Enhancing Sustainability in Wine Production: Evaluating Winery Wastewater Treatment Using Sequencing Batch Reactors ^†

Abstract

Wine production generates a high volume of wastewater with a significant fraction of biodegradable organic matter that must be removed before safe release into surface waters. Aerobic sequencing batch reactors (SBRs) have been successfully applied in the treatment of a wide range of wastewaters. However, only a few studies have described the use of the SBR process for the treatment of winery wastewater (WW). The effectiveness of using an aerobic SBR process was investigated for the treatment of WW using two activated sludge concentrations (i.e., 2 and 4 g_VSS L⁻¹) and nutrient-supplemented conditions. In nutrient-deficient conditions, COD removal efficiencies varied between 70% and 97% depending on the organic loading rate (OLR). In nutrient-supplemented assays, COD removal efficiencies remained above 91% in all conditions tested. However, the effluent quality decreased due to the increase in the total suspended solids concentration. Furthermore, the COD concentration of the treated effluent was unable to meet legal requirements (<0.150 g L⁻¹) for safe wastewater discharge. Therefore, longer aeration periods and settling phases may be required in order to improve effluent quality under high organic loadings. Overall, these findings demonstrate the potential of SBR as a biological WW treatment process.

1. Introduction

From an economic standpoint, wine production plays a fundamental role in many countries. However, the viticulture and winemaking sectors have long overlooked the environmental concerns associated with wine production. In Europe, several wineries still face challenges in wastewater management and fail to meet the legal limit requirements for water discharge or reuse due to a lack of adequate treatment practices [1].

For every litre of wine produced, between 0.5 L to 14 L of wastewater can be generated from several processes, including grape rinsing and de-stemming, pressing grapes into must, cleaning installations, fermentation barrels, and even wine losses into the waste stream [2]. These waste streams are rich in biodegradable organic matter, with typical chemical oxygen demand (COD) concentration varying between 0.3 and 49 g L⁻¹ [3]. Nutrients such as nitrogen and phosphorus can also be present in concentrations ranging from 10 to 415 mg L⁻¹ and 2.1 to 280 mg L⁻¹, respectively [3]. However, the characteristics of winery wastewater (WW) strongly depend on the winemaking stage and the technology applied. Although the oscillatory characteristics of WW streams emphasize the complexity and challenges associated with its treatment, preventing the discharge of untreated or partially treated WW is crucial to mitigate the risks of nutrient leaching to groundwater and the eutrophication of surface waters.

The choice of a specific treatment process depends on several factors, such as the size and location of the wineries, the volume of wastewater generated, and its organic content, as well as the capital investment and operating costs [4,5]. Several treatment processes, including physical, chemical, and biological methods have been studied to treat WW [6]. Advanced oxidation processes (AOPs) have emerged as a promising treatment process to successfully achieve efficient WW treatment [7,8]. However, the high cost of reagents and energy associated with AOP hinder their widespread application [5]. Biological treatment is widely recognized as a cost-effective and environmentally friendly approach to treating WW [9]. Aerobic biological processes, such as aerated storage tanks and conventional activated sludge systems, are commonly employed for the treatment of these waste streams due to their simplicity and high efficiency [9]. The sequencing batch reactor (SBR) has been successfully used to treat several industrial wastewaters, due to low infrastructure and energy requirements [10]. Furthermore, SBR processes offer significant advantages due to their simple automation, flexible operation, and low operating costs when compared to conventional activated sludge systems [6,9,11]. Despite these advantages, only a few studies have reported the use of SBR for WW treatment [12,13,14]. Therefore, the present study aims to determine the viability of treating winery effluents in SBR.

2. Materials and Methods

2.1. Winery Wastewater

The WW was obtained from a winery located in the Douro region in the north of Portugal. The WW was characterized in terms of chemical oxygen demand (COD), biological oxygen demand (BOD), total suspended solids (TSSs), volatile suspended solids (VSSs), total nitrogen (TN), and total phosphorus (TP), in accordance with the standard methods [15] (

Table 1. Composition of the winery wastewater.

	COD (g L−1)	BOD (g L−1)	TSS (g L−1)	VSS (g L−1)	TN (g L−1)	TP (g L−1)	COD:N:P
Winery Wastewater	68	55	14.7	12.8	0.663	0.258	100:1:0.4

). Five distinct COD_influent concentrations were tested to assess the treatment process efficiency using an SBR system. In order to obtain COD_influent concentrations from 3.0 g L⁻¹ to 18.5 g L⁻¹, the WW was diluted with tap water and then neutralized (pH 7) with sodium hydroxide.

2.2. Sequencing Batch Reactor Set-Up and Experimental Conditions

A total of 15 identical sequencing batch reactors (SBRs), each with a working volume of 4 L, were used in this work. The SBRs were operated in cycles of 23 h, including an aeration phase (21 h) and a settling phase (2 h). The feeding and effluent withdrawal phases were fast and had a negligible impact on the overall cycle time. During the aeration phase, air was introduced at the bottom of the reactors with a variable airflow rate, aiming to achieve a minimum dissolved oxygen concentration of 2 mg L⁻¹. All reactors were operated for at least 35 cycles to achieve steady-state, before concluding each experiment.

All reactors were inoculated with activated sludge from a local municipal wastewater treatment plant (Vila Real, Portugal). Two sludge concentrations were evaluated, namely 2 g_VSS L⁻¹ (X₂) and 4 g_VSS L⁻¹ (X₄). Specifically, reactors R1–R4 and R9–R11 were inoculated with 2 g_VSS L⁻¹ of activated sludge, while reactors R5–R8 and R12–R15 were inoculated with 4 g_VSS L⁻¹. Additionally, due to an imbalanced COD:N:P ratio in the WW, the effect of nutrient supplementation on the treatment process was also evaluated. Hence, to ensure suitable nutrients for cellular synthesis, nitrogen (15.4 mg_N g_COD⁻¹) and phosphorus (2.0 mg_P g_COD⁻¹) were added to the feed of R9–R11 (X_2+N) and R12–R15 (X_4+N). The volume exchange ratios (VERs) in R1–R8 and R9–R15 were 50% and 25%, respectively. A summary detailing the conditions applied to each reactor is presented in

Table 2. Main operating conditions applied to the SBRs.

Reactors	CODinfluent (g L−1)	OLR (g L−1 d−1)	Sludge Concentration (g L−1)	VER %	HRT (d−1)	Nutrient Supplementation
R1	3.0	1.6	2	50	1.9	No
R2	6.1	3.2
R3	8.9	4.6
R4	12.2	6.4
R5	3.0	1.6	4	50	1.9	No
R6	6.1	3.2
R7	8.9	4.6
R8	12.2	6.4
R9	8.9	2.3	2	25	3.8	Yes
R10	12.2	3.2
R11	18.5	4.8
R12	6.1	1.6	4	25	3.8	Yes
R13	8.9	2.3
R14	12.2	3.2
R15	18.5	4.8

.

3. Results and Discussion

3.1. COD Removal Efficiency

In this work, a total of 15 distinct SBR operating conditions were evaluated to determine the viability of using SBR technology for the treatment of WW, focusing on COD removal efficiency and effluent quality (i.e., COD and TSS). After obtaining steady-state in each tested condition, the influence of the OLR and F/M ratio on the COD removal efficiency was assessed (Figure 1). In nutrient-deficient conditions, (i.e., without nutrient supplementation), the COD removal efficiency exhibited a decreasing trend, from 97% to 70%, with increasing OLR regardless of the initial sludge concentration (Figure 1a). However, in assays with nutrient supplementation, the COD removal efficiency remained above 91% with OLR up to 4.8 g_COD L⁻¹ d⁻¹ (Figure 1a). Similarly, the increase in the F/M ratio resulted in a reduction in the COD removed, in assays without nutrient supplementation (Figure 1b). Moreover, at a similar F/M ratio, COD removal efficiency was considerably lower in assays performed with 4 g_VSS L⁻¹, suggesting the occurrence of mass transfer limitation at a higher sludge concentration. However, this limitation was not observed in nutrient-supplemented assays. In fact, a COD removal efficiency of 97% could be sustained for F/M ratios up to 1.6 g_COD g_VSS⁻¹ d⁻¹, which is higher than F/M ratios commonly applied in conventional activated sludge processes [16]. It has been suggested that F/M ratios up to 1.4 g_COD g_SS⁻¹ L⁻¹ can be applied in the treatment of high-strength organic wastewater in SBR processes [17]. In this work, a COD removal efficiency of 91% was attained at an F/M ratio of 2.4 g_COD g_VSS⁻¹ d⁻¹ with nutrient supplementation.

3.2. Effluent Quality and Sludge-Settling Properties

The final treated effluent quality of each assay was evaluated in terms of COD_effluent and TSS_effluent concentrations (Figure 2). In assays performed without nutrient supplementation, a clear degradation of the effluent quality was observed with the increase in the OLR applied, as shown by an increase in COD_effluent concentration (Figure 2a). At the highest OLR tested, the COD_effluent concentration reached 3.5 g L⁻¹ at SBR steady-state. Nonetheless, at an OLR of 1.6 g_COD L⁻¹ d⁻¹, the COD_effluent concentration was below the legal discharge limit (i.e., 0.150 g_COD L⁻¹) without the need for nutrient supplementation and with both sludge concentrations (Figure 2a). On the other hand, in nutrient-supplemented assays, the COD_effluent concentrations were rather stable, remaining, for the most part, below 0.5 g L⁻¹, with the only exception observed in the assay performed at an OLR of 4.8 g_COD L⁻¹ d⁻¹ and 2 g_VSS L⁻¹ of sludge (Figure 2a). In this condition, the COD_effluent reached 1.67 g L⁻¹ due to a mass transfer limitation caused by a high F/M ratio (i.e., 2.4 g_COD g_VSS⁻¹ L⁻¹), as discussed previously (Figure 1b and Figure 2a).

Nutrient supplementation had a negative impact on effluent quality in terms of the TSS_effluent concentration, which increased significantly with the increase in organic loading (Figure 2b). These results suggests that, in nutrient-supplemented assays, a higher availability of nutrients stimulated sludge growth. The newly formed sludge was unable to settle fast enough in order to remain inside the system, contributing to an increase in TSS_effluent concentration (Figure 2b). In the nutrient-deficient assay, TSS_effluent concentration increased slightly, but still remained below 0.6 g L⁻¹ (Figure 2b).

The sludge-settling properties were monitored in all reactors, in terms of sludge-settling velocity and sludge volume index (SVI) (Figure 3). Assays performed with a sludge concentration of 2 g_VSS L⁻¹ exhibited the highest settling velocities across all organic loading conditions. Furthermore, in nutrient-supplemented assays, the highest sludge-settling velocity (i.e., 2 m h⁻¹) was reached at an OLR of 4.8 g_COD L⁻¹ d⁻¹, which was four times higher than in nutrient-deficient assays for similar organic loading (Figure 3a). Assays performed with a sludge concentration of 4 g_VSS L⁻¹ exhibited low settling velocities (i.e., <0.14 m h⁻¹), regardless of the OLR applied (Figure 3a). These results suggest that in assays with a high sludge concentration, the settling velocity was hindered due to the development of sludge with a poor floc structure. In all conditions tested, the SVI decreased with the increase in organic loading, varying from 219 mL g⁻¹ to 77 mL g⁻¹ (Figure 3b). A steeper improvement in the sludge-settling properties was observed in assays performed with 2 g_VSS L⁻¹ of activated sludge and in nutrient-supplemented conditions. In activated sludge systems, severe sludge-bulking problems may occur when SVI is above 250 mL g⁻¹, while values over 150 mL g⁻¹ generally indicate sludge with poor settling properties [18]. In this study, in assays performed with 4 g_VSS L⁻¹ and nutrient supplementation, the SVI values reached above 170 mL g⁻¹ (Figure 3b). Furthermore, limited bulking conditions (above 150 mL g⁻¹) were observed in the other assays, particularly at an OLR below 3 g_COD L⁻¹ d⁻¹, although severe sludge bulking problems were not detected in this study (Figure 3b).

4. Conclusions

WW was efficiently treated using an aerobic SBR process. High COD removal efficiencies (>97%) were attained at an OLR of 4.8 g_COD L⁻¹ d⁻¹ in nutrient-supplemented assays performed using 4 g_VSS L⁻¹ of sludge. While nutrient supplementation allowed for high COD removal efficiencies across all OLRs applied, the effluent quality was unable to meet legal requirements for safe wastewater discharge. Further work can be carried out by adjusting the SBR cycle length in order to improve the effluent quality in high-rate organic loading conditions and/or adding a chemical process as a tertiary treatment.