The Streamliner Artificial Heart

1
The Streamliner Artificial Heart

• Brad Paden
• University of California, Santa Barbara
• LaunchPoint Technologies LLC

2
Outline

• LVADs for artificial heart assist
• Background
• Next generation devices
• Design Prototypes
• Actuators
• Sensor
• Control
• Commercialization

3
Need for Mechanical Circulatory Assist

• ?15,000,000 heart disease deaths/yr.
• 5-10 could be saved with circulatory assist
• Several options
• Transplant (limited supply)
• Ventricular assist device
• Total artificial heart (not needed in general)

4
Heart Transplants in the US
2500/yr
2000/yr
1500/yr
1000/yr
500/yr
5
Left Ventricular Assist Devices (LVAD) are the
Leading Alternative to Transplants
6
(No Transcript)
7
Lumped-element model of the cardiovascular system
8
1st Generation LVADs are in use and are pulsatile
9
1st Generation Devices

• Increase 2-year survival from 8 to 23 in
  end-stage heart failure patients
• Issues remain
• Thrombus (clot) formation,
• Mechanical reliability.
• Energy efficiency.

Rose et al, Long-term use of a left ventricular
assist device for end-stage heart failure, The
New England Journal of Medicine, Vol 345(20),
2001
10
1st Generation (pulsatile)
2nd Generation (rotary)
3rd Generation (maglev)
11
(No Transcript)
12
Background on 3rd GenerationLVADs

• Extracorporeal Prototypes (Olsen and Bramm, 1981
  Allaire, Maslen, and Olsen, 1995 Chen et al,
  1998)
• Implantable devices in animal trials (StreamLiner
  1998, TCI/Sulzer 1999, Berlin Heart ?)
• Human Trials (Berlin Heart AG, June 16th 2002)

13
Utah/UVA Mag-Lev LVAD
14
Cleveland Clinic/Mohawk LVAD
15
LVAD Design Objectives

• Avoid mechanical shearing of the blood
• 6 Liters/min and 100 mmHg
• High reliability and efficiency
• hence magnetic bearings
• low power
• 10 g loading

16
Shear-Induced Hemolysisa design constraint
L.B. Leverett et al, Red Blood Cell Damage by
Shear Stress, Biophysical Journal, Vol. 12, pp.
257-273, 1972.
17
1st Streamliner Concept (HemoGlide 1)
18
Conical Bearing Prototypea wonderful 8x8,
10-state nonlinear multivariable control problem.
Stabilized using static linear decouplers and and
5 SISO lead-lag controllers.
19
This is too complicated! Can we just use
permanent magnets?
Earnshaws Theorem (1842)

• In a divergence-free electric field there are no
  stable
• equilibria for charged particles.
• Similarly for ideal permanent magnets in a
• static magnetic field.

20
more formally
21
Design Corollary We cant use all permanent
magnet levitation...
22
HG3 concept
But we can eliminate all but one active axis...
23
Final Design
JA Holmes
24
Section View and Final Device
25
Jarvik-7, Novacor LVAD, HG3b
26
HG3b Animal Trial (July 98)first fully maglev
pump sufficiently compact and energy efficient
for implantation
27
34 Day Animal Trial (August 24, 1999)
28
Design Approach Computer Modeling and
Optimization
29
Design Procedure
TOPOLOGY SELECTION
LUMPED PARAMETER MODELS
FINITE ELEMENT MODEL
OPTIMIZATION
RAPID PROTOTYPE
OPTIMIZATION
IMPLANTABLE PROTOTYPE
30
Topology Selection (via design grammar)
(FH,AO) Sp - PRB-DCBM-ATB-PRB-Sp

sb - ib
- sb
31
Lumped-Element Modeling and Finite-Element
Analysis

• Motor Thrust Actuator
• Lumped reluctance analysis w/FEA-derived
  Correction Factors
• Some FEA optimization
• PM Bearings
• closed form solution of maxwells equations
• FEAanalysis
• Rotor
• rigid body model
• linear fluid damping

• Controller, Actuator, Sensor
• finite-dimensional models
• Pump
• Meanline Analysis
• Empirical Formulae
• Computational fluid dynamics (CFD)

32
PM Bearing Design
33
z
34
z
35
(No Transcript)
36
PM Bearing Model
37
Motor Design
STATOR
ROTOR
38
Motor Parameterization
R5 13.31
R3 5.936
R4 9.58
Ls 14.66
W1 3.73
39
Motor Optimization
h 90 V 20.3
N 10000 RPM V 12.4
P 8W V 18.2
N 5000 RPM h 85 P 4W V 16.1
N 15000 RPM V 11.0
P 16W V 20.8
h 95 V 40.2
40
Motor redesign and re-optimization to reduce
radial instability
41
Pump Design (CFD)(James Antaki Greg Burgreen)
42
Final impeller design

• 5 impeller blade refinements
• 4 internal flow path refinements
• 6 aft stator blade refinements
• 18 month development effort

43
Flow visualization of early design
44
Hydrodynamic performance
Efficiency
0.16 0.15 0.14 0.13 0.12 0.11 0.10 0.09 0.08 0.07
0.06 0.05 0.04 0.03 0.02 0.01
8000 RPM
PRESSURE mm-Hg
FLOW RATE (LPM)
45
Control system design

• Linear actuator with optimized force/watt1/2
• Virtual Zero Power (VZP) axial control
• (1.5W coil power while pumping)
• Ultra low-noise eddy-current sensors
• Sensorless Motor Control

46
Sensor System
1 MHz Osc.
-1
(-90 deg)
Current Driver
s-a
sb
-90º
90º
LPF
Voltage Sense Amp
V(x)

offset adjust
mixer
x
L(x)
1/(2?(L(x0)C) ½) 1MHz
C
47
PID Controller Structure (for reference only)
impeller axial disturbance force
heat
Eddy-Current Sensor
Ka
(Ms2 -Kb )-1
Ks
force
displacement
Rotor Mass Bearing Negative Stiffness
Linear Motor 2 N/ root watt
Pos. Reference 0
noise 1Å / root Hz
-
KpKd s
coil current
PID Controller
Ki/s
48
Virtual Zero Power (VZP) Controller Structure
impeller axial disturbance force
less heat
Ks
Ka
(Ms2 -Kb )-1
displacement
force
Eddy-Current Sensor
Rotor Mass Bearing Negative Stiffness
Linear Motor 2 N/ root watt
noise 1Å / root Hz
current reference 0
KpKd s
-
coil current
s(KpKd s)
Ki/s
Ki Kp (1-Kd Ki)s
Anti-windup included
VZP Controller
J. Lyman, Virtually zero powered magnetic
suspension, US Pat. 3,860,300, 1975.
49
Axial Disturbance Force
4 Newtons
50
19 August 1999 Streamliner HG3C sn001 pre-implant
What is next?
51
Commercialization

• Teamed with MedQuest Products Inc, Salt Lake
  City, Utah.
• commercially competitive engineering, clinical,
  and business team.
• a large maglev patent portfolio and has acquired
  the Streamliner patents
• Moved to a centrifugal pump design to maximize
  efficiency

52
Optimized Clinical HeartQuest VAD
Height 34.7 mm Diameter 76.4 mm Weight 550
grams
53
Conclusions

• Control engineers have much to offer
• System optimization is at the center of the
  design process
• The language of mathematics, objectives and
  constraints is essential
• Clever control design makes low-power maglev
  possible (Lyman Patent 1975)
• Physiologic control is next...
• Responsive to condition of heart and body

54
Streamliner Team

• Mechatronics
• Brad Paden, Control engineering
• Chung-Ming Li, Analog Design
• Tom Dragnes, Electrical
• Dave Paden, Mechanical
• Randy Crowsen, Mechanical
• Lina Arbelia, bio-coatings
• Nelson Groom, mag-lev

• Fluids/Biological
• James Antaki, Streamliner Director
• Greg Burgreen, CFD
• Jon Wu, exp. fluids
• Marina Kameneva, blood damage
• Phil Litwak, veterinary surgery
• Bartley Griffith, surgery

• Funding
• McGowan Foundation

55
MedQuest Products Team

• Business
• Pratap Khanwilkar, CEO
• Tim Walker, Marketing

• Mechatronics
• Brad Paden, Garrick McNey,control engineering
• Jed Ludlow, dynamics
• Chung-Ming Li, electronics
• Dirk Cooley, electronics
• Dave Paden, Mechanical
• Randy Crowsen, Mechanical

• Fluids/Biological
• James Antaki, LVAD design
• Jon Wu, exp. fluids
• Gordon Jacobs, experimental
• Jim Long, surgery
• Don Olsen, veterinary surgery

• Funding
• NIH
• Venture Capital

56
The End