1 Biosystems Control Design Chapter 23 addresses a variety of synthesis problems in the field of biosystems: Pharmaceutical Operations Bioreactors Crystallizers.

1

Biosystems Control Design

Chapter 23 addresses a variety of synthesis problems in the field of biosystems:

• Pharmaceutical Operations• Bioreactors• Crystallizers• Granulation

• Drug Delivery• Type 1 Diabetes• Blood Pressure Control• Cancer Treatment• Controlled Treatment for HIV/AIDS• Cardiac Assist Devices

Ch

apte

r 23

Ch

apte

r 15

2

Ch

apte

r 23

Figure 23.1 Schematic of a typical industrial fermentor.

Ch

apte

r 15

3

Ch

apte

r 23

0 20 40 60 80 1004

4.5

5

5.5

6

6.5

7

Time (h)

Bio

ma

ss (

g/L

)

Dilution +10%Dilution -10%

Figure 23.2 Step response of fermentor model to symmetric changes in dilution of magnitude 10% from the nominal

value of D=0.202 h-1.

Ch

apte

r 15

4

Ch

apte

r 23

0 20 40 60 80 1005.5

5.6

5.7

5.8

5.9

6

6.1

6.2

6.3

6.4

6.5

Time (h)

Bio

ma

ss (

g/L

)

Dilution +10%Dilution -10%

Figure 23.3 Step response of fermentor model to symmetric changes in dilution of magnitude 10% from the

nominal value of D=0.0389 h-1.

Ch

apte

r 15

5

Ch

apte

r 23

Solvent Addition

CrystallizerCooling Jacket

PrimaryTemperature

Controller

ConcentrationController

SecondaryTemperature

Controller

Jacket Temperature

Jacket TemperatureSetpoint

Steam Flowrate

Temperature

Concentration

Figure 23.4 Flowsheet of a typical industrial crystallizer showing concentration and temperature controllers, including cascade control for temperature.

Ch

apte

r 15

6

Ch

apte

r 23

Hierarchy of Process Information

Plant operating variable

trajectoriesand

parameters

Intrinsicprocess

properties (MWD, CSD,

PSD)

Product performance

properties(optical,

flowability,dustiness)

Customer requirements

Measurable, trackable, properties

Controllable, adjustable parameters

Need relationships between levels to deliver product which consistently meets the customers needs

Ch

apte

r 15

7

Ch

apte

r 23

Specific Challenges in Agglomeration Control

Various process operations– granulation– polymerization– spray drying

Common characteristics– real-time analysis required for distribution– complex heat/mass/momentum transfer problems– multiple attributes of interest (size, shape,

concentration, etc.)– high dimension, stiff models– underactuated

Ch

apte

r 15

8

Ch

apte

r 23

Granulation Control – Challenges

Multivariable interactions (5th and 90th percentiles)

No target (setpoint) for 5th and 90th percentiles– One-sided limits

Data acquisition is difficult

Suggested Operating Objectives:– Track bulk density to reference– Minimize control effort if PSD within limits– Strong action if PSD out of limits

Ch

apte

r 15

9

Ch

apte

r 23

ProductGranulationDrum

Dryer

Crusher

Screens

Undersize Particles

Binder

Dry GranulesWet Granules

Fresh FeedHot Air In

Oversize Particles

Conveyer

RecycleGranules

Figure 23.5 Process flowsheet for granulation circuit with recycle.

Ch

apte

r 15

10

Ch

apte

r 23

Case Study [Pottman, Ogunnaike, Adetayo, and Ennis, Powder Tech., 1999]

GranulatorDryer Classifier

Bulk densitymeasurement

Spray nozzles

Oversized particles

Undersized particles

FeedProduct

Correlated Process Variables– particle size distribution (5th, 90th percentiles)

– bulk density

Manipulated Variables– 3 spray nozzles (flow rate) [bounded]

Ch

apte

r 15

11

Ch

apte

r 23 ProductGranulation

DrumDryer Screens

Undersize Particles

Binder Spray Nozzles

Fresh Feed

Crushed Oversize Particles

u1 u2 u3

y1

y2

y3

Figure 23.6 Simplified process flowsheet for granulator example. Here u1, u2, u3 are, respectively, nozzles 1,2, and 3, and y1, y2, y3 are, respectively, bulk density, d5 and d90.

Ch

apte

r 15

12

Ch

apte

r 23

0 50 100 150-4

-2

0

2

4

6

8

10

12

Ou

tpu

t

Time0 50 100 150

-40

-20

0

20

40

60

80

100

Inp

ut

Time

Figure 23.7 Closed-loop response of granulator to +10 step change in set point for y1 – left plot is outputs, right plot is

inputs (dashed line, y1 and u1; dotted line, y2 and u2; dash-dot

line, y3 and u3).

Ch

apte

r 15

13

Ch

apte

r 23

0 50 100 150-10

0

10

20

30

40

50

60

Ou

tpu

t

Time0 50 100 150

-100

-50

0

50

100

150

200

Inp

ut

Time

Figure 23.8 Closed-loop response of granulator to +50 step change in set point for y1 with constraints enforced on the

inputs. The left plot is outputs, right plot is inputs (dashed line, y1 and u1; dotted line, y2 and u2; dash-dot line, y3 and u3).

Ch

apte

r 15

14

Ch

apte

r 23

Diabetes Mellitus

World’s most common and costly disease

About one in every 400 to 600 children and adolescents has type 1 diabetes mellitus (T1DM)

National Diabetes Fact Sheet, 2005, Centers for Disease Control and Prevention

Complications of T1DM reduce life expectancy by ~15 years through micro- and macro-vascular disease

– Heart disease and stroke– Blindness– Kidney disease– Nervous system disease

Evidence that intensive insulin therapy (IIT) reduces complications

Diabetes Control and Complications Trial Research Group, 1993

Increased hypoglycemic events with IITDiabetes Control and Complications Trial Research Group, 1993

Ch

apte

r 15

15

Ch

apte

r 23

Model-Based Control Approach for Diabetes[Parker, Peppas, Doyle III, IEEE Trans Biomed. Eng. 1999]

Controller Patient

Model

UpdateFilter

Model-basedAlgorithm

Compartmental Model

Kalman Filter

DesiredGlucose Level GlucoseInsulin

-

-

Ch

apte

r 15

16

Ch

apte

r 23

InsulinController Insulin

Pump

Glucose Sensor

GlucoseSetpoint

Meal Disturbance

Gsp

Patient

Gm

G

BloodSugar

Figure 23.9 Block diagram for artificial -cell, illustrating the meal as the most common disturbance. G denotes the blood sugar of the patient, Gm is the output of the glucose sensor, and

Gsp is the glucose setpoint.

Ch

apte

r 15

17

Ch

apte

r 23

0 50 100 150 200 250 300 350 4000

100

200

300

400

500

600

700

Glu

cose

(m

g/d

L)

Time (min)

Figure 23.10 Open-loop response of patient’s blood glucose when the insulin pump is turned off.

Ch

apte

r 15

18

Ch

apte

r 23

0 50 100 150 200 250 300 350 40080

90

100

110

120

130

140

150

160

170

180

Glu

cose

(m

g/d

L)

Time (min)

Figure 23.11 Open-loop response of the patient’s blood glucose to a constant infusion rate of 15 mU/min from her insulin pump.

Ch

apte

r 15

19

Ch

apte

r 23

0 50 100 150 200 250 300 350 40040

60

80

100

120

140

160

180

Glu

cose

(m

g/d

L)

Time (min)

Figure 23.12 Open-loop response of patient’s blood glucose to a constant infusion rate of 25 mU/min from her insulin pump.

Ch

apte

r 15

20

Ch

apte

r 23

0 10 20 30 40 50-15

-10

-5

0

5

10

15

20

25

Time (min)

Mean Arterial PressureCardiac Output

Figure 23.13 Closed-loop response of patient’s mean arterial blood pressure and cardiac output to a -10 mmHg change in the MAP set point.