Elsevier

Ecological Engineering

Volume 149, 15 April 2020, 105744
Ecological Engineering

Hydrodynamics of a hydroponic bed bioreactor with different substrate media

https://doi.org/10.1016/j.ecoleng.2020.105744Get rights and content

Highlights

  • Longer mean residence times do not always indicate greater mixing in a bioreactor.

  • Water dispersion within small fractured rock decreases as water velocity increases.

  • Dispersion in LECA-filled bioreactors is unaffected by water level and flow rate.

  • Mixing in media beds with LECA occurs faster than with fractured rock or lava rock.

Abstract

Water flow through hydroponic media beds has seldom been characterized, and little is understood about the effects of different substrate media and flow rates as they pertain to mixing within. The hydrodynamics of four hydroponic substrate media were assessed in a small-scale hydroponic media bed bioreactor using tracer tests and mixing tests. Four media commonly used in hydroponic media beds, including 12.7-mm (0.5 in) fractured rock, 25.4-mm (1.0 in) fractured rock, 25.4-mm (1.0 in) lava rock and 12.0-mm light expanded clay aggregate (LECA) were selected and placed into identical rigid plastic containers to test the mean hydraulic residence time (mean residence time (MRT)) and vessel dispersion numbers (dispersion numbers) at two water levels and two flow rates. Sodium chloride tracers were used in these tracer tests and electroconductivity (EC) was measured every second with an EC probe at the outlet to generate residence time distribution curves for each bioreactor under each set of conditions. Expanded clay had a 16–29% longer mean residence time in the low water level, low flow treatment compared to the other media. It was also the most consistent overall and did not show any significant differences between flow rates or water levels. The mean residence times of the small fractured rock treatments got longer as the water velocity decreased, with the highest mean residence time achieved at an estimated water velocity of 0.38 cm s−1. Dispersion through the expanded clay or large fractured rock was not affected by flow rate or water level. Dispersion in the small fractured rock treatment only appeared to be greater at the lowest water velocity tested. Mixing time was also significantly shorter in expanded clay treatments at the lower water level. In the mixing tests, small fractured rock was comparable to the expanded clay at the low water level, low flow treatments but mixed 59% slower than in expanded clay treatment when the velocity increased. Overall, expanded clay performed better than the other substrate media tested, while the small fractured rock only performed well at lower estimated water velocities. The large fractured rock was substandard to both of those media and the lava rock was inconclusive but appeared to slow the movement of the tracer due to its intricate structure.

Introduction

Media selection is an important consideration for flow dynamics in a hydroponic media bed or “hydroponic bioreactor.” The structure, size and type of media all affect the flow of water through the bed and thus influence nutrient uptake and availability for plants as well as the rate of adsorption and removal of contaminants. While light expanded clay aggregate (LECA) and lava rock have been commonly used in hydroponic media beds, various gravel and rock sizes have been used in subsurface flow constructed wetlands (Billore, Singh, Sharma, Dass, and Nelson, 1999; Vymazal, 2018; Wolverton, Mc Donald, and Duffer, 1983). LECA and lava rock have also been identified as potential filter media in subsurface flow constructed wetlands (Van Deun, Van Dyck, and Kempen, 2008; Dordio, Carvalho, Teixeira, Dias, and Pinto, 2010; Özengin and Elmaci, 2016), in part due to their adsorption capabilities.

While hydrodynamics have been modeled for subsurface flow constructed wetlands, few published studies have been completed for hydroponic media beds (Bodin, Persson, Englund, and Milberg, 2013; Crohn, Ruud, Decruyenaere, and Carlon, 2005; Garcia et al., 2004; Giraldi, de Michieli Vitturi, Zaramella, Marion, and Iannelli, 2009; Guo, Cui, Dong, and Liu, 2017). By understanding the hydrodynamics of water flow through a small-scale bioreactor at different flow rates and water levels, we can determine efficiency under various flow regimes as well as reach a conclusion about the best type(s) of media to select. If mixing in the hydroponic bioreactor is slow or poor, then nutrients will not be distributed evenly. Similarly, if nutrient-rich water entering the bed is not dispersed throughout the bed as it flows through, then plants may not receive adequate nutrients or could experience uneven growth. Studying the hydrodynamics will also help model hydroponic media beds so they can be scaled up without sacrificing performance and select the appropriate flow rates based on bed dimensions. The bioreactor designed for this study was intended for use in treatment of wastewater and for removing contaminants. Therefore, the hydrodynamics may be an even more important consideration than cost which may be the case for someone just building a hydroponic media bed to grow vegetables.

One method that is commonly used to characterize and evaluate the dispersion within a bioreactor is the residence time distribution (RTD). Simply put, the RTD is the distribution of times it takes for suspended particles to move through a continuous-flowing ‘reactor’, which in this case is the hydroponic media bed. The RTD particles will take various paths through a reactor, and each corresponding path will take a specific time to travel from inlet to outlet. The average time a particle spends in the reactor is known as the hydraulic residence time (HRT). The mean HRT can be obtained from the RTD function (Gao, Muzzio, and Ierapetritou, 2012; Garcia et al., 2004):τ=0tEttwhere τ = the MRT and E(t) represents the RTD function as a function of time and equates to:Et=QtCt0QtCttwhere C(t) = the concentration at time t and Q(t) = the flow rate at time t.

Models for continuous flow reactors typically assume ideal flow in either plug flow (PFR) or continuously-stirred (CSTR) reactors. However, we know that most bioreactors are non-ideal and experience both axial (up and down) and radial mixing (side to side). The axial dispersion model was developed to better describe flow in tubular PFR (Kovo, 2008; Mayer, Meuldijk, and Thoenes, 1996; Mills and O'Connor, 1992). Horizontal subsurface flow wetlands, which are similar to hydroponic media beds, are generally characterized as plug flow and can be modeled with the axial dispersion model, also referred to as the dispersion plug flow model (DPFM) (Chazarenc, Merlin, and Gonthier, 2003). The Peclet number, Pe, is a dimensionless number that describes the mixing in a reactor and is represented by:Pe=uL/Dwhere Pe is the Peclet number, u is the mean fluid velocity (m s−1), L is the length of the reactor (m) and D is the dispersion coefficient (m2 s−1). Therefore, a high Peclet number would indicate lower dispersion. The vessel dispersion number (Nd) is simply the reciprocal of the Peclet number (D/ uL) and is more commonly used as a measure of axial dispersion. Therefore, a higher vessel dispersion number indicates greater dispersion. The normalized variance can be used to calculate the dispersion coefficient (D) which in turn can be used to ultimately determine the vessel dispersion number:σθ2=0tτ2Ettτwhere σθ2 is the normalized variance:σθ2=2DμL2DμL21expμLD

In order to determine the RTD curve for a given reactor, a tracer must be used to measure particle residence time. Tracer tests involve injecting a measurable tracer of known concentration and volume at the inlet of the reactor and continuously measuring its concentration at the outlet. Common tracers include dyes, radioactive molecules and other nonreactive chemicals (Gao, Muzzio, and Ierapetritou, 2012; Newell, Bailey, Islam, Hopkins, and Lant, 1998). Tracers must be easily measurable and should be soluble in the reactor so that they do not settle out (Fogler, 2006). One substance that works well in a hydroponic reactor is sodium chloride. Sodium chloride is highly soluble in water, is cheap and can be easily measured in real time using an electrical conductivity (EC) meter. The resulting graph of effluent concentration over time is known as the ‘C curve’ and can be integrated to get the ‘E curve’ or RTD curve (Fogler, 2006). From the RTD, we can determine the Peclet number.

Hydrodynamics for horizontal subsurface flow constructed wetlands have been modeled on both commercial and pilot-scale using various approaches (Chazarenc, Merlin, and Gonthier, 2003; Garcia et al., 2004; Seeger, Maier, Grathwohl, Kuschk, and Kaestner, 2013; Zahraeifard and Deng, 2011). For constructed wetlands, gravel sizes are chosen within a recommended range that minimizes clogging over time, typically 5–20 mm in diameter (Vymazal, 2018). However, there are seldom studies to date that examine the differences in flow dynamics between acceptable media sizes of different shapes or materials such as fractured rock, LECA and lava rock.

RTD curves are a good starting point for examining the hydrodynamic behavior within a hydroponic media bed, and vessel dispersion numbers help us understand how nutrients and contaminants are spreading out within a substrate media. However, these parameters alone do not give us a clear picture about why one media could be better than another. Since the water in a hydroponics system will be continuously circulated, we can also characterize the mixing by measuring the total mixing time in the reactor. To do this, a tracer is injected at the input of the reactor and continuously measured at the output until the reactor becomes completely mixed or the tracer is uniformly distributed throughout the growbed. Faster mixing times will also indicate greater dispersion within the reactor and therefore better distribution of nutrients (and contaminants).

A careful examination of RTD curves, vessel dispersion numbers and mixing times at different flow rates and water levels within a small-scale hydroponic bioreactor will determine which media substrates are best as well as help optimize parameters for future hydroponic media bed designs.

Section snippets

Bioreactor construction

In order to model the hydrodynamics of a hydroponic bed bioreactor, a physical model was developed. Rubbermaid® 53-Liter Brute® storage containers (70 cm × 42.5 cm × 27.3 cm) were used as the bioreactors. Screens were positioned in the containers 13 cm from the outlet in order to hold the media in place while leaving space near the outlet for a pump and electrical conductivity (EC) probe. Screens were constructed using 0.64 cm mesh stainless steel hardware cloth fastened to contoured frames

Residence time distribution

Three runs for each treatment were completed. However, for the high volume, high flow tracer tests with the large rock media, one run was very different than the other two, so a fourth run was conducted. The EC data for each run were normalized against the background EC to determine the mean residence times. Resulting mean residence times with standard deviations can be seen in Table 3 with significant differences between treatment conditions identified by differing subscript letters. While

Conclusions

This experiment helped to better understand the flow dynamics through four substrate media that are commonly used in hydroponic and aquaponic media beds. Evaluation of the flow through various hydroponic media at different water volumes and pumping rates identified that LECA may be better suited for hydroponic plant growth than the other substrate media tested because it allowed for greater dispersal of incoming nutrient flows under a wider variety of parameters than did fractured rock and lava

Author contributions

No new authors contributed to this revision.

References (25)

Cited by (1)

  • Evaluation of a recirculating hydroponic bed bioreactor for removal of contaminants of emerging concern from tertiary-treated wastewater effluent

    2021, Chemosphere
    Citation Excerpt :

    The plant bioreactor was modeled after a horizontal subsurface flow constructed wetland (see Fig. S2), but instead of single pass treatment, the effluent would be continuously circulated and aerated in batch treatments to increase the probability of plant uptake and substrate degradation of contaminants. In addition, each bioreactor would be filled with light expanded clay aggregate (LECA), a commonly-used hydroponic substrate media, which allows for better dispersion than the media used in typical constructed wetlands (Recsetar et al., 2020). Furthermore, LECA has been reported to be an adsorbent material, which may serve as an additional removal pathway for CECs from tertiary-treated effluent (Dordio et al., 2010; Li et al., 2014).

View full text