Crossflow velocity

As most of the membrane separation processes arc operated in the crossflow mode for the reasons discussed earlier (Section 5.5.1 Crossflow Configuration), the crossflow velocity has marked effects on the permeate flux. A higher crossflow velocity typically results in a higher flux. The rate of flux improvement as a function of the crossflow velocity usually can be described by the following equation with specific units  

---

There are also two terms to account for the flow of entrained liquid carried around the bundle by crossflow of gas. Two terms are necessary because if a particular gas crossflow is reversed, then the subchannel from which the flow originates (the donor subchannel) has a different concentration of entrained liquid drops in the gas phase. It is assumed here that the crossflow of gas carries liquid droplets with it, and that the droplets adopt the same crossflow velocities as the gas. It is certain that the droplets do not behave in this way, but the effect of their actual behavior is not known. The equations then are... 

Flux. The film model (Equation 6.6) illustrates that increasing flux has an exponential effect on CP. If we accept that fouling is a consequence of CP the impact of excessive flux is obvious. As a result high flux membranes tend to be short lived and foul unless improved fluid management is able to enhance k. Selection of the appropriate flux and crossflow velocity is a trade-offbetween capital and operating costs (see cost of fouling below). 

Figure 11.3 Partial fluxes of isoamyl alcohol, ethyl acetate, isoamyl acetate and ethyl hexanoate as a function of their feed crossflow velocity (bottom axis) and Reynolds number (top axis) in a singlechannel module, using a POMS-PEI composite membrane. Notice that external mass-transfer limitations are not fully overcome when soluteswith a high affinity towardsthe membrane are processed (Adapted from Ref. 32.)...

Figure 21.15 Droplet formation at the pore opening. Side view with vectors indicating the unit vectors, M and m at the pore perimeter and the advancing (0a) and receding (0r) contact angles. Vrj represents the crossflow velocity of continuous phase at height equal to droplet radius, and vm is the mean disperse-phase velocity.

In the microfluid dynamics approaches the continuity and Navier-Stokes equation coupled with methodologies for tracking the disperse/continuous interface are used to describe the droplet formation in quiescent and crossflow continuous conditions. Ohta et al. [54] used a computational fluid dynamics (CFD) approach to analyze the single-droplet-formation process at an orifice under pressure pulse conditions (pulsed sieve-plate column). Abrahamse et al. [55] simulated the process of the droplet break-up in crossflow membrane emulsification using an equal computational fluid dynamics procedure. They calculated the minimum distance between two membrane pores as a function of crossflow velocity and pore size. This minimum distance is important to optimize the space between two pores on the membrane... 

Figure 4.21 Effect of crossflow velocity on water flux of an alumina membrane for oil-water separation [Bhave and Fleming, 1988]...

Some more drastic measures than increasing crossflow velocity need to be taken to help reduce the flux decline over an economically acceptable period of operating time. One such technique frequently adopted in many porous inorganic membrane systems today is... 

Successful performance of inorganic membranes depend on three types of variables and their interactions. The first type is related to the characteristics of the feed stream such as the molecular or particulate size and/or chemical nature of the species to be separated and concentration of the feed to be processed, etc. The second type is membrane dependent Those factors are the chemical nature and pore size of the membrane material and how the membrane and its accessory processing components are constructed and assembled. The third type is processing conditions such as pressure, transmembrane pressure difference, temperature, crossflow velocity and the way in which the membrane flux is maintained or restored as discussed earlier in this chapter. 

Parametric studies of the effects of TMP, temperature and crossflow velocity on the permeate flux and protein retention rate have been conducted using 0.8 pm alumina membranes at a pH of 4.4. The maximum steady state flux is observed at a TMP of 3 bars. As expected, a higher crossflow velocity increases the steady state permeate flux, as illustrated in Figure 6.3 under the condition of 50 C, TMP of 5 bars and pH of 4.40 [Attia et al., 1991b]. The protein retention rate also improves with the inciease in the crossflow velocity. The permeate flux reaches 175 L/hr-m, accompanied by a protein retention rate of 97.5% when the crossflow velocity is at 3.8 m/s. This improvement in the flux corresponds to a reduction in the thickness of the external fouling layer. 

Shown in Figure 6.4 are the effects of the process stream temperature (35 to 50X) on the permeate flux at a crossflow velocity of 3 m/s and a TMP of 5 bars for a feed pH of 4.40 [Attia et al., 1991b]. The steady state flux increases from 20 to 130 L7hr-m as the temperature increases from 35 to 50 C. Correspondingly, the protein retention rate decreases from 98.6 to 95.6 %. This increase in the flux and decrease in the protein retention may be explained by the reduced viscosity, increased diffusion coefficients and increased solubility of the constituents in the membrane and solution as a result of raising the temperature. 

Figure 6.3 Effect of crossflow velocity on permeate flux and protein retention for an 0.8 pm alumina membrane [Adapted from Attia et al., 1991b]...

Figure 6.5 Permeate flux of Osh processing effluent sueam as a function of crossflow velocity and TMP [Quemeneur and Jaouen, 1991]...

Typical membrane flux data using a 0.2 pm alumina membrane for clarifying a red and a white wine is displayed in Figure 6.8. The TMP is 2 bars and the crossflow velocity is 4.5 m/s. The flux declines in the first two hours or so before reaching steady state values. 

The impact of membrane filtration is not limited only to the quality of red wines. It is also ol rved with white wines. Given in Table 6.8. are the microfiltration results of a white and a red wine using 0.2 pun alumina membranes [Castelas and Serrano, 1989]. The TMP and crossflow velocity are 2 bars and 4.5 m/s, respectively. The changes on the wine quality are very obvious in almost every property. 

Meunier [1990] tested the feasibility of this new approach with 0.8 pm alumina membranes under a transmembrane pressure of 3 bars and a crossflow velocity of 3 m/s at 5 C. The initial permeate flux declines to a level of about 20 L/hr-m. ... 

Alumina and other ceramic membranes of various microfiluaiion pore sizes have been used for the separation of yeast (saccharomyces cerevisiae) from the broth and the clarification of thin stillage [Cheryan, 1994]. A typical flux of 110 L/hr-m can be obtained with a crossflow velocity of 4 m/s and a transmembrane pressure of 1.7 bars. The crossflow velocity is found to markedly affect the membrane flux. Concenuation factors (ratios of final to initial concentrations) of 6 to 10 for both the broth and the stillage can be achieved. Backflushing with a frequency of every 5 minutes and a duration of 5 seconds helps maintain the flux, particularly in the initial operating period. The permeate flux for both types of separation reaches steady state after 30 to 90 minutes. 

---