Page 206 - Membranes for Industrial Wastewater Recovery and Re-Use

S,ystetn design aids 1 7 5

when entering data into the design package. It is also invariably the case that
experimentally-determined ionic composition data do not yield an electrically
balanced specification. RO design software packages tolerate, to a certain extent,
an ion imbalance in the feed composition. Exceeding some predetermined limit of
this imbalance normally prompts a request to artificially adapt the total ion
balance by compensation (anion or cation). The total ion balance is then usually
corrected through the addition of sodium cations or chloride anions, since these
ions do not have a substantial impact upon the RO process design and operation.
Aggregate aspecific determinants, such as total organic carbon and total
dissolved solids, cannot normally be usefully entered into the design package.
Most design packages allow entry of data for the most common scalants, such as
the sulphates or carbonates of calcium, magnesium, barium, strontium and iron,
but not the more unusual compounds, such as sulphides of many divalent
metals, that can arise in some industrial effluents (Table 2.14).
The composition of the feed determines both osmotic pressure and scaling
propensity. These are increased both by the conversion, the ratio of permeate to
feed flow, and the degree of concentration polarisation (CP). In most software
packages a limit is placed on the CP coefficient (normally denoted as p-
coefficient). For example, for the RoPro package produced by Koch-Fluid Systems
the maximum p-coefficient value is 1.1 3. This limit is based on experience, rather
than anything derived from first principles. A warning is normally given when
the retentate scalant concentration exceeds its equilibrium solubility limit,
demanding remedial measures in the form of appropriate pretreatment (Section
2.4.3) or else reduced flux.
Other specific feedwater quality determinants of importance are the pH
and temperature. The feed pH is of particular significance since it has an impact
upon both scaling propensity and membrane integrity (see below). Precipitation
of hydrolysable scalants is suppressed at low pH levels and most RO CAD
packages allow for pretreatment with acid or base. The feedwater temperature
has a direct impact upon the permeate flux through the viscosity (Section 2.3.1 ),
as well as on scalant solubility and biological growth. Calcium carbonate
solubility and biological growth both decrease with increasing temperature,
whereas the viscosity (and hence the overall permeability) increases. Since the
latter is normally the most important temperature effect, it is crucial, when
designing the plant, to base the design on the lowest temperatures likely to
be encountered.
Membrane stability has already been discussed (Section 2.1.3). Reverse
osmosis membrane materials are all polymeric and have varying levels of
chemical stability. The two membranes materials most frequently employed are
cellulose acetate (low cost, chlorine resistant but limited rejection and subject to
alkaline hydrolysis at pH levels above 6) and aromatic polyamide (high rejection,
resistant to hydrolysis but more expensive and susceptible to oxidation by
chlorination), The choice of both the membrane material and module will
obviously depend upon the application, since parameters such as requirement
for sanitisation. feedwater temperature and pH fluctuations and fouling
propensity can be critical.