Please download to get full document.

View again

of 17

Study of the effect of feed location on mixing behaviour in semi-batch reactors using PIV techniques

Study of the effect of feed location on mixing behaviour in semi-batch reactors using PIV techniques
13 views17 pages
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Documenttranscript
  Khan, Rielly & Muller: ISMIP 5 (2004)   STUDY OF THE EFFECT OF FEED LOCATION ON MIXING BEHAVIOUR IN SEMI-BATCH REACTORS USING PIV TECHNIQUE F.R. Khan 1  , C.D. Rielly 1*  , F.L. Muller  2 1  Dept. of Chemical Engineering, Loughborough University, Loughborough, LE11 3TU, UK 2  Syngenta plc, Global Specialist Technology, Huddersfield, HD2 1FF, UK ABSTRACT Semi-batch agitated reactors are commonly used in the process industries to manufacture various fine chemical, specialty chemical and bio-chemical products. For fast competitive reactions, imperfect mixing can lead to by-product formation and can reduce the selectivity of a desired product. It is well known that the flow behaviour and local mixing rates vary significantly throughout a mechanically agitated vessel, and the complexity of the process may be increased by the presence of large-scale, macro-instabilities which are not well represented in time-averaged simulations or measurements. The present study addresses these issues using particle image velocimetry (PIV) in a stirred semi-batch reactor, where one of the reactants is added continuously through a feed pipe. The majority of the flow field studies reported here have been carried out using 2-D PIV in the regions close to the blade of a Rushton disk turbine and at two feed locations and ( ) T r T  z 4.0,28.0  ==  ( ) T r T  z 25.0,57.0  == . Various local turbulence properties have been estimated including the turbulent kinetic energy, energy dissipation rate and vorticity. Results indicate that because of the physical presence of the pipe, there was a wake formed and this had an effect on the local hydrodynamics and turbulence just downstream of the feed, where micro-mixed reactions occur. Keywords: PIV, stirred vessels, semi-batch reactor, Rushton disc turbine, mixing, turbulence, macro-instabilities * Email: C.D.Rielly@lboro.ac.uk , Fax: +44 (1509) 223923 – 1 –    Khan, Rielly & Muller: ISMIP 5 (2004)   INTRODUCTION Flows in stirred vessels have been a point of focus for researchers, because of their extensive use for mixing and reaction in the process industries. The production of the majority of fine chemicals, specialty chemicals and bio-chemicals is carried out in multi-purpose plants, using stirred tank reactors (STRs) for different product ranges. In the case of fast, competitive reactions, imperfect mixing can lead to by-product formation and changes in the product distribution. These types of reaction are influenced by micro-, meso- and macro-concentration gradients in the reactor (Baldyga and Bourne, 1992). It is well known from the literature that these reactions do not occur uniformly throughout the tank contents, but are highly localised (Bourne and Hilber, 1990), often occurring close to feed pipes and thus the product distribution is affected by local hydrodynamic conditions. In previous work, Khan et al . (2003) have shown that the turbulence kinetic energy and its dissipation rate vary significantly throughout a stirred vessel. There are strong spatial gradients within the mean and turbulent velocity fields, and close to the impeller the flow exhibits significant periodic variations, due to the blade passage. Away from the impeller, there are weaker velocity gradients, but the presence of flow obstructions, such as a dip pipe, could have an important effect on local hydrodynamic conditions. Therefore, this study examines the effects of the presence of a dip pipe on the local flow conditions and on local turbulence properties in the regions close to a feed. In addition, the presence of large scale macro-instabilities (MI) in these flows imparts more complexity into the process. Low frequency, long time and length scales fluctuations away from the impeller region, appear to be responsible for particle pick-up and suspension close to the tank bottom; they are likely to influence dispersion and mixing of the fluid in the zones near feed pipes. These are unsteady effects and are not described well using mean velocities, turbulent kinetic energy or dissipation rates. The MI have time scales much longer than the blade passage interval (various frequencies have been reported by Montes et al (1997), Roussinova et al. (2003) and Nikiforaki et al . (2003)). Furthermore, information about the effect of the MI on reagent – 2 –    Khan, Rielly & Muller: ISMIP 5 (2004)   dispersion is missing from time-averaged measurements and from Reynolds Averaged Navier Stokes (RANS) CFD simulations of the flow in a STR. Particle image velocimetry, however, enables full field instantaneous experimental measurements at instants in time, from which the spatial structure of the MI may be determined. As a preliminary to work on the spatial structure of MI, PIV has been applied to examine the effects of a dip pipe on the local mean flow and turbulence and to deduce the effect on the selectivity of a micro-mixing influenced reaction.. EXPERIMENTAL SET-UP The experiments performed here used a two-dimensional PIV system with 1k ×  1k TSI PIV 10-30 CCD camera to capture image pairs and a 50 mJ Nd:Yag pulsed laser and cylindrical lens to provide illumination. A six-bladed Rushton disc turbine (RDT) with a diameter of 97 mm (  D  = T   /3) was investigated in a fully-baffled tank of 290 mm diameter; the impeller and tank geometries are shown in Figure 1(a). A rectangular Perspex tank surrounds the cylindrical mixing tank, to minimise refractive effects. TSI Insight software was used to interrogate the images with two frame FFT cross-correlation between 32 ×  32 pixels windows and 50% overlap between interrogation windows. Throughout these experiments water was used as the working fluid and the impeller speed was maintained at 300 rpm which corresponds to = 47,500 and a tip-speed of . Hollow spherical glass spheres of 10 µ m diameter were used for seeding and the seeding density was chosen to give about 10 particles in each 32 × 32 pixel interrogation window. µ  ρ   /  2  ND Re  = m/s51.1 = tip V  – 3 –    Khan, Rielly & Muller: ISMIP 5 (2004)         H      T    =  T   = 290 mm 20 1221  (a) (b) Figure 1: (a) Stirred tank with 6-blade RDT and (b) feed locations 1 and 2 To study the flow behaviour in semi-batch reactor, a syringe pump (syringe size = 50 cm 3 ) was used to transport fluid through the feed pipe. Seeded water was used as a fluid in the feed pipe, so as to keep the seeding density constant during the experiment. The outlet of the feed pipe was radially inward to face the tank centre. The feed locations where measurements were taken are shown in Figure 1(b). The feed pipe had a diameter of 1mm and the flow rate supplied was 1 ml / min, giving a linear velocity of 0.021 m/s and a pipe Reynolds number of 21.2. These feed locations and flow conditions were chosen to match a parallel study of meso- and micro-mixing being carried out by Professor Jaworski's research group at Szczecin University, Poland. The parallel studies aim to compare experimental measurements of the selectivity of a fast competing reaction scheme, with CFD simulations performed using RANS and LES models. Thus, here the feed flow rate is slow to ensure micro-mixed conditions. RESULTS AND DISCUSSIONS Angle-resolved velocity fields close to the Rushton disk turbine blades The instantaneous velocity fields generated by the RDT close to the blade were examined. Angle-resolved measurements were made at blade positions 3.33 °   apart for the impeller geometry, as shown in Figure 2; the 0 °  position corresponds to the blade being in the PIV measurement plane. The field of view was approximately 34 ×  34 mm and using the interrogation windows described above, this gave a set of velocity vectors, each of which was averaged over a length – 4 –    Khan, Rielly & Muller: ISMIP 5 (2004)   scale of 1 mm. Khan et al . (2003) noted that various workers had inferred that close to the blade, the integral length scale is about 0.3 to 0.5 of the blade width, which in this case, corresponds to   ; the Kolmogorov length scale, corresponding to the mean energy dissipation rate is about 40 µ m. Thus the velocity vector resolution is in the inertial sub-range, about 10 times less than the integral length scale, and about 20 times greater than the Kolmogorov length scale. mm102 /   ≈= W  L  Figure 2: PIV field of view , close to the RDT Angle-resolved mean velocities were obtained by averaging over 100 image pairs, taken at each blade angle between 0 and 60 o . Animations of the periodic flow structure have been obtained by combining the angle-resolved velocity fields at each of the blade positions into a sequence of frames. Examples of two such frames at 0 °  and 20 °  are shown in Figure 3 (the projected blade position is shown by the solid lines). The vector maps presented in fig. 3 show the typical flow field generated by a Rushton disk turbine. Liquid is discharged in a radial direction in the form of swirling jet. At all blade positions, it is clear that the velocity is a maximum on the central line of the impeller near the blade outer edge and is mainly radial. The maximum velocity near the impeller is 1.4-1.5 m/sec, which is close to the tip velocity ( V  tip  = 1.5 m/s). There are a few velocity vectors in fig. 3(a) with velocity magnitude greater than 1.5m/sec, which are in error, due to the blade obstructing the light sheet. The discharge from the central section of the blade moves towards the wall and the velocities decrease slowly as surrounding fluid is entrained. – 5 –  
Advertisement
MostRelated
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x