastewater application compared to a drinking water application (2 g of SS/L for surface water of 12 NTU and 3 g of SS/L for municipal wastewater of 130 NTU (Lapointe and Barbeau, 2017; Lapointe et al., 2017)). Apart from the BM concentration, its particle size distribution is also expected to impact ballasted media performance. A medium of smaller particles translates into a higher number of available BM particles for a given mass concentration. Flocculation performance is known to be a function of G×t×No, where G is the mean velocity gradient, t the flocculation time, and No, the initial number of particles to flocculate. However, large BM media can produce larger flocs, which settle more rapidly. Consequently, similar to the selection of a BM SG, there is an optimum selection of the medium size to achieve the best overall flocculation/settling performance.

In summary, the selection of an appropriate BM medium is dictated by its SG, size, and the need to minimize un-ballasted flocs. This is because the latter are not well removed during settling at high superficial velocity. In addition, BM availability and cost, its resistance to abrasion, and the effectiveness of its recovery by hydrocyclone are important criteria to consider (Desjardins et al., 2002; Sibony, 1981; Young and Edwards, 2003). However, no systematic approach has yet been proposed to compare and select an appropriate BM with respect to its SG and size for a given application. In order to facilitate this procedure, this research project explored the hypothesis that flocculation performance is controlled by the surface area of BM available for ballasted flocculation, expressed as m2 of surface per L of water. We hypothesized that differing media density and particle size distributions impact ballast media performance by increasing/reducing the reaction surface. This hypothesis was tested at laboratory scale by evaluating five ballast media with differing SG, all of which were sieved to produce three different particle size distributions. Flocculation kinetics were monitored by measuring floc size by microscopy and with a camera installed directly on the jar-test beaker. Settling performance was monitored using turbidity measurements. Recommendations are made for guiding the selection of an adequate BM to achieve turbidity removal for a specific design superficial velocity.

Section snippets

Water characteristics

All experiments were conducted at laboratory scale at 21 ± 1 °C using surface water from the Sainte-Rose drinking water treatment plant (10 ± 2 NTU; pH 6.9), which is fed by the Mille-Îles River (Quebec, Canada). The tested surface water exhibits a relatively low alkalinity (30 mg CaCO3/L) and a significant dissolved organic carbon (DOC) concentration (6.9 mg C/L). The raw water was collected just past the 10 mm influent bar screens and refrigerated at 4 °C during the experiments (2 weeks).

Jar test procedure

The

Selection of the optimal ballast medium concentration

In past research efforts, BM optimal dosages were selected on the basis of mass base concentration (expressed in g/L). However, identical mass base concentration BM media with differing average diameters or densities will generate different particle concentration and surface area. Fig. 2 compares the impact of the BM media (expressed in g/L in Fig. 2A and m2/L in Fig. 2B) on settled water turbidities. To prepare this figure, the smallest/lightest GAC (80–125 μm) and largest/densest MS medium

Conclusions

This study demonstrated that ballasted flocculation is best described as contact between surfaces (microflocs and BM). Consequently, the BM concentrations should be normalized with respect to surface in order to select an optimal and appropriate BM. For very high settling rates, the BM surface concentration, SG, and size had important effects on turbidity removal. On the other hand, as the superficial velocity applied in the clarifier decreased, it was shown that the BM SG and size were less

Acknowledgments

These experiments were conducted as part of the Industrial-NSERC Chair in Drinking Water (Polytechnique Montreal) research program, which benefits from the financial support of Veolia Water Technologies Canada, Inc., the City of Montreal, the City of Laval, the City of Repentigny and the City of Longueuil. Experiments were conducted at the CREDEAU, a research infrastructure supported by the Canadian Foundation for Innovation and the Ministry of Education of Quebec.

References (28)

  • APHA et al.

    WEF Standard Methods for the Examination of Water and Wastewater

    (2012)
  • T.R. Camp et al.

    Velocity gradients and internal work in fluid motion

    J. Boston Soc. Civ. Eng.

    (1943)
  • F. De Dianous et al.

    Advantages of weighted flocculation in water treatment

    Water Supply

    (1991)
  • A. Ghanem et al.

    Mechanisms of ballasted floc formation

    J. Environ. Eng.

    (2007)
  • Cited by (30)

    View all citing articles on Scopus
    View full text