Hydraulic modelling of liquid-solid fluidisation in drinking water treatment processes

Research objectives

Developing a better understanding of the hydraulic principles affecting liquid-solid fluidisation in drinking water treatment processes; making improved modelling easier and more accessible; making full-scale implementation more achievable, sustainable and profitable.

Knowledge gap
The main objective is to elucidate the liquid-solid fluidisation system in a fundamentally better way. Starting points concern improving our understanding of the fluidisation principles of natural particles in full-scale pellet-softening reactors and understanding the dependency of the chemical phase to the fluid bed state.

Necessary models
A main aim is to produce a more accurate prediction model of the inner particle-water movements and phenomena of the bed.

Implementation challenges
To improve the performance of full-scale reactors, constraints and sensors have to be developed.
This will help water engineers design more efficient reactors and tackle current challenges.

Project outline

Introduction
In the Netherlands, an annual 400 million m³ water is softened in drinking water treatment plants applying fluidised bed pellet-reactors. Here, sand is generally used as seeding material and marble pellets are produced as a by-product. To improve sustainability, calcite pellets are dried, crushed and sieved and re-used as seeding material. To predict the fluidisation behaviour of particles in fluidised bed reactors, theoretical knowledge is generally used that pertains to spheres that are perfectly round. For natural and imperfectly shaped particles, numerous semi-empirical models have been published, but there is no general agreement on which equation is the most accurate. In many cases, shape factors are introduced for particle diameter to improve numerical results. The particle diameter, an important variable in prediction models is often obtained through traditional sieve analyses. This method is not suitable to quantify particle dimensions of irregular shaped particles.

Research objectives
The performance of the chemical process in pellet-softening reactors is proven dependent on the state of the fluidised particle bed. In the transition from a fixed to a fluidised bed state, after increasing water throughput, the drag force on particles increases. This research will show the dependency of the drag of the actual particles size and change in orientation. It will demonstrate that the shape of the particle will not decline, but that the re-orientation will cause the drag force to decrease considerably. This revisited approach will result in a better understanding and prediction of the fluid bed state.
In the Moody chart, the friction factor is plotted against the Reynolds number, with an emphasis on the turbulent flow. In liquid-solid systems, the flow regime is generally assumed to be laminar. In an improved approach, the friction factor is represented without using the default log-log method.
More advanced morphological particle properties will be obtained using image analyses techniques which makes prediction models become more accurate.
Improved prediction models are calibrated using pilot plant experiments and validated in full-scale reactors.

Approach
We aim to improve our knowledge on the hydraulics of the liquid-solid fluidisation phenomena to optimise the softening process in fluid bed reactors. Research will take place at the Weesperkarspel facility in Amsterdam, in the science lab of the applied university in Utrecht as well in the Process and Energy department laboratory at the 3ME faculty.

Results

Scientific relevance: substantial fundamental knowledge regarding hydraulics of liquid-solid fluidisation phenomena. The use of the drag force substitutes the prolonged usage of the indistinct particle shape factors. Many frequently used approaches and models can now be compared using the modified Moody’s drag-Reynolds diagram without the log-log scales; the laminar and turbulent regimes can be explained better.

Models
Accurate empirical using symbolic regression techniques and also theoretical prediction models for liquid-solid fluidisation reactors increase the opportunity to better maintain and optimise chemical process circumstances while improving water quality. The models will be implemented in the softening process.

Opportunities
New knowledge and improved models are starting points for the optimisation and development of similar water treatment processes, e.g. active carbon and sand filtration processes, and are even useful for other industrial processes and interdisciplinary fields.

Economic relevance
If optimal fluidised bed conditions are successfully implemented in the full-scale facility of Waternet, at least 10% of caustic soda can be saved, i.e. 200 k€/y.

Education
More than a dozen students cooperating in this project will gain theoretical knowledge and practical scientific skills which they, as young professionals, can use on the professional market.

Literature

Powder Technology
Water Research, a Journal of the International Water Association
Water Science and Technology: Water Supply
International Journal of Multiphase Flow
Particuology
Chemical Engineering Science
Journal of Water Supply: Research and Technology – AQUA
Drinking Water Engineering and Science (DWES)