DESIGN
Membrane flux. The design maximum membrane fluxFlux: Flux is the flow rate of water applied per unit area of the membrane and has units of volume/unit area/time. is a critical decision since it will determine the required membrane area and the number of membrane modules which is a significant system capital cost. A lower flux will increase the capital cost but too high of a flux will increase operational costs and issues, e.g., much more frequent backpulsingBackpulse or backwash: 1) Backwash is a procedure in which periodically the flow direction is reversed through the membrane for a short period of time in order to remove particulates accumulated at the membrane surface. 2) Backwash also refers to the waste water produced as a result of the backwash procedure. and chemical cleaningClean-in-place (CIP) or in-situ chemical cleaning: Clean-in-place is a procedure performed periodically to clean a membrane more thoroughly than backwashing can achieve in order to restore the permeability of the membrane towards baseline levels. The process uses chemicals such as citric acid and chlorine or others to remove accumulated foulants on the membrane.. Until a few years ago, the maximum flux was typically determined through pilot testingPilot testing: Pilot testing is the use of treatment units significantly smaller than the full scale plant but which would mimic the treatment of the full scale plant in order to determine design and operational factors. or the state may have set a value. However, there is now greater experience with membrane systems such that many are designed based on experience with similar systems and water qualities. However, pilot testing can be useful for determining an appropriate flux that balances capital and operational costs. The upper bound of acceptable fluxes is often called the "critical flux" above which unacceptably more frequent backpulse and cleaning is required. Importantly, the water quality going onto the membranes is a significant variable dictating the design flux. The worse the water quality, e.g., higher turbidityTurbidity: Turbidity is a measure of the cloudiness of a water as well as a gross measure of the amount of suspended solids in a water. and suspended solids, the lower will be the flux rate. If the expected turbidity exceeds 10 NTUNTU: NTU stands for nephelometric turbidity unit and is a measure of turbidity, i.e., the cloudiness of a water, and is a gross measure of the amount of suspended solids in a water. on a "sustained basis," then pretreatmentPretreatment: In membrane treatment systems, pretreatment encompasses all treatment processes prior to the membrane, e.g., strainers, flocculation, and sedimentation. should be evaluated for MFMicrofiltration (MF) membranes: Microfiltration membranes are typically hollow-fibers with a pore size range of approximately 0.1 – 0.2 μm (nominally 0.1 μm)./UFUltrafiltration (UF) membranes: Ultrafiltration membranes are typically hollow-fibers with a pore size range of approximately 0.01 – 0.05 μm (nominally 0.01 μm). systems. Additionally, a significant organic carbon level in the feed waterFeed water: The feed water is the water stream applied to the membrane unit. may warrant pretreatment, e.g., coagulationCoagulant aids: Coagulant aids are often polymers used to assist primary coagulants such as alum or ferric salts in order to achieve better settling or better filtering flocs. and possibly sedimentation. Dissolved solids are usually not a consideration for MF and UF systems since these systems are not intended to remove them unless it is a softening system with hardness precipitation upstream. And notably, if the water contains significant iron and/or manganese which may be precipitated deliberately or inadvertently and increase fouling of the membrane, then this should be considered in the design.
Temperature compensation. MF/UF systems are typically operated in a constant flux mode using variable speed pumpsVariable speed pumps: Variable speed pumps can adjust their rpm's in order to match the pumping rate with demands. to adjust to increasing transmembrane pressure (TMP)Transmembrane pressure (TMP): The transmembrane pressure is the difference in pressure between the feed water side and filtrate side of a membrane., and thus maintain constant flow (AWWA 2008). However, water viscosityViscosity: Viscosity is a measure of the internal resistance of fluids to flow, e.g., honey has a much higher viscosity than water. The viscosity of water will control how much pressure or head loss is needed to force it through a membrane. The viscosity of water will increase with decreasing temperature and thus winter operation will see higher transmembrane pressures to maintain the same flux. significantly affects the TMP with higher viscosity causing higher TMPs. In turn, water viscosity is a function of water temperature with colder temperatures having higher viscosity. Therefore, the choice of design flux must consider water temperature/viscosity variations during the year. If colder temperatures/higher viscosities cause the TMP to exceed upper limits, then more membrane area will be required in order to reduce the flux and thus the TMP.
Theoretical equations for membrane operation are presented in the section "Process description." Equations 1 and 5 from that section can be combined to give the equation below that shows the relationship between flux, flow rate, membrane area, viscosity and TMP:
Equation 1
Where:
Am = surface area of membrane, m2 (ft2)
Jt = flux at temperature T, L/hr/m2 (Lmh) or gal/d/ft2 (gfd)
Qp = filtrateFiltrate: Filtrate is the water that has passed through the membrane. flow rate through membrane, L/hr (gal/d)
Rt = total membrane resistanceResistance: Resistance or membrane resistance is a measure of the difficulty of passing water through a membrane due to the nature of the membrane itself or to foulants accumulated on the membrane surface., psi/gfd-cp
TMP = transmembrane pressure, psi
µT = viscosity in centipoises at water temperature T, cp
Viscosity varies with temperature according to the equation below:
Equation 2
Where:
T = water temperature, oC
If the membrane produces at a constant flow rate, Qp, and constant transmembrane pressure,TMP, and if the resistance, Rt, is also constant, then equation 1 gives the following relationship between the required membrane area at two different temperatures:
Equation 3
Where:
ADes = design surface area of membrane, m2 (ft2)
ARefT = surface area of membrane at a reference temperature T, m2 (ft2)
µDesT = viscosity in centipoises at design water temperature T, cp
µRefT = viscosity in centipoises at reference water temperature T, cp
In equation 3, the reference temperature could be the temperature at which the pilot testing was done, or as is commonly done, 20oC. In the latter case, the membrane area based on pilot plant testing would be adjusted to 20oC via equation 3 and then equation 3 used to find the required area at another temperature.
The EPA Membrane Filtration Guidance Manual (2005) recommends the following approach to determining the design membrane area:
1. Tabulate the average daily flow and average temperature over each calendar month.
2. Use equation 2 above and the data from step 1 to estimate the average viscosity for each month.
3. Use the paired average flow rate and viscosity for each month along with a TMP and Rt in equation 1 above to determine the required membrane area for each month.
4. Take the largest membrane area from step 3 and its associated flow rate to determine the flux for that worst case month.
5. If the worst case flux from step 4 is less than the maximum permitted flux, then use this area as the design area. If this flux is greater than the maximum permitted flux, then use the maximum permitted flux to determine the design membrane area.
Note that a somewhat more conservative design variant of the approach above has been suggested which is to use the maximum flow rate and minimum temperature in each month in step 1 above. Also note that the approach above requires values for TMP and Rt to use in equation 1. Good values for TMP should be readily available but values for Rt may not be clear especially since the total membrane resistance, Rt, is the sum of the intrinsic membrane resistance, Rm, and the resistance of the foulant layer, Rf, as in equation 6 in the section "Process description." Because of the difficulty in selecting values for the resistance of the foulant layer, Rf, the EPA Guidance Manual recommends using the intrinsic membrane resistance, Rm, as the estimate of total membrance resistance, Rt. But since Rm, would be the membrane resistance of a clean filter, i.e., at the start of a filter cycle just after backpulse/cleaning, the Guidance Manual also chooses the TMP at the beginning of the filter cycle rather than the maximum TMP at the end of the run. The filter manufacturer may of course have their own method for determining membrane area in which case the utility, state, manufacturer, and engineer should consult in choosing the design approach.
Cross-connection control. Cross-connectionCross-connection: A cross-connection occurs when untreated or non-potable water can enter the treated water system. In membrane systems, cross-connections must be prevented with the clean-in-place chemical system although there may be other potential cross-connections as well. controls are applied to MF/UF systems in order to prevent chemicals used in cleaning from contaminating either the filtrate or feed water. The most common method to prevent this is called the double block and bleed arrangement as shown in Figure PD.6. The system is designed such that valves V-1A and V-3A in the feed water line are closed in order to block the flow of chemicals into the feed water. Valve V-5A in the clean in place line (CIP) is also closed and Valves V-4A and V-6A are opened to allow the flow of chemicals to the membrane rack. Importantly, valve V-2A is also opened and serves as a bleed valve so that if valve V-3A leaks, then the leakage is discharged through valve V-2A rather than potentially entering the feed water line through valve V-1A. A similar arrangement is installed on the filtrate side.
Figure PD.6. Double Block and Bleed Valving Arrangement (USEPA, 2005)
Other system capacity considerations. In selecting the capacity of the membrane system, routine operational requirements such as backpulsing, chemical cleaning and integrity testing are typically accounted for. However, the engineer and utility must also consider the reliability of the membrane treatment system to meet water demands when units are down for repair and maintenance for longer periods. This is especially important for smaller systems where taking one unit out of service has a much greater impact on the system's ability to supply water. In order to provide a satisfactory level of system reliability, it is common to either:
1) Oversize the design flux so that the units in service can operate at a higher flux and thus make up for part of the lost capacity when a unit is down, or
2) Provide an additional membrane unit.
Of course redundant capacity must also be considered for other critical components such as pumps, blowers, compressors, and chemical feed equipment.
A valuable resource for the design, procurement, installation and commissioning of membrane systems is ANSI/AWWA's standard B110-09 (AWWA 2010).
