Elsevier

Desalination

Volume 443, 1 October 2018, Pages 198-211
Desalination

Cost-optimal design of a batch electrodialysis system for domestic desalination of brackish groundwater

https://doi.org/10.1016/j.desal.2018.05.010Get rights and content

Highlights

  • Capital cost dominated energetic cost; hence, high current densities were favored.

  • High velocities (11–21 cm/s) optimally balanced mass transfer and pressure drop.

  • Channel heights were minimized to 0.30–0.33 mm.

  • Pumps contributed up to 46% of the total capital cost.

  • Active area scaled linearly with feed concentration, but varying recovery between 60–90% minimally affected the optimal design.

Abstract

This study presents the pareto-optimal design of a domestic point-of-use batch electrodialysis (ED) system. Specifically, the optimal geometry, flow-rates, and applied voltage for total cost minimization were explored for varying production rate (9–15 L/h) and product concentration (100–300 mg/L) requirements, while feed concentration and recovery ratio were maintained at 2000 mg/L and 90%, respectively. Capital cost dominated over energetic cost; hence, optimal designs maximized current density. Capital cost was significantly higher for 100 mg/L systems, than 200 and 300 mg/L: $141 vs. $93 and $79, at 12±0.5 L/h of production. Pumps were an important consideration, contributing up to 46% of the total cost. Large membrane length-to-width aspect ratios (3.5:1 to 6:1) and thin channels (0.30–0.33 mm) promoted high current densities, and 11–21 cm/s velocities optimized mass transfer against pressure drop. Optimal voltages were 0.9–1.3 V/cell-pair at 9 L/h, and decreased at higher rates. Lastly, higher production was obtained primarily by increasing cell-pair area rather than number of cell-pairs (36–46). It was additionally observed that active area increased linearly with feed concentration (1500–2500 mg/L), while recovery (60–90%) minimally affected design. This work also suggests that voltage control during the batch process, and less expensive pumps, can further reduce cost.

Introduction

Domestic reverse osmosis (RO) systems are widely used in Indian homes to desalinate groundwater to a total dissolved salt (TDS) content that is suitable for drinking (less than 500 mg/L [1]), but they recover only 25–40% [2] of the feed. The domestic scale addressed here refers to point-of-use (POU) systems that typically produce 8–15 L/h of drinking water, store 7–10 L, weigh 8–11 kg, and are usually wall-mounted or placed on kitchen counters in individual homes [3,4]. Since the market for POU RO devices at this scale is forecast to grow at a compound annual growth rate of 18.2% between 2016 and 2024 [5], there is also a commercial incentive for developing more efficient desalination solutions that operate at the same scale.

Given that the concentration of the groundwater underlying a majority of India is under 2000 mg/L, electrodialysis (ED) can provide a higher recovery and more energy-efficient desalination compared to RO for this domestic application [6,7]. Similarly, growing concern over water scarcity and the need for more energy-efficient desalination has also recently revived an interest in the possibility of using ED for brackish water desalination and tap-water softening in European cities [8].

Despite the interest surrounding the use of ED for domestic purposes, little work has been performed to characterize the design of an appropriate ED system for the application. Pilat developed and piloted more than 200 domestic ED units before 2001, but little information regarding cost or the design of the system was provided [9]. More recently, Thampy et al. investigated a hybrid approach whereby ED was used to initially desalinate 2000–4000 mg/L water to 500 mg/L and further desalination to 120 mg/L or lower thereafter was achieved using RO [10]. Given that their small-scale system operated in a continuous process, without the recirculation of product water, only 50–60% of the feed supply was recovered. Instead, Nayar et al. showed that it was feasible to implement ED solely in a batch architecture (Fig. 1), where product water is recirculated, to desalinate from 3000 mg/L to 350 mg/L, at a competitive production rate of 12 L/h while providing 80% recovery [11]. However, their system was not designed to minimize capital cost which was an estimated $206 for the entire system, $138 of which was attributed to the ED stack and pumps.

While Nayar et al. have demonstrated that batch ED is a viable technology for satisfying household desalination needs, further cost reduction is required to be competitive with existing RO devices which are priced between $200–$300. Therefore, in this work we investigated the pareto-optimal design of the proposed domestic batch ED system considering production rate, product water concentration, and cost using simulation. In particular, we aimed to address the following:

  • 1.

    How should a domestic ED system be designed to minimize cost?

  • 2.

    How do water quality and production requirements affect the design?

  • 3.

    What are the primary contributors to cost?

  • 4.

    What developments are necessary for further cost reduction?

Prior design and optimization work has been performed for large-scale systems which are typically operated in a continuous architecture [12] for industrial applications. For these systems, the pump cost and energy consumption are often neglected because they are low relative to cost of the ED stack and the energy consumed by desalination [13,14]. Optimization at the domestic scale presents a different scenario where the pumps were found to strongly affect the cost, energy consumption, and performance of the ED system.

In addition, minimization of operating costs is often the most important consideration in industrial applications whereby the energy consumption can not be neglected [15]. In the present study, it was found that capital cost was the dominant factor affecting the affordability of the domestic system.

Section snippets

System description

The batch ED system (Fig. 1) proposed by Nayar et al. [11] and analyzed here consists of two primary flow circuits: one for the diluate, and the other for the concentrate. At the start of each batch process, both tanks hold feedwater at the same concentration. The relative volume of water in the diluate versus the concentrate circuits governs the recovery ratio of the process. During desalination, a voltage is applied and fluid is recirculated through the stack until the desired concentration

Models

The models used in this analysis have been thoroughly described and validated by Wright et al. [20]. However, a brief overview of the theory relevant to this optimization problem is presented herein to facilitate the reader's understanding of the work. For a more detailed description of the mass transfer processes in electrodialysis, Ortiz et al. [19], Strathmann [21], and Tanaka [22] also are recommended.

Following common practice, this work models desalination assuming a sodium-chloride

Optimization

The optimization problem of identifying the geometry and operating parameters which provided the lowest-cost system is presented in the following section. In the primary investigation described in 4.1 Problem formulation, 4.2 Variables and bounds, 4.3 Objective function, 4.4 Constraints, 4.5 Parameters, the feed concentration and recovery ratio are maintained at 2000 mg/L and 90%, respectively. The sensitivity to these parameters is then explored separately in a second problem formulation,

Results and discussion

In this section, cost-optimal designs obtained for varying production rate and product concentration requirements are first discussed at a fixed feed concentration and recovery ratio of 2000 mg/L and 90%, respectively.

Sensitivity to feed concentration & recovery ratio

In the previous section, the detailed design of an ED stack is investigated for a fixed feed concentration and recovery ratio of 2000 mg/L and 90%, respectively. Now we consider how the optimal design is affected when the feed concentration and recovery ratio are varied while the product concentration and production rate are maintained at 200 mg/L and within 12 ± 0.5 L/ h, respectively.

Conclusions

Cost-optimal designs of batch ED systems targeted at production rates of 9–15 L/h and product concentrations of 100–300 mg/L, from a fixed feed concentration of 2000 mg/L at 90% recovery, were first investigated. Voltage and flow-rates were held constant during the batch desalination process for each design.

In all cases, capital cost was found to dominate over the operating cost due to the upper-bound on the ion removal rate imposed by the limiting current density. Furthermore, capital cost was

Acknowledgement

This work was supported by the Tata Center for Technology and Design at MIT and Eureka Forbes Ltd. We also thank Professor Olivier L. de Weck for providing guidance and feedback on the optimization tasks.

References (45)

Cited by (0)

View full text