**3. Volumetric mass transfer coefficient corresponding with energy consumption**

In the event of an inflection point (marked in red in **Figure 4**), it is allowed to truncate the

, where *Ci*

ing to the inflection point. If the curve does not have an inflection point, the data can be trun-

To obtain time delay, the remaining data are extrapolated through the intersection of the fitted curve at the time axis (**Figure 5**). The primary data are corrected by shifting the curve to the initial moment that the concentration of dissolved oxygen is zero—see **Figures 5** and **6**.

**Figure 5.** Extrapolation of experimental data to obtain the delay time, for MP 1.6, at Q = 360 l/h.

**Figure 6.** Estimation of Kla and *Cs* parameters by non-linear regression for MP 1.6, at *Q* = 360 l/h.

represents the concentration correspond-

http://dx.doi.org/10.5772/intechopen.76696

355

Experimental Study of Standard Aeration Efficiency in a Bubble Column

curve up to the concentration at *<sup>C</sup>* <sup>=</sup> 1.5 <sup>⋅</sup> *Ci*

.

cated to 20% of the *Cs*

To obtain the oxygen transfer the following configurations are tested:


For all series of metallic plates, the following injected air flow rates are tested: *Q* = 180, 360, 480, 600, 720, 960, 1140 l/h. The results are then compared with the aeration performance of the CP and GP.

The method of measuring dissolved oxygen (DO) in clean water, according to the standard for measuring oxygen transfer in water [9], requires removing DO from water (using Na2 SO<sup>3</sup> ) and then reoxygenating up to at least 90% of the saturation concentration value. To obtain standard oxygen transfer rates (SOTR), the water in which testing takes place must be qualitatively equivalent to drinking water. Concentration of dissolved oxygen in time (*C*) is measured while maintaining constant air flow injected into the system. The measurements are repeated for each of the air flow rates with the abovementioned plates. After each set of measurements, the standard procedure is applied for the removal of DO from water and reoxygenation up to 90% of the saturation concentration value.

#### **3.1. Processing of the experimental data**

The following is an example of estimating the Kla and *Cs* parameters for the plate MP 1.6 operating at an injection air flow rate of *Q* = 360 l/h. The concentration of DO in time *C* = f(*t*) is shown in the figure (**Figure 4**).

**Figure 4.** *C* = f(*t*) and the inflection point detection.

In the event of an inflection point (marked in red in **Figure 4**), it is allowed to truncate the curve up to the concentration at *<sup>C</sup>* <sup>=</sup> 1.5 <sup>⋅</sup> *Ci* , where *Ci* represents the concentration corresponding to the inflection point. If the curve does not have an inflection point, the data can be truncated to 20% of the *Cs* .

To obtain time delay, the remaining data are extrapolated through the intersection of the fitted curve at the time axis (**Figure 5**). The primary data are corrected by shifting the curve to the initial moment that the concentration of dissolved oxygen is zero—see **Figures 5** and **6**.

**Figure 5.** Extrapolation of experimental data to obtain the delay time, for MP 1.6, at Q = 360 l/h.

**Figure 6.** Estimation of Kla and *Cs* parameters by non-linear regression for MP 1.6, at *Q* = 360 l/h.

**Figure 4.** *C* = f(*t*) and the inflection point detection.

**3. Volumetric mass transfer coefficient corresponding with energy**

To obtain the oxygen transfer the following configurations are tested:

• a ceramic plate (CP) with volume porosity in the range 45–50%; and

• series 2 of two metallic plates (MP 0.2 and MP 0.5);

genation up to 90% of the saturation concentration value.

The following is an example of estimating the Kla and *Cs*

**3.1. Processing of the experimental data**

is shown in the figure (**Figure 4**).

• series 1 of five metallic plates (MP 0.2, MP 0.3, MP 0.5, MP 0.9, and MP 1.6);

• a fritted (sintered) glass plate (GP) with porosity controlled in the range 0.25–0.315 μm.

For all series of metallic plates, the following injected air flow rates are tested: *Q* = 180, 360, 480, 600, 720, 960, 1140 l/h. The results are then compared with the aeration performance of

The method of measuring dissolved oxygen (DO) in clean water, according to the standard for measuring oxygen transfer in water [9], requires removing DO from water (using Na2

and then reoxygenating up to at least 90% of the saturation concentration value. To obtain standard oxygen transfer rates (SOTR), the water in which testing takes place must be qualitatively equivalent to drinking water. Concentration of dissolved oxygen in time (*C*) is measured while maintaining constant air flow injected into the system. The measurements are repeated for each of the air flow rates with the abovementioned plates. After each set of measurements, the standard procedure is applied for the removal of DO from water and reoxy-

operating at an injection air flow rate of *Q* = 360 l/h. The concentration of DO in time *C* = f(*t*)

SO<sup>3</sup> )

parameters for the plate MP 1.6

**consumption**

354 Desalination and Water Treatment

the CP and GP.

The chart (**Figure 5**) is replotted with corrected data using the mathematical model described

*C* = *CS* − (*CS* − *C*0) *e* <sup>−</sup>*klat* (6)

The procedure is repeated for all the test plates for the air flow rate range: *Q* = 96–1140 l/h.

estimation of MP 0.5 and MP 1.6 at injected

http://dx.doi.org/10.5772/intechopen.76696

357

Experimental Study of Standard Aeration Efficiency in a Bubble Column

parameters are obtained (**Figure 6**).

**Figure 8.** Estimation of Kla and *Cs* parameters by non-linear regression for MP 0.5.

**Figures 7** and **8** show the graphs for Kla and *Cs*

air flow rates, *Q* = 240÷1140 l/h.

by Eq. (6):

The Kla and *Cs*

**Figure 7.** Estimation of Kla and Cs parameters by non-linear regression for MP 1.6.

The chart (**Figure 5**) is replotted with corrected data using the mathematical model described by Eq. (6):

$$\mathbf{C} = \mathbf{C}\_{\rm s} - \left\{ \mathbf{C}\_{\rm s} - \mathbf{C}\_{\rm u} \right\} e^{-\text{kat}} \tag{6}$$

The Kla and *Cs* parameters are obtained (**Figure 6**).

air flow rates, *Q* = 240÷1140 l/h.

The procedure is repeated for all the test plates for the air flow rate range: *Q* = 96–1140 l/h. **Figures 7** and **8** show the graphs for Kla and *Cs* estimation of MP 0.5 and MP 1.6 at injected

**Figure 8.** Estimation of Kla and *Cs* parameters by non-linear regression for MP 0.5.

**Figure 7.** Estimation of Kla and Cs parameters by non-linear regression for MP 1.6.

356 Desalination and Water Treatment

The regression curves for the estimated parameters Kla and *Cs* for all nine tested plates at the

Experimental Study of Standard Aeration Efficiency in a Bubble Column

http://dx.doi.org/10.5772/intechopen.76696

359

From the aeration point of view (dissolved oxygen transfer), an improvement is observed between the first and second series of the MP, the second giving a better performance as the interphase area is higher. However, the MP aeration characteristics Kla and *Cs* are lower com-

Krishna and van Baten [10] illustrate the influence of column diameter on Kla (**Figure 10**), assuming that a homogeneous flow regime prevails (with dispersion consisting of 5 mm of small-sized bubbles). In this study, the diameter of the aeration device is equal to the diameter of the tank. A strong reduction of aeration characteristics is observed. The bubble column is confined by the walls and the flow velocity induced in the ascending column produces an increased velocity in the air column that tends to accelerate the bubbles in the central core,

In our study the water column is higher than the bubble column, *L/D* = 6.8, in order not to have the water column constrained by confinement. For this reason, Kla is less than in the

/s] is the air flow rate through the aerator and *A* [m<sup>2</sup>

surface of the plate (active area). The hypothesis of the uniform initial ascending velocity is a rough estimation, which is more realistic for MP (because of uniform orifice losses) but less so for GP and CP. Krishna and van Baten [10, 11], obtained from CFD, using the effective area of transfer, with large bubbles corresponding to 20 mm are represented in (**Figure 11**) by filled circles, and the 5 mm sized bubbles, are represented by filled squares. The open circles relate

A decrease in the Kla coefficient is observed with an increasing bubble diameter, as transfer performance is related to the interfacial area (for the same air flow rate). However, it is observed that Kla is comparable for two experiments. The difference is explained by the limitation of the bubble column size and the acceleration of the bubble column in Krishna's experiment.

**Figure 10.** Influence of column diameter DT (by numerical simulations) on Kla coefficient for operation in the

] is the emission

Krishna experiment, for equivalent bubble diameters (MP 0.5) around 5 mm.

to experimentally determined values and *Utrans,* is the transition velocity.

The initial theoretical ascending velocity of the bubbles can be calculated using:

injection flow rate *Q* = 360 l/min are presented for comparison in **Figure 9**.

pared with CP and GP. GP provides the best aeration characteristics.

reducing gas–liquid contact time.

*U* = *Q*/*A*, where *Q* [m3

homogeneous flow regime [10].

**Figure 9.** Estimation through nonlinear regression of Kla and Cs parameters for Q = 360 l/h. (a) MP 0.2, series 1. (b) MP 0.3, series 1. (c) MP 0.5, series 1. (d) MP 0.9, series 1. (e) MP 1.6, series 1. (f) MP 0.2, series 2. (g) MP 0.5, series 2. (h) Ceramic plates with volume porosity in the range 45÷50%. (i) Glass plates with porosity controlled in the range 0.25÷0.315 μm.

The regression curves for the estimated parameters Kla and *Cs* for all nine tested plates at the injection flow rate *Q* = 360 l/min are presented for comparison in **Figure 9**.

From the aeration point of view (dissolved oxygen transfer), an improvement is observed between the first and second series of the MP, the second giving a better performance as the interphase area is higher. However, the MP aeration characteristics Kla and *Cs* are lower compared with CP and GP. GP provides the best aeration characteristics.

Krishna and van Baten [10] illustrate the influence of column diameter on Kla (**Figure 10**), assuming that a homogeneous flow regime prevails (with dispersion consisting of 5 mm of small-sized bubbles). In this study, the diameter of the aeration device is equal to the diameter of the tank. A strong reduction of aeration characteristics is observed. The bubble column is confined by the walls and the flow velocity induced in the ascending column produces an increased velocity in the air column that tends to accelerate the bubbles in the central core, reducing gas–liquid contact time.

In our study the water column is higher than the bubble column, *L/D* = 6.8, in order not to have the water column constrained by confinement. For this reason, Kla is less than in the Krishna experiment, for equivalent bubble diameters (MP 0.5) around 5 mm.

The initial theoretical ascending velocity of the bubbles can be calculated using:

*U* = *Q*/*A*, where *Q* [m3 /s] is the air flow rate through the aerator and *A* [m<sup>2</sup> ] is the emission surface of the plate (active area). The hypothesis of the uniform initial ascending velocity is a rough estimation, which is more realistic for MP (because of uniform orifice losses) but less so for GP and CP. Krishna and van Baten [10, 11], obtained from CFD, using the effective area of transfer, with large bubbles corresponding to 20 mm are represented in (**Figure 11**) by filled circles, and the 5 mm sized bubbles, are represented by filled squares. The open circles relate to experimentally determined values and *Utrans,* is the transition velocity.

A decrease in the Kla coefficient is observed with an increasing bubble diameter, as transfer performance is related to the interfacial area (for the same air flow rate). However, it is observed that Kla is comparable for two experiments. The difference is explained by the limitation of the bubble column size and the acceleration of the bubble column in Krishna's experiment.

**Figure 10.** Influence of column diameter DT (by numerical simulations) on Kla coefficient for operation in the homogeneous flow regime [10].

**Figure 9.** Estimation through nonlinear regression of Kla and Cs parameters for Q = 360 l/h. (a) MP 0.2, series 1. (b) MP 0.3, series 1. (c) MP 0.5, series 1. (d) MP 0.9, series 1. (e) MP 1.6, series 1. (f) MP 0.2, series 2. (g) MP 0.5, series 2. (h) Ceramic plates with volume porosity in the range 45÷50%. (i) Glass plates with porosity controlled in the range 0.25÷0.315 μm.

358 Desalination and Water Treatment

**Figure 11.** Volumetric mass transfer coefficient as a function of U, from the homogeneous to the heterogeneous flow regime, (a) our experiments, (b) Krishna experiment's [9].

#### **4. Air flow influence on the aeration efficiency**

The experimental data were post-processed following standard procedures [9] and reduced to the same temperature and pressure conditions (*t <sup>w</sup>* = 20°C and *patm* = 1 atm) in order to ascertain the influence of hole size on the aeration efficiency.

As per the standard process the following steps was considered:

• volumetric mass transfer coefficient, estimated by regression and corrected at 20°C

$$\text{Kla}\_{20} = \text{Kla} \cdot \theta^{|\text{20} + i|} \,\text{[1/min]} \tag{7}$$

The power consumed for the injection of air through the aerator

injection.

of dissolved oxygen.

and glass plate GP).

*P* = *Q* ⋅ (*dp* + *gH*)/1000 [kW] (12)

Experimental Study of Standard Aeration Efficiency in a Bubble Column

http://dx.doi.org/10.5772/intechopen.76696

361

**Figure 12** shows the pressure losses measured for all MPs, CPs, and GPs. The injection losses on MPs are 10 times inferior to the CPs and GPs. The losses decrease with increasing diameter of the orifice but in a reduced report compared to CPs and GPs. The arrangement of the holes plays a role, too, with the second series of MPs exhibiting smaller losses at the point of air

Aeration parameters compared with CPs and GPs from the two series of MPs are presented in **Figures 13**–**18**. The ceramic plate has volume porosity in the range 45–50% and the fritted

These experiments show that increases in the air flow rate lead to increased dissolved oxygen transfer and Kla up to a certain value, after which Kla remains constant. Increasing the air flow rate induces the losses of air admission increases. In such as way the maximum aeration

Following this analysis, the MP gives the best results compared to the equivalent configurations in CP or GP. The pressure losses are 10 times less important compared with the CP and GP. Reducing the distance between the holes from 10 to 7 diameters, increasing the number of holes for the same active surface, and thus reducing the bubble size lead to an increase in the air-water contact surface and the retention time in water and thereby improves the transfer

**Figure 12.** Pressure drop on the aerator variation for different types of plates (MP series 1, MP series 2, ceramic plate CP

(sintered) glass plate has a porosity which is controlled in the range 0.25–0.315 μm.

and minimum energy consumption is optimized, reflected by SAE (**Figure 18**).

• concentration of DO at the measuring point, at saturation, corrected at temperature 20°C, at standard pressure 101 kPa, and relative humidity conditions are 100%

$$\mathcal{C}\_{\text{c20}} = \mathcal{C}\_{\text{s}} \left( \frac{1}{\text{r}\Omega} \right) \tag{8}$$

• standard oxygen transfer rate (*SOTR*)

$$\text{SOTR} = kla\_{20} \cdot \mathbb{C}\_{\varsigma 20} \cdot V \text{ [mg/min]} \tag{9}$$

• standard oxygen transfer efficiency (*SOTE*)

$$\text{SOTE} = \frac{\text{SOTR}}{W\_{\alpha}} \text{l-l} \tag{10}$$

• standard aeration efficiency (*SAE*)

$$SAE = \frac{SOTR}{P} \left[ \text{kg}\_{\text{CO}} / \text{kWh} \right] \tag{11}$$

The power consumed for the injection of air through the aerator

**4. Air flow influence on the aeration efficiency**

to the same temperature and pressure conditions (*t*

regime, (a) our experiments, (b) Krishna experiment's [9].

360 Desalination and Water Treatment

*Cs*<sup>20</sup> = *Cs*(

• standard oxygen transfer rate (*SOTR*)

• standard aeration efficiency (*SAE*)

• standard oxygen transfer efficiency (*SOTE*)

*SAE* = \_\_\_\_\_

*SOTE* = \_\_\_\_\_

tain the influence of hole size on the aeration efficiency.

As per the standard process the following steps was considered:

The experimental data were post-processed following standard procedures [9] and reduced

**Figure 11.** Volumetric mass transfer coefficient as a function of U, from the homogeneous to the heterogeneous flow

*Kla*<sup>20</sup> = *Kla* ⋅ *θ*<sup>|</sup>20−*t*<sup>|</sup> [1/min] (7)

• concentration of DO at the measuring point, at saturation, corrected at temperature 20°C,

*SOTR* = *kla*<sup>20</sup> ⋅ *Cs*<sup>20</sup> ⋅ *V* [mg/min] (9)

*SOTR*

*SOTR WO*<sup>2</sup>

\_\_\_1

• volumetric mass transfer coefficient, estimated by regression and corrected at 20°C

at standard pressure 101 kPa, and relative humidity conditions are 100%

*<sup>w</sup>* = 20°C and *patm* = 1 atm) in order to ascer-

*<sup>τ</sup>*Ω) (8)

[−] (10)

*<sup>P</sup>* [kgOD/kWh] (11)

$$P = Q \cdot (dp + \rho gH) / 1000 \quad \text{[kW]} \tag{12}$$

**Figure 12** shows the pressure losses measured for all MPs, CPs, and GPs. The injection losses on MPs are 10 times inferior to the CPs and GPs. The losses decrease with increasing diameter of the orifice but in a reduced report compared to CPs and GPs. The arrangement of the holes plays a role, too, with the second series of MPs exhibiting smaller losses at the point of air injection.

Aeration parameters compared with CPs and GPs from the two series of MPs are presented in **Figures 13**–**18**. The ceramic plate has volume porosity in the range 45–50% and the fritted (sintered) glass plate has a porosity which is controlled in the range 0.25–0.315 μm.

These experiments show that increases in the air flow rate lead to increased dissolved oxygen transfer and Kla up to a certain value, after which Kla remains constant. Increasing the air flow rate induces the losses of air admission increases. In such as way the maximum aeration and minimum energy consumption is optimized, reflected by SAE (**Figure 18**).

Following this analysis, the MP gives the best results compared to the equivalent configurations in CP or GP. The pressure losses are 10 times less important compared with the CP and GP. Reducing the distance between the holes from 10 to 7 diameters, increasing the number of holes for the same active surface, and thus reducing the bubble size lead to an increase in the air-water contact surface and the retention time in water and thereby improves the transfer of dissolved oxygen.

**Figure 12.** Pressure drop on the aerator variation for different types of plates (MP series 1, MP series 2, ceramic plate CP and glass plate GP).

**Figure 13.** Kla20 variation function of the air flow rate under standard conditions and comparison between the two MP series (right side).

**Figure 14.** Kla variation function of aerator pressure drop for MP and comparison with GP and CP.

**Figure 15.** Standard oxygen transfer rate (oxygenation capacity) variation function of injected air flow rate.

Experimental 2D particle image velocimetry (PIV) measurements, with uniform background lighting and laser-induced fluorescence (LIF) of the tracking particles, were performed in order to characterize the air-water biphasic flow and the 2D bubble column rising velocity in static water—see works by Murgan et al. [12]. For complete characterization of the flow, the

velocity field induced by the column of bubbles and the bubbles features are simultaneously determined using image processing technics. The bubbles features include: ascension veloc-

**Figure 18.** Standard aeration efficiency at different injection air flow rates. Comparison between the two series of MP

**Figure 16.** Standard oxygen transfer rate variation function of the power consumed for the air injection through the

Experimental Study of Standard Aeration Efficiency in a Bubble Column

http://dx.doi.org/10.5772/intechopen.76696

363

ity, diameter variation, interfacial area and shape factor.

**Figure 17.** Standard oxygen transfer rate variation function of the aerator pressure drop.

(right side).

aerator.

**Figure 16.** Standard oxygen transfer rate variation function of the power consumed for the air injection through the aerator.

**Figure 17.** Standard oxygen transfer rate variation function of the aerator pressure drop.

Experimental 2D particle image velocimetry (PIV) measurements, with uniform background lighting and laser-induced fluorescence (LIF) of the tracking particles, were performed in order to characterize the air-water biphasic flow and the 2D bubble column rising velocity in static water—see works by Murgan et al. [12]. For complete characterization of the flow, the

**Figure 15.** Standard oxygen transfer rate (oxygenation capacity) variation function of injected air flow rate.

**Figure 14.** Kla variation function of aerator pressure drop for MP and comparison with GP and CP.

**Figure 13.** Kla20 variation function of the air flow rate under standard conditions and comparison between the two MP

series (right side).

362 Desalination and Water Treatment

**Figure 18.** Standard aeration efficiency at different injection air flow rates. Comparison between the two series of MP (right side).

velocity field induced by the column of bubbles and the bubbles features are simultaneously determined using image processing technics. The bubbles features include: ascension velocity, diameter variation, interfacial area and shape factor.
