The use of membranes for solids removal in potable water treatment schemes has been known for hundreds of years. In the past, however, the supply of freshwater sources and the high cost of membrane treatment have made it impractical. Currently, with water shortages due to drought conditions or the limited availability of freshwater sources, and the reduced cost of reverse osmosis (RO) membrane technology treatment, municipal systems worldwide are deciding whether they can use seawater or brackish waters for drinking water sources.
Prior to making this decision, there are a few things that need to be considered. When designing treatment plants with seawater as the source, it must be remembered that RO membranes are intended for removing only salt and dissolved ions. Any particulate in the source water can potentially foul the membrane and create operational problems that will result in poor-quality effluent and increased cost.
The particulate material can be in the form of suspended solids that could be colloidal, precipitated, biological or organic by nature. The particulate material can deposit and become trapped on the surface and in the pores of the membrane. The biofouling occurs when biology forms colonies that grow into slime deposits in the feed spacer. All of these materials can foul membranes and cause systems to fail due to the buildup of increased pressure that consumes more energy, requires more cleaning, reduces flux and decreases recovery.
This situation can be improved with the use of pretreatment equipment that consistently provides raw water that is suitable as feedwater for RO membrane systems. Each raw water source is different and requires a custom approach. Gravity filtration is one of the pretreatment options that has proven to be effective. It can consistently remove solids so the membranes can perform as designed. Adding a coagulant to the raw water feed can precipitate soluble material and/or reduce the charge on the particles and flocculate the colloidal material. This particulate material will be removed by the gravity filter media prior to the effluent being fed to the RO membrane system.
In order to have some measure of the degree of this fouling problem, a testing procedure resulting in a silt density index (SDI) is used. The test procedure described in ASTM 4189-95 is performed using a 0.45-micron, 47-mm diameter filter. The water to be tested is supplied to the filter at a constant pressure of 30 psi. The test involves measuring the time it takes to collect a 500-ml sample through the filter at the start of the test and comparing it with the time it takes to collect a 500-ml sample after water has flowed through the filter (at 30 psi) for 15 minutes.
The sample times are applied to a formula, and the resulting value, SDI15, indicates the potential plugging of the membrane in percent-per-minute. On waters with high SDI, it is often useful to measure the SDI at 5- and 10-minute intervals. The resulting values, SDI5 and SDI10, can provide a better indication of the rate at which the membrane is plugging.
Utilizing the Red Sea
One system that is currently on line is in Jeddah, Saudi Arabia. This system uses the Red Sea for its source water. Its 30 pretreatment gravity filters are 52.5 by 13.9 ft and designed to process 126 mgd of water. The loading rate is designed for 4 gpm/ft2 with 750 mm of 0.5-mm sand and 1,500 mm of 1.2-mm anthracite for media.
The filters are designed with dual parallel lateral underdrains that have a concurrent air-water backwash system with air being delivered from a chlorinated polyvinyl chloride air header pipe located in the center flume of the filter. The air-water backwash provides “collapse pulse” cleaning of the media bed to remove solids deep in the media. The dual parallel lateral design ensures equal water pressure and flow at each opening in the underdrain.
The flow is uniformly distributed through a large number (23 per ft2) of 1/4-in. dispersion orifices. The 1/8-in. recess at the dispersion orifice prevents support material from resting directly on the orifice, allowing greater distribution and reduced headloss.
During normal conditions, the raw water inlet SDI15 is 4.75, and the plant feeds sodium hypochlorite followed by 0.3 mg/L of ferric chloride coagulant and 0.4 mg/L of 40% active polyDADMAC cationic polymer coagulant aid to treat the <3 NTU raw water turbidity. Sulfuric acid is added to the filter effluent to bring the pH down to 6.5, and sodium metabisulfite is injected continuously for seven hours of each eight-hour shift to dechlorinate the RO membrane feedwater. This protects the membranes from oxidation of soluble heavy metals that might foul the membrane.
To avoid biofouling, 0.2 mg/L of residual chlorine is allowed to pass through the membranes for one hour every shift. Under these circumstances, the gravity filters reduce the SDI15 to 2.85. When there are excursions where the raw water SDI15 climbs to 10 (usually due to biology in the source water), the ferric chloride dosages can be increased as high as 1 mg/L. Even under these conditions, the effluent water out of the gravity filters is consistently below 0.3 NTU turbidity with an SDI15 consistently below 4. This has resulted in suitable feed-water to the RO membrane process.
Eliminating fouling, decreasing cost
At the example plant, the gravity filter pretreatment has helped eliminate membrane fouling by removing biological solids and turbidity (suspended and colloidal solids), eliminating or reducing the physical deposits and trapped particles. Without the gravity filtration process, both conditions would have reduced recovery rates through the membrane process, reduced flux and increased cleaning, significantly increasing the cost of operations and decreasing the life of the membranes.