Elsevier

Chemical Engineering Science

Volume 111, 24 May 2014, Pages 153-169
Chemical Engineering Science

Bubble characteristics measured using a monofibre optical probe in a bubble column and freeboard region under high gas holdup conditions

https://doi.org/10.1016/j.ces.2014.02.024Get rights and content

Highlights

  • Local gas holdups, rise velocities and chord lengths measured using 1C optical probe.

  • High gas holdup conditions achieved with elevated pressures and surfactant addition.

  • Probe measurements were comparable to global values in water up to 9.0 MPa.

  • Significant bubble size reduction with surfactant diminished probe׳s effectiveness.

  • Freeboard and bubble column global hydrodynamics comparable at high gas holdups.

Abstract

Local bubble characteristics, including gas holdups, bubble rise velocities, and chord lengths, were measured using a monofibre optical probe manufactured to withstand elevated pressures. Previous studies have validated the use of single tip probes for simultaneous measurement of local bubble properties at atmospheric conditions; however no study has been currently reported for these probes at elevated pressures. Experiments were conducted in a 101.6 mm diameter column operating at pressures up to 9.0 MPa. Surfactant addition and operating pressure were studied to simulate high gas holdups observed in many industrial reactors containing liquid mixtures with surface-active compounds. Experiments were hence completed using two liquid phases: tap water and a 0.5 wt% aqueous ethanol solution. Liquid and gas superficial velocities were varied between 0–90 mm/s and 0–150 mm/s, respectively. Radial profiles at atmospheric conditions validated the probe measurements in water. Local holdups, rise velocities and chord lengths were adequately measured in water up to 9.0 MPa. The probe struggled in the aqueous ethanol solution due to its physical constraints (i.e., tip diameter and sensing length) when compared to the significant bubble size reduction (chord lengths below 0.5 mm). Comparisons with fluidized bed freeboard measurements demonstrated that flow through the bed enhanced bubble breakup for a coalescing system, but had a negligible impact with the added surfactant.

Introduction

Gas–liquid–solid flow is frequently encountered in chemical engineering processes. The fluid dynamic behavior of these systems must be studied to predict heat and mass transfer, flow behavior, and particle mixing. Bubble characteristics (e.g. bubble size distributions, bubble rise velocities, and local gas holdups) in industrial bubble columns and/or gas–liquid–solid fluidized beds are generally difficult to measure on-site due to their operating conditions. Vessels for such processes typically require materials that can withstand elevated temperatures and pressures, consequently limiting visual observations. The unit of interest for this study is the LC-FinerSM hydroprocessor which operates at pressures and temperatures of approximately 11.7 MPa and 440 °C, respectively (McKnight et al., 2003). The hydroprocessor׳s liquid recycle pan in the freeboard region was previously redesigned with the aid of CFD simulations, where the goal was to reduce the quantity of recycled gas (McKnight et al., 2003). The size of all bubbles for the simulation was assumed to be 1 mm based on a force balance. As computational times for CFD modeling are continually reduced and measurement techniques are improved, the objective of this study is to measure local bubble properties under high gas holdup conditions to improve future gas–liquid separation predictions and techniques.

Bubble characteristics in gas–liquid and gas–liquid–solid systems have been previously investigated using various measurement devices (Boyer et al., 2002). These techniques are commonly categorized as non-invasive or invasive, where the former do not interfere with the flow conditions inside the studied system. Non-invasive techniques can be used to measure some of the desired bubble parameters for this study. For example, global phase holdups in bubble columns operated at elevated pressures have been determined using differential pressure transducers via pressure profiles (Behkish et al., 2007, Jin et al., 2004, Rudkevitch and Macchi, 2008). Non-invasive techniques (e.g. dynamic gas disengagement, photography, radiography, NMR, particle image velocimetry, laser Doppler anemometry) have also been used to measure bubble size distributions or phase velocities (Chaouki et al., 1997). As discussed by Boyer et al. (2002) however, these techniques are limited by the operating conditions, low gas holdup requirements, and/or relative costs. Invasive techniques were thus examined to measure the desired bubble characteristics using the available experimental system under high gas holdup conditions.

Needle, heat transfer, and ultrasound probes are invasive measurement techniques which have been previously used to measure bubble properties in gas–liquid and/or gas–liquid–solid systems (Boyer et al., 2002). Ultrasounds probes measure the desired bubble properties based on the laws of ultrasound wave propagation, either through transmitted or reflected waves in a gas–liquid system. Unfortunately, the technique does not normally allow the simultaneous measurement of all desired bubble characteristics. Furthermore, the effectiveness is reduced at gas holdups above 10–20% due to repeated reflections/scattering of the signal (Broering et al., 1991, Macchi et al., 2001). Gas holdups in a bubble column operating at elevated pressures were previously shown to exceed these limitations (Rudkevitch and Macchi, 2008), consequently restricting the use of ultrasound probes.

Heat transfer probes have been used to determine local gas holdups based on the heat exchanged between an electronically heated probe and the surrounding liquid medium. As a bubble interacts with the probe, the quantity of heat exchanged is reduced causing a noticeable signal change (Abel and Resch, 1978). By examining the magnitude and slope of the signal, Utiger et al. (1999) found that local bubble holdups determined with the heat transfer probe were comparable to those obtained with an optical probe. The main advantage for the heat transfer probes is the measurement of the liquid phase velocity and root mean square (RMS) fluctuations. Current heat transfer probes do not measure all desired bubble characteristics for this study.

Needle probes are capable of simultaneously measuring local gas holdups, bubble chord lengths and rise velocities. Two types of needle probes have been previously used for measurements in bubble columns: optical fiber and impedance/conductive probes (Boyer et al., 2002). Optical and impedance probes operate based on the differences in the refractive index or conductivity, respectively, of the liquid and gas phases. Signal fluctuations due to phase changes at the probe tip allow the measurement of local gas holdups and bubble frequency. Dual tipped probes were developed to measure the bubble chord length and rise velocities, where the distance between both probe tips is typically in the range of 0.5–5 mm (Chabot et al., 1998, Chaumat et al., 2007, Magaud et al., 2001, Moujaes, 1990, Shiea et al., 2013). Four-point optical probes were also developed and validated to improve bubble velocity vector measurements compared to dual tip configurations (Xue et al., 2008, Xue et al., 2003). Previous studies have shown that bubble size distributions become narrower while increasing the pressure up to 15 MPa (Lin et al., 1998), where a notable fraction of bubbles diameters can be in the range of 1 mm and lower. The distance between multiple tips may consequently be an issue under high gas holdup conditions, where bubbles pierced by the front tip may be less likely pierced by subsequent tips. As shown in Table 1, few studies were found in the open literature where bubble characteristics were measured at elevated pressures and/or temperatures using an invasive device. In addition, these studies were conducted with no liquid flow which does not effectively simulate an ebullated bed reactor.

An innovative monofibre optical probe developed by Cartellier (1992) eliminated the requirement that a bubble must be pierced by two consecutive tips to measure the rise velocity and chord length. Previous experiments demonstrated that the bubble rise velocity was inversely proportional to the signal rise time between the liquid and gas voltage levels when the probe was normal to the gas–liquid interface (Cartellier and Barrau, 1998a, Cartellier and Barrau, 1998b). This relation is a function of the probe sensing length, a unique characteristic for each probe that must be determined prior to experiments. The monofibre optical probe׳s ability to simultaneously measure local gas holdups, bubble rise velocities and chord lengths has been validated at atmospheric conditions in a bubble column (Barrau et al., 1999, Cartellier, 1992) and in three-phase flow with particles of similar density to the liquid (Mena et al., 2008).

The main objective of this study is to investigate the use of a monofibre optical probe for bubble characterization in a bubble column and the freeboard region of an ebullated bed when operating under high gas holdup conditions (i.e., varying pressures up to 9.0 MPa, with and without the addition of a surfactant). In order to investigate the coalescing/heterogeneous and dispersed/homogeneous bubble flow regimes, data is gathered using water and a 0.5 wt% aqueous ethanol solution, respectively. Radial profiles at low pressures are compared to globally measured values to provide a partial validation of the device. Global and local gas holdups as well as photos are compared to local holdups while varying gas/liquid flow rates, increasing the pressure and adding a surfactant. The impact of the previous parameters on bubble chord length distributions and rise velocities distributions are also discussed. Lastly, global and local measurements in the freeboard region of an ebullated bed are compared to bubble column results at equivalent operating conditions.

Section snippets

Experimental setup

Experiments were carried out in a gas–liquid–solid fluidization system (Fig. 1) built by Zeton Inc. and capable of reaching pressures up to 10 MPa. The stainless steel fluidization column has an inner diameter of 101.6 mm and a maximum expanded bed height of 1.8 m. Glass viewing windows with dimensions of 118.8 mm×15.6 mm are located at heights of 244 mm, 603 mm, and 956 mm above the top of the distributor plate. At the top of the column, an expanded overflow section was designed as the primary

Measurement techniques

During the experiments, local bubble properties were measured using a monofibre optical probe, global phase holdups were determined using a differential pressure transducer and photos were taken at selected conditions. Once the system reached steady state, all measurements were taken to ensure proper comparison between global and local values.

Radial gas holdup profiles

The 1C optical probe performance was first investigated by measuring radial gas holdup profiles in a bubble column, where the results are presented in Fig. 4. Measurements were taken in both water and the 0.5 wt% aqueous ethanol solution while varying the gas and liquid superficial velocities to study the homogeneous and heterogeneous bubble flow regimes. The system was also operated at 1.0 MPa to study the initial radial profile changes when elevating the pressure.

Local holdups obtained with the

Ebullated bed results

The probe was placed at the center of the column (r/R=0) in the freeboard region above an ebullated bed containing 4 mm glass beads. Though bubble characteristics in the bed were of interest, the probe tip material could not withstand collisions with the studied glass beads. Mena et al. (2008) previously used a similar optical probe in three-phase flow using calcium alginate beads; however their studied particle density (1023 kg/m3) and size (dP=2.1 mm) minimized the likelihood of damaging the

Conclusions

A custom made monofibre optical probe was tested for the measurement of bubble characteristics in two phase flow under high gas holdup conditions, obtained via elevated pressures and/or the addition of a surfactant. Local gas holdup radial profiles were validated using global values obtained via the dynamic pressure drop. The average and maximum relative errors in water when integrating the profiles were 9% and 16%, respectively. Probe measurements also illustrated the modified profile

Nomenclature

    cB

    bubble chord length (m)

    dC

    column inner diameter (m)

    dP

    particle diameter (m)

    g

    gravitational acceleration (m/s2)

    hB

    bed height (m)

    LS

    sensing length (μm)

    m

    mass of the particles (kg)

    n

    refractive index (dimensionless)

    P

    pressure (Pa)

    ΔP

    dynamic pressure drop (Pa)

    r

    probe radial position (m)

    R

    column radius (m)

    T

    temperature (°C)

    tB

    bubble residence time (s)

    tR

    signal rise time (s)

    tT

    total measurement time (s)

    UG, UL

    gas and liquid superficial velocities (m/s)

    vB

    bubble rise velocity

    VG, VL

    gas and liquid voltage levels (V)

    Δz

Acknowledgments

The authors are grateful to Craig McKnight and Jason Wiens (Syncrude Canada Ltd.) as well as Stéphane Gluck and Nicolas Zuanon (A2 Photonic Sensors) for their valuable insights. The authors would like to acknowledge the Natural Sciences and Engineering Research Council of Canada, the Canadian Foundation for Innovation, the Ontario Innovation Trust and Syncrude Canada Ltd. for financial assistance.

References (39)

Cited by (60)

  • Measurement of droplet size and velocity based on a single tapered fiber optical reflectometer

    2022, Measurement: Journal of the International Measurement Confederation
  • Radial distribution of solid holdup and its fluctuation characteristics in a three-phase scrubbing-cooling chamber

    2022, Chemical Engineering Science
    Citation Excerpt :

    The post-processing of the voltage–time series values measured by the optical fiber probe can be used to obtain the radial distribution of εg and εs in region a under different operating conditions, which can be analyzed in combination with the pressure difference variation curve when bubbles escape measured by the DGD technique, as shown in Fig. 9(a)-(f), which can verify the inference in Section 3.2 that bubbles affect the particle distribution properties. The bubble movement under the “bubble scale” will affect the fluid flow pattern and the overall gas holdup under the “reactor scale”: The lift force will push the larger bubbles to the center of the bubble column when the bubble coalescence is promoted, and the increased bubble rise speed reduces the residence time, leading to the overall gas holdup reduced; the lift force will push smaller bubbles to the wall area when the bubble coalescence is suppressed, and the decreased bubble rise speed increases the residence time, leading to the overall gas holdup increased (Pjontek et al., 2014; Besagni and Inzoli, 2017). To connect the overall scale (εg1) with the bubble scale, the average bubble diameter Dave and the average bubble velocity vave were measured (h = 475 mm) using a dual-tip conductivity probe (The instrument description and operation method is detailed in the study of Peng et al. (2018a, 2018b)).

View all citing articles on Scopus
View full text