Reliability and ef ﬁ ciency of an advanced tertiary treatment process for wastewater reclamation

A reliability study for the reclamation of wastewater treatment plant ef ﬂ uent using continuous sand ﬁ ltration-multimedia ﬁ ltration (CSF-MMF) combined with ultra ﬁ ltration (UF) and reverse osmosis (RO) has been conducted. The objectives of the research are two-fold: (1) ef ﬂ uent of CSF-MMF can be used for surface water supplementation and (2) permeate of UF-RO can be applied as greenhouse irrigation water. The removal ef ﬁ ciencies for nutrients and electric conductivity (EC) as well as some operational parameters of the pilot plant were investigated. The concentration of T-N, COD and turbidity in the ﬁ ltrate of CSF-MMF with external C-course methanol dosage and FeCl 3 dosage could be kept at less than 2.2 mg/L, 35 mg/L and 0.9 NTU respectively. Average EC exceeded the required surface water standard by 10% and it was dif ﬁ cult to meet the low surface water standard for T-P (below 0.15 mg/L). The EC of RO permeate was below 20 μ S/cm, which was much lower than the standard for greenhouse irrigation. With frequent back ﬂ ushes, cleaning in place (CIP) and enhanced cleaning the UF could be operated with a constant permeability of 100 L/(m 2 ·h·bar). An appropriate CIP resulted in a recovery of 47 – 52% of the RO. The protective cartridge ﬁ lter prior to the RO should be replaced every 2 weeks.


INTRODUCTION
Increasing scarcity of freshwater resources and growing environmental awareness give rise to the use of reclaimed wastewater as an additional source of water supply on a worldwide scale, especially in areas where the climate is (semi) arid and/or the population and economic growth is fast (Hochstrat et al. ; Yang & Abbaspour ). Municipal sewage is promising as an alternative water resource and can satisfy non-potable water requirements such as streams in recreational parks, toilet flushing, irrigation and so on (Hidaka et al. ). Advanced treatment at existing wastewater treatment plants (WWTPs) is expected to provide a more feasible solution for wastewater reclamation compared to optimizing operation and upgrading of existing conventional biological nutrient removal steps.
It is indispensable to develop a cost-effective, compact and easy to maintain treatment scheme for sewage reclamation, which is capable of the removal of both nutrients and ions to some extent according to the required quality for reuse. In order to produce reliable and reusable water from WWTP effluent for European Water Framework Directive (WFD) and/or greenhouse irrigation water, an advanced effluent treatment scheme on a pilot scale was developed. This pilot consists of two steps: (1) continuous sand filtration (CSF) followed by multimedia filtration

Description of pilot process
The schematic diagram of the pilot process is shown in Figure 1. Effluent of WWTP is pumped to a high level buffer tank to supply water to the pilot. The effluent flows by gravity to the CSF and the flow can be adjusted.

CSF
The characteristics and main relevant operational parameters of the pilot CSF are given in Table 1 Table 2.

UF
The filtrate of the MMF is pumped to the UF by a feed pump. The UF unit is applied as pre-treatment prior to the RO to remove fine SS, coagulated colloidal materials and bacteria (Qin et al. ). The UF is operated in the dead-end mode. The UF unit contains a Capillary multibore *7 module (Innovative Membrane Technologies B.V.,

RO
To upgrade the water quality after UF, monovalent ions like Na þ and Cland other compounds have to be removed. In the whole pilot plant, every unit can be controlled independently from other parts. The only dependency is the availability of water from the preceding process unit.

Sample analysis
The

Pilot setup
The CSF and MMF were operated to check for leakage and that components were working properly after being placed on 5th February and 9th February respectively.
After 25th February, methanol (5%) and ferric chloride solution (40% FeCl 3 ) were dosed into the inflow of the CSF and MMF, respectively. The flow of CSF and MMF were controlled at 9 and 5 m 3 /h, respectively. After three weeks a removal of NO 3 --N in the CSF was observed, which indicated the growth of denitrifying bacteria in the CSF.
The UF and RO units were in operation from 10th March and 3rd April, respectively.

RESULTS AND DISCUSSION
Characteristics of the raw water The effluent of WWTP served as pilot feed water. The quality is presented in Table 3. The values in Table 3 for  Target water quality The objectives of the pilot are to produce surface water with CSF-MMF and greenhouse irrigation water with the dual membrane process. Some of the general maximum tolerable risk (MTR) standards for surface water in the Netherlands are summarized in Table 4. The main limits of the MTR for surface waters in relation to effluent reclamation are T-N below 2.2 mg/L and T-P below 0.15 mg/L. Greenhouse irrigation water requires EC below 500 μS/cm.

NO 3 --N removal in CSF and methanol dosage
After the CSF was placed and operated well, methanol (5%) was dosed into the inflow to cultivate a biofilm around the sand particles including denitrifying bacteria and heterotrophic bacteria. The flow rate and dirty water rate were controlled at 9 and 0.7 m 3 /h, respectively. From 15th March more than 50% NO 3 --N removal was observed in the CSF, which meant the denitrification in the CSF had been accomplished gradually. The NO 3 --N removal rate fluctuated from 14.3 to 96% and the average NO 3 --N removal rate reached over 60.9% in the four-month study period. The average effluent NO 3 --N was kept below 1.0 mg/L, as shown in Figure 2. Pure methanol dosages were 14.4 and 28.8 mg/L before 23rd April and after 24th April respectively. It also indicates that the CSF had a higher and more stable NO 3 --N removal rate when the influent NO 3 --N is over 4 mg/L. In some commercial CSF it was found that a high influent NO 3 --N concentration (above 10 mg/L) may be the reason for high removal rates of more than 85-95% NO 3 --N removal.
The DO of the effluent from the clarifiers ranged from 3.6 to 5.5 mg/L. According to Table 3   PO 4 3--P removal was studied in CSF and MMF. Figure 4 shows PO 4 3--P removal with and without FeCl 3 dosage According to Table 3 approximately 50% of the total phosphorus in the effluent is in the form of orthophosphate.  This means that with a complete removal of orthophosphate the remaining total phosphorus concentration is still twice as high as the required standard when WWTP effluent PO 4 3--P exceeding 0.3 mg/L.

Permeability of UF unit
The permeability of the UF unit during the study period is shown in Figure 5. During April, membrane fouling was programmed by a back flush after every 625 L of filtrate production, a chemical enhanced back flush with sodium hypochlorite solution (12.5% NaClO, CIP caustic) after every five back flushes. Manual caustic, acidic or caustic and acidic was applied frequently because of an unstable UF process. In May, the second running month, only automatic water backflush and caustic CIP was used for permeability recovery. The relatively high frequency of the caustic CIP means that this cleaning was not effective for the permeability recovery. On 14th and 17th June, a combined caustic/acidic CIP was carried out twice, which was followed by a caustic CIP 7 days later. After 24th June, the permeability of UF could be kept at 100 L/(m 2 · h·bar) for more than one month at a recovery rate of 60-80% only with programmed back flush. Before 21st June the UF  Performance of RO unit Figure 6 shows the post cartridge pressure and TMP of the RO unit during the study period. From May to July, the RO performed very well with one cartridge replacement nearly every 2 weeks and a CIP at 47-52% recovery rate.
The post cartridge pressure was maintained at 2.5-3 bar.
When TMP reached 1 bar, CIP ran automatically.
EC and pH of dual membrane Figure 7 shows the EC of UF feed, UF permeate and RO permeate during the study period. The EC of the UF feed ranged from 807 to 1,106 μS/cm. UF has no effect on the EC, which is logical because UF does not remove ions.
The EC of the RO permeate was very low at 9-17 μS/cm.
Average values for the pH of raw water, CSF effluent, MMF effluent, UF permeate and RO permeate are shown in Figure 8. The pH of CSF effluent increased slightly due to the formation of hydroxyl ions during denitrification.
The pH of RO permeate was 5.5-5.9, which was 1.

Targets accessibility
Filtration (CSF and MMF) and dual membrane filtration    The EC of the RO permeate varied from 7 to 15 μS/cm with an average of 10.5 μS/cm, which was well below the required value for greenhouse irrigation water.