Pilot study and modelling of remineralization of low-temperature desalinated water by calcite filtration

Abstract

To prepare for future challenges, such as possible upcoming organic micro pollutants in the source water, including traces of medicines, pesticides, and industrial by-products, it is expected that the conventional treatment does not ensure a reliable quality of drinking water. As a result, during the past few years, Oasen started to research a new treatment concept based on 100% reverse osmosis (RO) membrane filtration to provide an excellent barrier for organic micro pollutants. However, the water produced by the RO membranes, called permeate, is corrosive, bitter in taste and does not comply with the drinking water regulation standards in the Netherlands. Therefore, a certain degree of remineralisation is crucial to solve these problems and improve the water quality. A commonly used remineralisation process is to filter the desalinated water through a calcite contactor, providing the appropriate amount of bicarbonate and calcium in the water. In order to properly design and operate the calcite filters as well as to predict the final water quality, it is essential to understand the processes that occur in the filter.
The aim of this study was to find the best kinetic calcite dissolution model in order to understand the calcite grains dissolution behavior inside the filter and subsequently to adequately design and operate the calcite filter. Therefore, extensive pilot research was conducted to investigate the effect of various parameters on calcite dissolution such as the calcite grain size, velocity and carbon dioxide concentration. On top of that, the dissolution was modelled based on a successful empirical expression given by Yamauchi et al. (1987). However, it was found that the effect of the flow rate on the diffusion boundary layer around the calcite grains has not been taken into account in the study carried out by Yamauchi et al. (1987). Therefore, the effect of velocity on the calcite dissolution coefficient was investigated at five different velocity ,i.e., 5, 10,15,20,30 m/h. From there, a function was developed to describe the correlation between flow rate and the dissolution rate coefficient. In order to calculate the equilibrium concentration, the chemical reactions were simulated using PhreeqPython (Phreeqc built in Python).
The main difference of this study compare to previous investigations was the low temperature of the water (12 oC vs 22-40 oC) and the smaller grain size of the calcite (0.5-1.2 mm vs 1-2 and 2-3mm) which was tested. Besides that a high range of CO2 dosing (1.45- 9.5 mmol/l) was tested. As expected from theory, the dissolution rates was strongly affected by the varied parameters. It is concluded that the smaller grain size of 0.5-1.2 mm reduced the required empty bed contact time (EBCT) to15 min where operating the filter with the larger grain size of 1-2mm needs a minimum EBCT of 25 min to reach calcite equilibrium. The CO2 dosing is recommended to be less than 3 mmol/l, since the CO2 efficiency will drop under the 60% at higher CO2 concentrations.

Eventually, the optimal design will be introduced for the remineralisation process at Oasen treatment plant De Hooge Boom located in Kamerik. For this purpose various operational scenarios were compared on capital and operational cost. The overall cost including, both Capital expenses (CAPEX) and Operational expenses OPEX, was estimated between € 0.048 and 0.064 per m3 for different scenario’s where 71% consists of investment cost. The total treatment cost of this design is 0.057 €/m3 and the investment cost was found to be € 1.351.000 which is 32% less than price estimated by previous study done by Oasen.