Giorgio Besagni

Politecnico di Milano,Department of Energy,Via Lambruschini 4a,20156 Milano,Italy

Keywords:Bubble column Scale-up Gas holdup Interfacial area LOPROX

ABSTRACT It is known that the performances of multi-phase reactors depend on the operating parameters(the temperature and the pressure of the system),the phase properties,and the design parameters (the aspect ratio (AR),the bubble column diameter,and the gas sparger design).Hence,the precise design and the correct operation of multi-phase reactors depends on the understanding and prediction of the fluid dynamics parameters.This paper contributes to the existing discussion on the effect of operating and design parameter on multi-phase reactors and,in particular,it considers an industrial process (e.g.,the LOPROX (low pressure oxidation) case study,which is typical example of two-phase bubble columns).Based on a previously-validated set of correlations,the influence of operating and design parameter on system performances is studied and critically analyzed.First,we studied the effects of the design parameter on the liquid-gas interfacial area,by keeping constant the fluid physical-chemical properties as well as the operating conditions;subsequently,we discussed for a fixed system design,the influence of the liquid phase properties and the operating pressure.In conclusion,this paper is intended to provide guidelines for the design and scale-up of multi-phase reactors.

1.Introduction

It is known that the performances of multi-phase reactors mainly depends on the operating conditions,on the design parameter and on the phase properties (i.e.,a change in the liquid phase properties during operation may led a change in the whole system performance) [1].Among the multi-phase reactors,two-phase bubble columns are widely used,owing to their advantages in operation as well as their simple design [2];a typical bubble column configuration consists in a vertical pipe,where the gas phase is dispersed into a liquid phase in the form of ‘‘non-coalescence-i nduced”bubbles or in the form of ‘‘coalescence-induced”bubbles.Unfortunately,despite the simple system design,complex fluid dynamics phenomena at the different scales—manifesting in the prevailing flow regimes—exist(see Ref.[3]).The different relationships between the fluid dynamics phenomena at the different scales and the many variables (i.e.,operating conditions,design parameter and phase properties) characterizing the system make it difficult to find general scheme of correlations for the precise estimation of the fluid dynamics parameters (i.e.,the gas holdup,the dispersed phase size,and the interfacial area) [4].In this respect,our recent paper contributes to the existing discussion and proposed a scheme of gas holdup correlation based on a formulation (i.e.,based on the ‘‘flow regime transitions”description of the system) able to take into account both the operating and the constructional variables [5].Taking into account the previous literature on the prediction of multiphase reactor performance[6,7] and the influence of the design criteria [8,9],this paper contributes to the existing discussion on the effect of operating and design parameter on multi-phase reactors.Among the different processes that employs bubble columns,WAO (wet air oxidation)has received particular interest in the field of wastewater treatment plant[10].In these systems,a reactor puts in contact the liquid phase (i.w.,organic effluent,caustic spent from refinery) with an oxidizer(i.e.,oxygen enriched air,or pure oxygen).This process can be catalyzed or not,and it generally operates at high pressure and high temperature.The LOPROX (viz.,‘low pressure oxidation’)technology is a WAO-based reactor,patented by Bayer Technology Services;it is characterized by relatively low operating pressure(approximately 3.5 MPa) and operating temperature (approximately 500 K).The LOPROX process is operated in bubble column reactors,owing to the large contact area between the gas and liquid phase properties (thus,providing high mass transfer coefficients between the phases),leading to a very good yield of product.Unfortunately,the design and the performance of a bubble column reactor mainly depend on the mass and heat transfer,hydrodynamics,fluid dynamics phenomena.For this reason a deep understand of those parameters are of fundamental importance.In particular,the study of scaling-up between the laboratory and the industrial scale is a matter of intense studies since the pioneering study of Wilkinsonet al.[11].In this paper,we study the influence of the operating and the design parameter on a LOPROX reactor performances,based on a previously-validated set of correlations;thus this paper is intended to provide guidelines for the design and scale-up of multi-phase reactors,as a continuation of ongoing research activities (see,for example,Refs.[8,12,13]).

The paper is structured as follows.First,the case study and the operating conditions of the LOPROX reactor are discussed;second,the methods and scheme of correlations employed are presented and discussed;third,the main outcomes of the analysis are presented and discussed.Finally,the main outcomes of this study are proposed.

2.The Case Study:The LOPROX Process

The LOPROX process,patented by Bayer AG,aims to improve the biodegradability of heavily contaminated production effluents(e.g.,containing organic substances,i.e.,phenol) before they are transferred to the traditional biological wastewater treatment facilities.In the LOPROX process,the contaminated wastewater(eventually,with the addition of a catalytic media)is pumped into a reactor(at approximately 500 K and 3 MPa)and oxygen is introduced through a gas sparger from the bottom.In the same time,a steam generator provides the required heat to the process.Once the process proceeds,the treated water is extracted from the top column and it is sent to the traditional biological wastewater treatment plant.Under the above-mentioned operating conditions the oxygen molecule react with the catalyst to split the long chain chemical contaminants into smaller components,which can be digested by bacteria in the plant.Despite this technology is not novel,a significant research activity has been carried on in the last decades;in particular,one of the greatest challenge is to make the LOPROX reactor works without the addition of strong acid,to reduce the overall cost of the whole material of the facility.

3.Methods

The proposed calculation procedure aims in computing the interfacial area,ai,of the bubble column depending on the operating conditions,on the design parameter and on the phase properties,starting from the system parameters of the LOPROX process.

3.1.Calculation procedure

To estimate the interfacial area,reliable correlation to estimate the gas holdup,εG,and the size of the dispersed phase,db,should be provided.The correlations applied in this study are described in the following of this section.The gas holdup correlation is a generalization of the well-known Wilkinsonet al.[11]as proposed by Besagni and Inzoli[5].This correlations,based on the‘‘flow regime transition”concept (This concept has the main assumption that systems having similar flow regime transition points exibits similar ‘‘reactor-scale”behaviour.),is able to take into account the many parameters characterizing the system:(a)the gas superficial velocity(UG);(c)the physical properties of the two phases;(c)the pressure and the temperature (pcandTc) (d) the bubble column design parameters (i.e.,bubble column aspect ratio,AR (AR=Ho/dc),inner diameter,dc,,and diameter of the gas sparger openings,do,see Fig.1).The scheme of gas holdup correlation reads as follows:

whereUcoalescenceinducedbubblesandUnon-coalescenceinducedbubblesare the rising velocity of the ‘‘coalescence induced”bubble and the‘‘non-coalescence induced”bubbles,respectively (please refer to Besagniet al.[14] for additional discussions concerning these concepts).In particular,the terms in Eq.(1) have been obtained by including in the original Wilkinsonet al.[11] correlation the mail design parameters,which reads as follows:

It should be noted that above-presented scheme of correlation(Eqs.(1)-(4)),employed to estimate the gas holdup,has been validated against a comprehensive experimental benchmark in our previous study[5];in addition,it should be noted that the this validated scheme of correlation holds on a previously validated scheme of correlation from Wilkinsonet al.[11].Also,it is worth noting that in Eq.(2),d0has a 0.01 exponential power of,which does has very low contribution rate.The non-dimensional groups in Eqs.(2) and (3) are defined as follows:

Dimensionless bubble column height (AR,aspect ratio).Defined as the ratio between the initial liquid height,H0,and the diameter of the bubble columns,dc(see Refs.[9,12]):

Dimensionless sparger openingDefined as the ratio between the bubble size produced by the bubble nucleation at the sparger,estimated by the well-known Gaddis and Vogelpohl correlation [15],as the diameter of the gas sparger opening:

Dimensionless diameterThe dimensionless diameter represents the quantification of the Rayleigh-Taylor instabilities at the ‘‘reactor-scale”(see Ref.[8]):

Fig.1. The bubble column design parameters.

Please note that the above scheme of correlation has been validated by Wilkinsonet al.[11] in its original implementation and in our previous paper as well.The bubble size correlation employed in the following is the correlation proposed by Wilkinsonet al.[16],based on dimensionless analysis and validated by experimental studies:

Hence,the interfacial area can be estimated,considering Eq.(1)and Eq.(8),through Eq.(9):

3.2.The cases studied

In this section,the cases studied are listed and commented.Starting from a baseline set of operating conditions(Table 1),first,the effects of the design parameter is studied,by keeping constant the fluid physical-chemical properties and operating conditions,then,once column diameter,column height and sparger nozzle diameter are established(based on the previous results),a discussion is made about the influence of presence of organic impurities(i.e.,a decrease the liquid surface tension.These cases have been considered as the concentration of the phenol in industrial waste-water (in the range of 1 %-10 %,mass fraction) was found to decrease the surface tension.) and of operating pressure on the liquid-gas interfacial area.It is worth noting that,in the following,for the sake of clarity,the influence ofUGis neglected and the results have been obtained at the flow regime transition point,computed using the correlation presented,discussed and validated in our previous study [5].

4.Results

The aim of this section is to describe the effects of column and gas distributor geometry on mass transfer parameters (in particular on gas-liquid interfacial area) in the LOPROX process for the wet air oxidation of organic pollutants in a bubble column reactor.It is worth mentioning that in the forthcoming paragraphs,the discussion refers to the scaling-up criteria discussed by Besagniet al.[8] (viz.,dc＞1 m,AR ＞5 anddc＞0.05 m).

4.1.The effect of bubble column aiameter and aspect ratio on gasliquid interfacial area (Case study#1)

The role of this section is to study the influence ofdcand AR on bubble column performances.These geometrical parameters are well-known scale-up parameters accordingly to Wilkisonet al.[11];the precise knowledge of their influence on bubble column performance is needed to understand how ‘‘laboratory-scale”experimental data can be representative of‘‘industrial-scale”reactors.Indeed,the bubble column diameter in practical reactors is very different from the one used in laboratory reactors,and may differ in one or two order of magnitudes.For this reason,in this analysis thedcis varied in the range of 0.1 m (‘‘laboratory-scale”)-3.0 m (‘‘industrial-scale”).A similar criterion is applied when considering the influence of AR,which is varied in the range of 1-20,in order to cover all kind of reactors.In this analysis the diameter gas sparger openings is selected,d0=0.05 m,as an intermediate value between the ‘‘laboratory-scale”and the ‘‘industrial-scale”.Based on the given operating conditions (temperature and pressure),and the physical-chemical properties (Table 1),the interfacial area displayed in Fig.2 have been obtained.The interfacial area increases withdc;conversely,the effect of AR is marginal.The increase inaiwithdcis reasonable because of the increase in εG.In addition,the increase rate is much higher atdc,while at largerdc,effect become negligible.According to Wilkinsonet al.[11],this is reasonable,because larger the diameter,the lower the wall effect on bubble becomes.Based on these results,it can be stated that the effect of AR onaiis lower compared to the effect ofdc.It seems thatdc＞1 and/or AR ＞5 leads to negligible effects on mass transfer and hydrodynamic parameters.The interested reader should also refer to the recent experimental study by Besagniet al.[8] concerning the influence of AR on bubble column fluid dynamics.

Fig.2. The effect of AR and dc on ai.

4.2.The effect of gas sparger openings aiameter and aspect ratio on interfacial area (Case study#2)

The role of this section is to study the influence ofdoand AR on interfacial area.In this case,dois varied in the range of 0.02 to 0.04 m,AR is varied in the range of 1-20 and whereasdc=1 m(equal to the baseline value,Table 1 as well as the results obtained in the previous Section).The results are presented in Fig.3.As previously observed,the influence of AR is marginal compared with the other geometrical parameters;conversely,ailargely increases when decreasingdo.Indeed,these conditions concerns the ones of a ‘‘fine gas sparger”and a mono-dispersed flow regime occurs(generally speaking,a small bubble size imposed by the gas sparger has positive effect on gas-liquid interfacial area,hence on mass transfer).The interested reader may refer to some studies from the literature for a comprehensive description of this flow regime[18-21].In particular,the reader may refer to our previous study for a detailed description of the flow regime instabilities that occurs when using low gas sparger openings (viz.the Ledinegg instability related to the hindrance effect of the uniform bubble side distribution).

4.3.The effect of gas sparger openings and bubble column diameter on interfacial area (Case study#3)

The role of this section is to study the influence ofdoand AR on interfacial area.In this case,dois varied in the range of 0.02 to0.04 m,dcis varied in the range of 1-2 m and AR=5(AR=5 accordingly to results obtained in the previous Section as well as the literature [8,11]).The results,displayed in Fig.4,are in agreement with Wilkinsonet al.[11];based on these results,it can be concluded that highdcand lowdoare needed to operate with highai.

Table 1 Description of the case studied-Baseline case presents fixed operating conditions which are changed in case studies #1-5,accordingly with the ranged defined

Fig.3. The effect of AR and do on ai.

It is worth noting that,in practical application,decreasingdois not feasible:firstly,impurities present in the liquid phase may cause fouling and obstruction of the nozzle,leading to violation of safety constrains,secondly;a high-pressure loss may occur when liquid phase passes through small orifices,due to a sudden expansion process at outlet.

Fig.4. The effect of dc and do on ai.

4.4.The effect of surface tension and pressure on interfacial area(Case study#4) and the effect of surface tension and pressure on interfacial area (Case study#5)

In the previous sections,the effect column design parameters,at fixed operating conditions and fluid properties,has been studied.In this section,the influence of the liquid phase properties and operating condition(viz.pressure and temperature)has been studied,for fixed geometrical design.Indeed,as a wet air oxidation technology,the two most important parameters for LOPROX are temperature and pressure.In particular,dc=1,do=0.05 and AR=5 are chosen accordingly to the previous results.When considering the range of surface tension and operating pressure to be investigated,it is worth noting that LOPROX reactors operates with wastewater;in particular,according to Lemoineet al.[6],it is likely that this process operates with liquid phases with variable surface tension,owing to the presence of phenol.The higher is the phenol concentration,the lower is the surface tension.In particular,in practical operating conditions,water surface tension is estimated to be approximately 0.03 N·m-1(see Table 1 value,‘‘baseline case”) and the presence of phenol up to 2%in mass would decrease the surface tension to approximately 0.01 N·m-1[22,23].The operating pressure of a LOPROX process,varies depending on the boundary and the load condition:according to Bayer,it may range between 0.3 and 3.5 MPa.The results of this analysis are presented in Fig.5(the effect of surface tension and pressure onai) and in Fig.6 (the effect of surface tension and temperature onai);ailargely increases with increasing the operating pressure.Indeed,increasing the pressure increases the gas holdup and decreases the bubble size:increasing the pressure increases the break-up rate and reduces the coalescence rate:the new equilibrium between coalescence/break-up leads to a decrease in the bubble size and delays the appearance of large bubbles.The cause is the propagation of Kevin-Helmholtz instability and internal gas circulation.Conversely,aidecreases with increasing the surface tension.In addition,it is worth noting that the increase in surface tension leads to decrease in foamability of the liquid phase (see the discussion concerning foaming discussed in Ref.[24]),hence the decrease of bubble stability,leading to coalescence.Practically,the increase of phenol content may raiseai,due to the decrease of surface tension.Both effects are in agreement with the previous literature [6,11].Regarding the effect of temperature(Fig.6),it is found that increasing the temperature slightly decreases the interfacial areas;however,such effect is marginal and might be neglected.

Fig.5. The effect of surface tension and pressure on ai.

Fig.6. The effect of surface tension and temperature on ai.

5.Conclusions

The effects of operating and design parameter on a commercial industrial process have been studied,based on a previouslyvalidated set of correlations.It was found that thatdc＞1 m,AR ＞5 anddc＞0.05 m,leads to negligible effects onaiand,thus,they be considered as possible scaling-up criteria in the present discussion (see,for example,Refs.[8,12,13]).Finally,ailargely increases with increasing the operating pressure;conversely,aidecreases with increasing the surface tension.In conclusion,the proposed paper propose a straightforward method to estimate the effect of operating and design parameter on bubble column performance.In addition,the present results and the describe procedure described can be interpreted as a simple method to scaleup‘‘laboratory-scale”bubble column,to estimate the effect of bubble column scale.It should be noted that,in the present approach,the correlation to estimate the gas holdup takes into account several parameters including design parameters.Conversely,the bubble size correlation does not concern the design parameters,which may causes some calculation bias for the design of multi-phase reactors.This limitation should be taken into account in future studies,and a comprehensive correlation for the bubble size should be proposed and validated as well.In addition,in the future,it is necessary to supplement the effects of the reaction appropriately including variation of concentration or temperature with the reaction going on.

Nomenclature

aiinterfacial area,mm-1

DHhydraulic diameter,m

non-dimensional diameter

dbbubble diameter,mm

dbsbubble size produced by the bubble nucleation at the sparger,mm

dcdiameter of the column,m

dogas sparger holes diameter,mm

gacceleration due to gravity,m·s-1

H0height of the free-surface before aeration,m

Usuperficial velocity,m·s-1

ε holdup

μ viscosity,N·s·m-2

ρ density,kg·m-3

Subscripts

b bubble parameter

c bubble column parameter

G gas phase

L liquid phase

o gas sparger parameter

trans flow regime transition point

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.