Performance investigation of a novel design of vertical micro-screen drum filter for a recirculating aquaculture system (RAS)

The present study proposes a new design of mechanical filter to suit the recirculating aquaculture system (RAS) as a high-efficiency alternative filter to the conventional swirl separator (CSS). The new filter is considered the developed version of CSS; thus, it is called thereafter as the swirl-vertical drum filter (SVDF) which is provided with a rotating vertical micro-screen drum equipped with a fixed vertical brush as a self-cleaning tool instead of the conventional backwash system. The performance of SVDF was evaluated under different drum rotational speeds, and drum rotation times using sinking and floating fish feed pellets and magnetic field device, and then, it was compared to CSS in terms of water quality and filtration efficiency under the same operational conditions. Experimental results showed better values of water quality including Dissolved oxygen (DO), pH and unionized ammonia (NH3). Also, the lowest value of COD was found to be 18.33 mg/L by using SVDF with floating feed pellets, whilst the highest value of 100.73 mg/L was found in the case of using the CSS with sinking pellets. Meanwhile, the magnetic field device had no effect on the filtration process. The lowest surface area of the vertical micro-screen drum of 0.217 ± 0.013 m2 was obtained at the drum’s rotational speed of 20 rpm (0.39 m/s), rotation time of 150 min with floating feed pellets. Experimentally, it is found that the new SVDF has higher filtration efficiency than the CSS by about 57.57% under the same operational conditions.


Introduction
Global fish consumption as food for human beings increased at an average annual rate of 3.1% in the period from 1961 to 2017; this rate is almost twice that of annual world population growth (1.6%) for the same period. Per capita food fish consumption grew from 9.0 kg (live weight equivalent) in 1961 to 20.5 kg in 2018, by about 1.5% per year (FAO 2020). In this context, aquaculture accounted for 67.7% of the total fisheries production in 2017, reaching a staggering 53 million tons (Nie et al. 2020). According to the latest worldwide statistics compiled by FAO (2020), global fish production reached about 179 million tons in 2018. Egypt is the first fish producer in Africa where its aquaculture production reached 1,561,500 tons in 2018 (FAO 2020). Traditional aquaculture uses a substantial volume of water and often the discharge contains substances that affect the environment negatively, conserving water resources is mandatory and desirable to protect the environment during this period of climate change that hits the globe (Estim et al. 2019), especially, the most cultivated aquatic species is freshwater fish, which account for 56% of the total aquaculture production (Ottinger et al. 2016).
Therefore, aquaculture ponds' water should be recycled using different types of filters to remove the solid wastes and for resolving the expected problem of shortage in freshwater resources. This process is called recirculating aquaculture system (RAS) (Guerdat et al. 2013;Yogev et al. 2017). Compared to other forms of aquaculture production, RAS reduces the potential environmental impacts such as eutrophication and water dependence (d' Orbcastel et al. 2009). Solid wastes in the RAS usually originate from feeds that are composed of organic matter with a low density and wide range of particle sizes (Lee 2010). The solid organic wastes from uneaten feed and fish feces can reach up to 23% of the total feed supplied and can have a large effect on total solids production (Lee et al. 2009). Managing the levels of solid wastes in the RAS is very essential to maintain high-quality water for aquaculture. The RAS processes include several unit operations, such as a culture tank, solid waste removal, anaerobic digestion, and disinfection (Ahmad et al. 2022). RAS is proven to be a viable solution to reduce water waste and minimize water usage for aquaculture (Martins et al. 2011), improved sludge management (Piedrahita 2003), and the possibility of decreasing the logistical problems in aquaculture facilities such as; organic pollution, ecological changes, and a deterioration of water quality (Gunning et al. 2016;Akizuki and Toda 2018). Understanding of sources and classification of waste in aquaculture systems is essential for solids management.
Aquaculture wastes are dependent upon many factors such as; the type of culturing system, water exchange rate, cultured fish species, stocking density, and feed quality (Mark and Alexander 2020). Solid wastes are often expressed as total suspended solids (TSS), total dissolved solids (TDS), total solids (TS), and total volatile solids (TVS) (Pfeiffer et al. 2008;Ali 2013;Schumann et al. 2017).
Total suspended solids (TSS) generation in a RAS is generally dependent on many physical and biological factors involved. This includes temperature, fish species and size as well as filtration components and efficiencies (Cripps and Kelly 1996). Solid loads in an aquaculture system generally range from 5 to 50 ppm. However, this range may be considered relatively low when compared to other industries, such as municipal wastewater treatment (Droste 1997). TSS in recirculating aquaculture system (RAS) tends to be maintained below the range 20-25 mg/l limits recommended in the literature and widely regarded as critical for some species like salmonids (Magor 1988;Becker et al. 2018). The maximum TSS concentration reported for salmonid (Sensitive Species) Although, the main disadvantage of rotating micro-screen filters is that a large amount of water is lost by backwashing the filter sieves which represents 0.2-2% of the total volume of filtrated water as well as the high energy consumption for backwashing, drum rotation, and water pumping to compensate the head loss through the filter. Moreover, the conventional micro-screen drum filter is partially submerged, wherein about 40% rotates above the water level for washing (Brinker and Bergheim 2005), which may lead to reducing the filtration area of the drum, consequently, unnecessary economic losses in the production process will occur. the micro-screen drum filter needs a high-pressure water jet washing system that easily damages of drum sieve as reported by Ni and Zhang (2007), and consequently, high maintenance costs as well as crushes large particles, resulting in producing fine particulates and reducing biological filtration efficiency. In light of the above, it was found that there is an urgent need to develop the design and enhance the performance of such important types of mechanical filters (i.e., micro-screen drum filter).
Thus, the novelty of the present work is to (i) fabricate and evaluate the performance of a new design of an integrated swirl separator with a vertical rotated micro-screen drum, thereafter called the swirl-vertical drum filter (SVDF), as a novel mechanical filter to be used for the first time for intensive aquaculture in RAS; (ii) avoid the use of complicated parts in constructing the SVDF in RAS to meet the needs of common consumers/operators in terms of cheapness and ease of operation; (iii) install a rotated vertical micro-screen drum inside the body of a conventional swirl separator to accomplish the filtration process in one device (i.e., one step) instead of the conventional two consecutive separated steps of sedimentation and micro-screen filtering (i.e., horizontal drum) to reduce the consumed power as well as the occupation area of RAS; (iv) increase the surface area of micro-screen drum as far as possible, because the common horizontal micro-screen drum losses about 40% of its filtration area due to its horizontal rotation; (v) set up a self-cleaning mechanism for the proposed vertical micro-screen drum in SVDF instead of using the conventional backwashing process by installing a fixed vertical brush alongside the new drum to remove the stuck solid particles and avoid the micro-screen clogging, and consequently, the required power and water to clean the micro-screen drum will be saved completely; (vi) investigate the effect of using a magnetic device on the filtration performance of the new SVDF in RAS.
Hence, the main objective of the present work is to construct a new, small-scale, simple in design, and inexpensive mechanical filter called SVDF by merging the vertical microscreen drum with the conventional swirl separator employed in RAS to be suitable for small aquaculture farms as unit terms of saving costs of construction and operation reducing the required power, increasing the filtration area of micro-screen drum, lastly, saving the backwashing power and water due to the self-cleaning system. Furthermore, the performance of the proposed SVDF was evaluated and compared to the conventional swirl separator solely on the same RAS for the intensive culture of Nile tilapia fish.

Description of RAS
The manufacturing of RAS's components including the new SVDF and the practical experiments of the present study was carried out at the Central Laboratory of Aquaculture Research (CLAR), Abbassa, Abo-Hammad, Sharkia Governorate, Egypt. The RAS was constructed as described by Fernandes et al. (2015) to investigate the performance of the proposed SVDF compared to CSS. As depicted in Fig. 1, RAS mainly consists of circular tanks for fish rearing, centrifugal pump, mechanical filter, bio-filter, magnetic device, aeration system, pipes, valves, and other fittings. Two circular tanks made of fiberglass were used as fish rearing units with 1 m 3 in capacity for each tank and a centrifugal pump (Power = 0.75 hp, H max = 40 m, Q max = 2.4 m 3 /h) used for recirculating the water in RAS.
A magnetic field device (model: Delta Water DW-DP80-1 INCH, Egypt) made of stainless steel with a diameter size of 25.4 cm, a flow rate of 12 m 3 , withstand temperature up to 80℃, pressure up to 7 bar, and salt grade water treated up to 800 ppm. This device was used for the water treatment process in RAS including, reducing the effect of water hardness, preventing the formation of lime-scale, removing rust; salt and chemical free…etc.
A central aeration unit is used for the oxygenation process by injecting the fresh air into the fish rearing tanks for 12 h a day intermittently via an air stone. A granular filter was used as a bio-filter for the removal of ammonia wherein; gravels were used as a filter medium for biological filters, in which water flows through the gravels from up to down for the nitrification process by the bacterial colonies consisting on the outer surface of gravels. Polyvinyl chloride (PVC) pipes with diameters of 2.54, 3.81, and 5.08 cm and valves of 3.81 and 5.08 cm in diameter were used for piping the water through the system.
In the present case, rearing tanks were stocked with a density of 50 fish/m 3 by mono-sex Nile tilapia (Oreochromis niloticus) with an average initial weight of 54 ± 0.5 g. Fish were fed manually twice a day at 9:00 am and 2:00 pm, respectively using two types of pelleted feeds, the sinking and floating pellets using the same standard ration, with 30% of crude proteins where the amount of feed given to fish was calculated as a ratio of 3% of total biomass according to Craig and Helfrich (2017).

The conventional swirl separators
The principle working theory of the conventional swirl separator (CSS) (called also as the hydro-cyclone) is that the particles are denser than water and this difference increases by the centrifugal force resulting in separation and followed by deposition of the solid particles suspended in the water in RAS. The hydraulic load on the swirl separator may be Fig. 1 Schematic diagram of the used RAS much higher, 20-25 m/h; however, the advantage of this type of filter unit is its quite simple construction with no movable parts; in addition, it is cheap to buy and is intensive. The major drawback of CSS is that it requires uniform water flow for optimal efficiency. If the water enters the CSS at a higher rate than the designated rate, the solid particles will flow out of the filter with the outlet water in the center. Additionally, it is not enough to remove all the wastes, particularly the tiny size particles or suspended solid waste, so there is an imperious necessity to use a more efficient design for the filtration process, and consequently high removal efficacy such as the micro-screen drum filter. The size of the CSS can be obtained by the following equation given by Lekang (2007).
where V s is the hydraulic load on a swirl separator (20-25 m/h), Q is the flow rate (L/h), and A is the circular section area of the swirl separator (m 2 ).
As seen in Fig. 2, the CSS that was used in the present study is made of steel in a cylindrical-shaped body at the upper part (70 cm in diameter and 80 cm in height) connected with a conical collector at the bottom part (60 cm in height and 60° angle of the cone) leading to the sludge collecting box (30 × 40 cm with 5.08 cm in diameter for outlet). A vertical pipe called as the central pipe of the treated water was installed at the middle of filter's body provided with bores at the upper end (3.81 cm in diameter) and it was extended to the outlet (5.08 cm in diameter) that allocated underneath the filter's body.

The novel swirl-vertical micro-screen drum filter
The novel swirl-vertical micro-screen drum filter (SVDF) was fabricated according to design calculations given by US Army (1978). In general, several factors should be taken into consideration in the hydraulic design of the filter containing a micro-screen drum including; the maximum flow rate, allowable head losses, the porosity of the medium, effective submerged surface area, drum speed, and characteristics of the feed. These factors are numerically combined in Boucher's filterability index (Boucher 1947) for water (Rushton et. al., 2000). The design steps for the vertical micro-screen drum used in SVDF can be described in the following steps: i. Input data ii. Wastewater flow: iii. Average flow, L/min iv. Peak flow, L/min v. Suspended solids concentration, mg/L vi. Effluent requirement, mg/L vii. Design parameters viii. Head loss across the micro-screen (m) ix. Initial resistance of clean filter fabric (m) at a given temperature and standard flow conditions. x. Filterability index obtained from a laboratory study. xi. Speed of drum (number of square meters of effective fabric that allows water to enter in the given time) xii. Design procedure xiii. Effective submerged area of the micro-screen drum can be calculated as follows: where A is the effective submerged area (m 2 ), M is a constant equal to 0.0267, Q is the total flow rate (L/min), C is the initial resistance of clean filter fabric at a given temperature, and standard flow conditions 0.206 m for 77-μm (manufacturer's requirements), n is a constant equal 0.1337, I is the filterability index of influent measured on fabric in use (from laboratory results) 0.5, S is the suggested speed of drum for enough cleaning (1.85m 2 /min), H is the head loss across micro-screen for optimum flow (suggested by the US Army 1978 ≅ 0.0152 m).
Afterward, the hydraulic rate of application can be calculated as follows where HR is the hydraulic rate (L/min.m 2 ), and A is the screen area (m 2 ).
The novel SVDF was fabricated based on the same geometry and dimensions of the filter's body of CSS mentioned previously in the "The conventional swirl separators" section. A micro-screen drum with woven mesh made of stainless steel with 77 μm (pore size) was installed on a PVC cylindrical frame with dimensions of 250 mm in diameter and 80 cm in length. The drum is mounted vertically inside the filter's body at the center using insulated bearings from up and down for smooth rotation, as depicted in Fig. 3.
Dim. in cm (1) micro-screen drum, (2) bearings (3) trestles, (4) outlet of treated water, (5) chain, (6) electric motor, (7) sludge collecting box, (8) outlet of sludge, (9) cleaning brush, (10) untreated water inlet, (11) by-pass pipe, (12) flange of the drum The drum rotates by an electric motor (gearbox 40 rpm,0.37kW,230/400V,1390 rpm) mounted on the outer surface of the upper part of the filter (the cylindrical part) and provides the rotation movement for the vertical drum through a chain movement mechanism which is consist of chain, sprockets, and tightener. Furthermore, a static cleaning brush consists of fiber bristles fixed on a stainless-steel frame of 80 cm in length, as shown in Fig. 4. The cleaning brush was fixed vertically along the periphery of the rotating microscreen drum to remove the stuck solid particles and prevent the drum's clogging. The brush is used in the SVDF as a self-cleaning mechanism for the drum instead of installing the conventional backwashing system.

Working principle of the novel SVDF
Based on pre-experiments for determining the proper flow rate of untreated water entering the new filter, the flow rate of 800 L/h was found to be optimal for the novel SVDF, because the flow rates higher than this mentioned rate led to over rising of water level inside the filter, consequently, spilling out the water from the filter.
The untreated water gets out from the bottom of the rearing tank carrying all wastes of rearing (faces and uneaten feed) by the recirculating centrifugal pump that forces the water with a maximum flow rate of 800 L/h through the RAS. By using two ball valves, the untreated water was allowed to flow through a magnetic device or directly to the new mechanical filter of SVDF. Water enters the filter tangentially generating centrifugal force that works together with gravity for separating the heavy solids. Suspended solids are removed by the vertical micro-screen drum, where wastes trapped on the outside screen surface area are removed by the static cleaning brush due to the rotation of the drum. The vertical drum rotates inside the water with direct contact with the cleaning brush which makes the cleaning process and removing wastes sticking to the screen of the drum easier inside the water. Wastes that were separated by centrifugal and gravity forces as well as the cleaning brush are collected and settled inside the sludge collecting box.
After the mechanical filtration process, water flows to a bio-filter for more cleanness, and then the treated (clean) water is returned to the rearing tanks in a closed loop. The sludge collector was discharged after filtering 10 m 3 of water about 15 to 20 lof sludge (0.15-0.2% of the total volume of water containing the sludge).

Experimental procedure and determinations
In the present work, the performance of novel SVDF is compared to CSS in terms of water quality including dissolved oxygen (DO), temperature, pH, unionized ammonia Fig. 4 The vertical cleaning brush 1 3 (NH 3 ), and chemical oxygen demand (COD) as well as the filtration efficiency. Preliminary experiments were conducted to determine the optimal rotational speed range of the micro-screen. It was found that speeds less than 5 rpm showed a sharp rise in the water level inside the filter, while speeds higher than 20 rpm showed stability in the water level. So, three different rotational speeds for the micro-screen drum of 5, 10, and 20 rpm corresponding to drum peripheral speeds of 0.13, 0.26, and 0.39 m/s were used in the present work.
According to the results of the preliminary experiments, SVDF filter was working for 150 min daily as full time (FT), which was the essential time to recirculate 2000 L of aquaculture water in the rearing tank used in the present study. Therefore, the performance of the new SVDF was evaluated using three different rotation times of the microscreen drum 75 min (50% of FT), 112.5 min (75% of FT), and 150 min (100% of FT) at a flow rate of 800 L/h. Sinking and floating fish feed pellets with a typical formula for the ration were used as a different diet for fish feeding under the effect of a magnetic field.

Water quality analysis
Water samples were collected daily at the inlet (untreated water) and outlet of the filter (treated water) either for CSS or the novel SVDF; afterwards, the samples were stored in refrigeration for further analysis. The equipment used in the present study for measuring water quality is shown in Table 1. Dissolved oxygen, temperature, and total ammonia were tested five days a week. COD was measured three times a week by the heat of dilution procedure according to Boyd (1979) and outlined in Boyd and Tucker (1992). Total suspended solids (TSS) were determined three times a week as dry mass concentration per volume (mg/L) according to Boyd (1979). Unionized ammonia (NH 3 ) is calculated by multiplying the total ammonia value by the appropriate factor according to the following relation: where A is a coefficient related to water pH and temperature at sampling time (Emerson et al. 1975).

Filtration efficiency
Filtration efficiency is commonly used to evaluate the removal efficiency of the mechanical filter. The efficiency of filtration is normally expressed as the percentage of the TSS removed by the filter by measuring the TSS concentration in the water entering and leaving the filter that which is calculated as: where η f is the filtration efficiency for total suspended solids (%), TSS m is the total suspended solids concentration at the inlet of the filter (mg/L), and TSS out is the total suspended solids concentration at the outlet of the filter (mg/L).

Screen surface area of the drum
The lateral drum surface area (m 2 ) of the drum was calculated daily during the operation of the filter, by monitoring the water level by using a vertical ruler fixed inside the filter according to the following relation: where A is the lateral screen surface area of the drum (m 2 ), r is the radius of the drum (m), and h is the height of the water level inside the filter (m).

Statistical analysis
The obtained data were subjected to three-way ANOVA, to test the effects of filter type, magnetic field, and feed type and explore their interactions on water quality parameters and filtration efficiency. The Duncan, multiple range tests was used as a posthoc test to compare between means at P ≤ 0.05. All the tests were done using SPSS software, version 15 (SPSS, Richmond, Virginia, USA) according to Dytham (1999).

Water quality
The ability to maintain water quality parameters including; dissolved oxygen (DO), temperature, pH, unionized ammonia (NH 3 ), and chemical oxygen demand (COD) within safety limits is the key factor to successful aquaculture production. All parameters of quality for water in the fish rearing units in RAS were measured throughout the experiment's period for CSS and the new SVDF. In the present study, the wastewater influent passed through both filters, CCS and SVDF, has the same characteristics by connecting the CSS to the fish rearing tanks in a typical experimentation day followed by connecting the SVDF to the same tanks in the next day, wherein all operating conditions were unified in terms of the same fish size, fish density, feeding regime, type of fish feed pellet (sinking-floating), feed formula…etc. The obtained results showed that DO during the experimental period was affected by the type of filter. As seen in Fig. 5a, the recorded values of DO using SVDF were found to be in the range of 5.67 to 5.98 mg/L which were higher than those of CSS which were found in the range of 4.55 to 4.93 mg/L under all other different operational conditions. Furthermore, water temperature inside the rearing tanks ranged from 25.33 to 26.09 °C either using CSS or SVDF, which indicated slight differences in water temperature for both filters as depicted in Fig. 5b. This means that the type of filter has no influence on water temperature in RAS under any operational parameter of the present study. The obtained data from the practical experiments show that the novel SVDF affected pH values in lower values for pH compared to those of CSS. As shown in Fig. 5c, pH values ranged from 7.53 to 7.78 and from 7.89 to 8.1 in the case of using SVDF and CSS, respectively under all other different operational parameters. It was found that NH 3 is affected by the filter type. In general, the NH 3 values in the case of using SVDF and CSS ranged from 0.0184 to 0.031 and from 0.0687 to 0.0989 mg/L, respectively, as seen in Fig. 5d.
The obtained data showed that using floating pellets gave less value for NH 3 than obtained by using sinking pellets as fish feed. Also, the flow of water through the magnetic field device in RAS showed a slight decrease in the NH 3 values under all of the operational parameters, which agrees with the result confirmed by Limbu (2015). This means that the SVDF has a high capability to reduce the toxic effect of the accumulation of ammonia in water that inhibits the Nitrosomonas and Nitrobacter bacteria that are considered essential bacteria for the nitrification process as described by Camargo and Alonso (2006), resulting in, more cleanness and proper environment for Nile tilapia fish by using the new filter.
Unlike the magnetic field device, the results of the present study showed that filter type and type of fish feed pellet had a clear effect on COD value. The lowest value of COD was 18.33 mgO 2 /L achieved in the case of using the new SVDF, floating pellets, and magnetic field, whilst the highest value of 100.73 mgO 2 /L was recorded using the CSS, sinking pellets, and without magnetic field, as depicted in Fig. 5e. This indicates the high removal efficiency of SVDF compared to CSS to remove a large number of organic solids from the aquaculture water in RAS.
To get deep insight into the performance of both filters, the CSS and the new SVD, the characteristics of the influent of wastewater passing through filters and the effluent in terms of water quality indicators are shown in Table 2. Table 2 indicates that the type of filter and fish feed pellet are very crucial parameters affecting the quality of the effluent delivered from the filters. As a general pattern, the best water quality of effluent was achieved by using the SVDF and floating pellet wherein; all values of water quality parameters are located in the appropriate level that was suggested be earlier works in literature, while the low water quality is observed by using CSS and the sinking feed pellets due to the instability and high losses of such types of pellets. Concerning the effect feed pellets on water quality parameters in the effluent, data show that using SVDF with floating pellets enhances DO level in effluent from 4.75 ± 0.15 (CSS with sinking pellet) to 5.89 ± 0.45 mg/l, pH was reduced from 7.96 ± 0.08 (sinking pellet using CSS) to 7.53 ± 0.07 and NH3 from 0.090 ± 0.06 to 0.01 ± 0.014, respectively.

Timmons and
Ebeling (2013) COD, mgO 2 /l 112.33 ± 1.03 98.56 ± 1.09 100.01 ± 1.05 92.90 ± 0.58 112 ± 0.97 20.18 ± 0.74 102.43 ± 0.81 18.33 ± 0.67 < 30 mgO2/l Svobodová et al. (1993) TSS, mg/l 27.83 ± 0.92 20.32 ± 0.7 24.33 ± 1.06 19.16 ± 0.42 27.83 ± 0.92 9.34 ± 0.78 24.88 ± 1.06 8.27 ± 0.65 < 20 mg/l Becker et al. (2018) The presented data in Table 2 can be explained as SVDF having an additional filtration step (micro-screen) that separates organic solids more than CSS. These organic solids are considered the main source of NH 3 and consume a large amount of oxygen for oxygenation processes, especially in the case of using the sinking type of fish feed pellet causing high concentrations of organic wastes due to its low stability and high loss in water, on the contrary of floating pellets that have high water stability and low loss of in fish rearing units as described by Thadeus et al. (2020). Additionally, the magnetic field device had no effective impact on the water quality parameters in both filters; this result is in agreement with Hassan et al. (2019) and Irhayyim et al. (2020).

The effect of rotational speed, rotation time, feed type, and magnetic field on the drum surface area of SVDF
Three rotational speeds of 5, 10, and 20 rpm corresponding to drum peripheral speeds of 0.13, 0.26, and 0.39 m/s respectively, three different rotation times of 75 min (50% of FT), 112.5 min (75% of FT), and 150 min (FT) using floating and sinking feed pellets, under magnetic field were investigated to investigate their effect on the drum surface area (the micro-screen surface, which is used for the filtration process of water per hour as design features). The obtained results showed that the rotational speed of the drum affected the drum surface area clearly, where it decreased linearly with increasing the rotational speeds from 5 to 20 rpm and rotation time from 75 to 150 min. Moreover, using the floating and sinking fish feed pellets led to a remarkable decrease in the drum surface area from 0.57 to 0.217 m 2 and from 0.679 to 0.314 m 2 , respectively, as shown in Fig. 6a and b.
This result is ascribed to the high rotational speed of 20 rpm and the continuous rotation for 150 min (FT) that would remove the accumulated solids wastes effectively, which works on the instability of solids trapped on the micro-screen, in addition to the continuous friction for a relatively long period of time between the cleaning brush and outer surface of the drum. So, the drum is cleaned appropriately which in turn leads to removing solids on the micro-screen immediately as soon as it reached the mesh, making its bores still open all the time for a better and continuous filtration process. Additionally, the continuous rotation and the high rotational speed of the vertical drum enhanced the centrifugal force inside the filter, consequently boosting the separation of heavy solids.
Floating and sinking fish feed pellets were used in the present study to evaluate their effect on the micro-screen drum surface area of the new SVDF. The obtained data showed that the feed type affected the drum surface area under all of the operational parameters. Based on the data obtained herein, floating feed pellets were found to be in the range of 0.217 to 0.57 m 2 which was less than those of sinking feed pellets which were found in the range of 0.314 to 0.679 m 2 under the same operating conditions as depicted in Fig. 7a and b.
The relevant results to the floating feed pellets having better water stability with no loss of pellets on the depth of the fish rearing tanks making water cleaner with less content of total solids. On the other hand, sinking feed pellets have less water stability than floating pellets, thus, the sinking feeds tend to be fractured and settled at the bottom of the fish-rearing tank making water more contaminated with feed wastes which increases the total suspended solid as approved by Thadeus et al. (2020). In summary, it is clear that the sinking fish feed pellets contributes to higher TSS in the influent that passed to the mechanical filters from the fish rearing tanks than the floating feed pellets whether in case of using CSS or SVDF wherein; the results of TSS presented in Table 2 confirms this behavior of the sinking pellets. This can be explained as the sinking pellets absorb a large amount of water, while feeding fish which can lead to rapid fractures and loose in texture of pellets, and consequently higher TSS in water compared to the floating pellets. As the presence of a high concentration of total solids in the water leads to rapid clogging of the micro-screen and increases the water level inside the filter which in turn leads to using the undesirable larger surface area for the drum.
As is known, the magnetic field affects the structure of water and hydrated ions as well as the physicochemical properties and behaviour of dissolved salts. In the present study, a magnetic field was used to examine its effect on the water quality parameters during the filtration process and the surface area of the micro-screen drum in case of using the new SVDF. According to the obtained data the magnetic field had no remarkable effect on the water parameters as the quality parameter indicated previously in Sect. 3.1, also it did not affect the drum surface area at all, as it is shown in Fig. 8a and b.  Fig. 6 The effect of rotational speed, rotation time and magnetic field on the drum surface area using fish feed: a floating pellets, b sinking pellets This result can be explained by the water used for RAS in the present study being tap water which has a salinity not exceeding 300 ppm. Moreover, it is clear that the concentration of salt in the used water in RAS has no bad effects on the water quality needs of Nile tilapia and did not cause any problems in the physicochemical properties of water. Hence, the magnetic field may be more effective in the case of using water containing a higher concentration of salinity such as brackish water and seawater in case of using RAS in areas suffering from lack of fresh water. Table 3 presents the interaction between filter type (CSS and SVDF), magnetic field (magnetic and non-magnetic), and feed type (floating and sinking fish feed pellets) and their effects on the water quality parameters including DO, pH, water temperature, NH3, and COD. The statistical analysis shows that the filter type affected significantly (P < 0.05) the DO, pH, NH3, and COD, and the feed type significantly (P < 0.05) affected only the COD. Meanwhile, the interaction between filter type × magnetic field, filter type × feed type, magnetic field × feed type, and filter type × magnetic field × feed type did not show significant (P > 0.05) effects on the water quality parameters at all.  7 The effect of feed type, rotational speed, and rotation time on the filtration efficiency of the SVDF significantly (P < 0.05) affected the drum surface area, but the contrary occurred for the magnetic field wherein; it has an insignificant effect on the drum surface area. The interaction between feed type × rotation time, feed type × rotational speed, rotation time × rotational speed, and feed type × rotation time × rotational speed showed a significant (P < 0.05) effect on the drum surface area. Meanwhile, the interaction between magnetic field × feed type, magnetic field × rotation time, magnetic field × rotational speed, magnetic field × feed type × rotation time, magnetic field × feed type × rotational speed,

Filtration efficiency
In the present study, the efficiency of the filter was determined by measuring the total suspended solids (TSS) concentration in the water entering and leaving the filter. A slight variation of Filtration efficiency of CSS was observed under all the operational parameters includes feed type (floating and sinking feed pellets) and using neither magnetic nor non-magnetic field is considered.  On the basis of presented TSS data in Table 2, using CSS reduced the TSS concentration from 27.49 ± 1.12 mg/l in the influent to be 20.32 ± 0.7 mg/l in the effluent in case of using the sinking pellets, and from 20.32 ± 0.7 to 20.32 ± 0.7 mg/l in the effluent using floating feed pellets. It was observed that the TSS in the discharged water from CSS still exceeded the appropriate/optimal level of < 20 mg/l that determined by Becker et al. (2018).On the other hand, The new SVDF shows a remarkable reduction in TSS from 27.83 ± 0.92 mg/l in the influent to 9.34 ± 0.78 mg/l in the effluent using sinking feed pellets, and from 24.88 ± 1.06 mg/l in the influent to 8.27 ± 0.65 mg/l in the effluent in case of using the floating pellets. It is clear that using SVDF reduced the TSS concentration obviously in the discharged effluent to be lower than the mentioned optimal level, thus, it is evident that SVDF has a potent capability to remove a considerable amount of the suspended solid wastes in effluent, and consequently, it is very appropriate for RAS.
According to the data of TSS during the practical experiments, filtration efficiency was calculated as an average using Eq. 6 which is found to be in the range of 27.6 to 28.37% (~ 28%), as seen in Fig. 9. The obtained results revealed that the maximum filtration efficiency for SVDF was found to be 66.87% using rotational speed of 20 rpm, rotation time of 150 min, floating feed pellets, and magnetic field. Whereas, the minimum efficiency was about 65.27% using the rotational speed of 5 rpm, rotation time of 75 min, sinking feed pellets and without magnetic field, as seen in Fig. 7.
It was observed that there is an unremarkable difference of about 2.30% between the highest and the lowest value filtration efficiency for the new SVDF. In general, operating SVDF at a high drum rotational speed of 20 rpm and full-time rotation (150 min) using sinking or floating feed pellets is recommended in the present study, as depicted in Fig. 9. This because they used less drum surface area and keep the water level stable inside the filter which is prevented the water from spilling out of the filter compared to other operational conditions. The obtained results showed a remarkable difference in filtration efficiency between CSS and SVDF, as illustrated in Fig. 9. This can be explained that the SVDF having an additional filtration device represented in the vertical micro-screen with pore size of 77 μm, in addition to centrifugal and gravity forces that can remove the big sizes of solids wasted efficiently. Moreover, the continuous cleaning process due to exist of the vertical cleaning brush that secures a clean surface of the micro-screen drum without clogging due that may cause by the suspended solid wastes. Therefore, SVDF is very potential for separating large amounts of waste compared to the CSS which depends on only the centrifugal and gravity forces for depositing the big size solid wastes. In the light of the above, the optimum filtration efficiency of SVDF was 66.87%, while it was about 28% for the CSS. It can be concluded that the new SVDF has higher filtration efficiency than the CSS by about 57.57% under the same operational conditions of RAS.
For effective treatment of total suspended solid wastes in the recirculate water, the commercial RAS must include a swirl/or radial separator followed by a micro-screen drum filter (horizontal type) for sufficient solids wastes removal efficiency.The swirl separator removes approximately 23% of the TSS according to Timmons et al. (2002), whereas the micro-screen rotating drum filter removes approximately 40-45% from TSS when using either settling device as a pretreatment (Twarowska et al. 1997;Summerfelt et al. 2004). This means that the filtration process is carried out through two devices in two consecutive stages with about 63-68% as total percentage of removing TSS.
On the hand, the new SVDF carries out the filtration process in one device through single stage to reach about 66% as total percentage of removing TSS for the floating and sinking feed pellet, as shown in Table 2. The compactness that featured new SVDF indicates to the less occupation area and lower consuming power, especially that the new filter does not use backwashing system that consumes more fresh water and power but uses a static vertical cleaning brush.

Conclusions
The present work proposed a simple and new design mechanical filter for a recirculating aquaculture system (RAS) as a closed loop for rearing Nile tilapia. The RAS consists of rearing tanks, mechanical filter, biological filter, recirculate pump, magnetic device, and connecting pipes and fittings. The new filter resulted from the development of the conventional swirl separator (CSS) by adding a rotating vertical micro-screen drum to the swirl separator provided with a self-cleaning tool represents in a fixed vertical brush instead of using a backwash system that consumes more power and water. The new filter is called the swirl-vertical drum filter (SVDF), which is used for the first time. The performance of SVDF was compared to CSS in terms of water quality and filtration efficiency under the same operational conditions. The performance of SVDF was investigated under three drum rotational speeds of 5, 10, and 20 rpm (0.13, 0.26, and 0.39 m/s), three rotation times for drum of 75 min (50% of FT), 112.5 min (75% of FT), and 150 min (100% of FT), using sinking and floating fish feed pellets and magnetic device. The major outcomes of this study are enlisted as follows.
i. Concerning SVDF, the lower surface area of the vertical micro-screen drum found to be 0.217 ± 0.013 m 2 that achieved at the rotational speed of 20 rpm (0.39 m/s), rotation time of 150 min (100% FT), and floating feed pellets. On the contrary, the higher value of 0.6799 ± 0.011 m 2 was obtained at the rotational speed of 5 rpm (0.13 m/s), rotation time of 75 min (50% FT) and sinking feed pellets. ii. The maximum DO values were found to be 5.98 and 4.93 mg/L for SVDF and CSS, respectively at water temperature in a range of 25.33 to 26.09 °C in rearing tanks. iii. pH and NH 3 values ranged from 7.53 to 7.78 and from 0.0184 to 0.031 mg/L, respectively for SVDF, whereas from 7.89 to 8.1 from 0.0687 to 0.0989 mg/L, respectively for CSS. iv. It was found that the filter type and type of fish feed pellet had a clear effect on COD value. Results revealed that the lowest value of COD was 18.33 mg/L by using SVDF with floating feed pellets, whilst the highest value of 100.73 mg/L was found in the case of using the CSS with sinking pellets. v. The magnetic field device has no effect on the water quality or the performance of either the SVDF or CSS. vi. The filtration efficiency of SVDF was found to be ~ 66% which was higher than of CSS (~ 28%). Hence, it is clear that the novel SVDF has filtration efficiency higher than the CSS by about 57.57%. vii. The total percentage of removing TSS for SVDF is ~ 66% obtained from a single device without backwashing system, whilst the percentage of removing TSS for the conventional RAS is 63-68% obtained from two consecutive devices according to earlier works in literature. Hence, SVDF featured with less occupation area and consumed power as well as saving the backwashing water.

Recommendation
It is recommended to use the new mechanical filter (SVDF) that designed and presented in the present work due to its potentiality for the RAS instead of CSS. The new filter can be used with high efficacy for removing the suspended solid wastes whether with the floating or sinking feed pellets. However, the magnetic field device may have a clear impact in case of using salty water for RAS in areas suffering of lack of freshwater. permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.