Dale Weber

1
Modeling and Control of Particle Incorporation
into Bacterial Cellulose in a Rotating Disk
Bioreactor

• Dale Weber
• M. Kuure-Kinsey, H.R. Bungay. J.L. Plawsky
• B.W. Bequette
• November 6, 2006

2
Agenda

• Introduction
• Bacterial Cellulose Production in a RDB
• Motivation
• Uses of BC and Benefits of Particle Incorporation
• Experimental tool for more practical reactors
• Incorporation Model
• Mathematical Model of RDB
• Results
• Conclusion
• Acknowledgments

3
Production of Cellulose Gel

• Agitation (e. g. RDB) decreases doubling time
• Static Culture 8-10 hrs
• Agitated 4-6 hours
• Cannon and Anderson (1991)

• In Philippines commonly produced in static
  cultures
• Known as Nata de Coco

4
Uses of Bacterial Cellulose

• Native Gel
• Dessert Nata de Coco
• Food Thickener and Additive
• Wound/ Burn Dressing
• 90 water retention
• Membranes
• Speaker Cone
• Biological Fuel Cell

• Benefits of Particles
• Medicated Wound Dressing
• Antibiotics, Proteins
• Mechanical Improvements
• Customize properties
• Conductivity, reactivity
• Insert Markers
• Counterfeit Prevention

5
Research Tool

• Geometry does not scale well
• Velocity bounds on gel formation
• Velocity bounds on particle uptake

Future Work Continuous sheet or trickle flow
reactor
6
Incorporation Model

• Infinite Fluid (Levich, 1962)

n - Kinematic viscosity (m/r) w- Rotation Rate D
- Diffusion Coefficient d0, d - Boundary layer
thicknesses j - mass flux to disk surface c0 -
particle concentration far from disk
7
Incorporation Model pt 2

• Laminar Flow over flat plate
• Liquid Phase
• Air Phase

kc - Mass Transfer Coefficient Z - Dimensionless
liquid film thickness Ca, Cas - Capillary
number h - Dimensionless surface tension R -
Dimensionless level c - Gravity Factor

• (Vijayraghvan and Gupta, Ind. Eng. Chem.
  Fundam.,1982)

8
SigmaCell 20 Uptake Model
Weighted Least Squares (w5 100)
9
CFD Simulation
R 3 cm disk
R 5 cm RDB
z-component of velocity field
10
Mathematical Model

• AAd - Available Disk Area
• TAd - Total Disk Area
• xpc, xpB - Particle Concentration in wet gel,
  RDB
• xgB - Glucose concentration in RDB
• Vl - Volume of wet gel
• e - Water hold up in gel ( vol)
• mm Zero order bacterial growth rate

11
EKF-based MPC
Augmented State Space Formulation
Predictor / Corrector
Prediction Horizon 15 Control Horizon
1 Sample Time 5
Disturbance Magnitude 10
12
Preliminary Control Simulations - Single Stage RDB
Extra DOF with continuous reactor allows better
control
13
Qualitative Gel Phase Results
14
Conclusions

• Develop quantitative gel analysis protocol
• Time-Varying Experiments
• Non steady state behaviors
• Collect Data from Continuous Rotating Disk
  Bioreactor
• Incorporation data with better knowledge of
  instantaneous concentrations
• Bacterial growth kinetic parameters

15
Acknowledgments

• BES-04
• AIChE

• Matt Kuure-Kinsey
• Lynn Bresnahan
• Dane Kuttron

Questions?
16
Control What?
17
Modeling
Results of Full and Reduced Models making an OL
step change
18
State Estimation

• Can perform discrete measurements
• Glucose Concentration
• Biomass Concentration

• Currently have online sensors for
• Temperature pH
• Liquid Phase Particle Concentration

• Use Optimal Kalman Observer to Infer
• Tank Volumes
• Gel Thickness
• Particle Incorporation

19
Particle Incorporation

• Previous Work Analyzed
• (SigmaCell 20)
• Von Karman Boundary Layer Analysis
• Adsorption is irreversible
• Surface coverage has no effect on adsorption
• Particles and Fluid outside the boundary layer
  are not moving
• Result Mass Transfer coefficient only a function
  of velocity and boundary layer thickness
• Adsorption Isotherm
• Purely empirical

Mormino, Doctoral Thesis, 2002
20
Current Particle Work

• Bulk Liquid

• Air
• Infinite Pulling Plate

Riley and Carbonell, 1993A and 1993B
21
Size Matters
22
Bacterial/Substrate kinetics
23
Control and Simulation
Optimal Control (regulator w/Min Fuel)
MPC (project) Kalman filter