Development, Fabrication, and Testing of a Miniature Centrifugal Compressor Thesis Defense by Kevin Gregory Finney November 4, 2003.

Development, Fabrication, and Testing of a Miniature Centrifugal

Compressor

Thesis Defense by

Kevin Gregory Finney

November 4, 2003

Presentation Content•Project Specifications•What is a Centrifugal Compressor?•What makes this compressor unique?•Components and procedures for development

–Motor–Bearings–Coupler–Compressor Part Modeling–Compressor Assembly

•The testing environment / Acquisition of data•Results and conclusions from testing

Project Requirements• Reverse Turbo-Brayton cryogenic cooling system

for space applications• Goal was to develop a compressor capable of

system specifications– Designed inlet conditions

• Working fluid was air• Inlet Pressure = 1 atm (14.7 psig)• Temperature = 25o C• Total-to-static pressure ratio of 1.7 required• Mass flow rate of 4.5 grams per second

• Led to the design of a centrifugal compressor– Restrictions imposed by entire system

• Size of a cola-can / impeller diameter of 4.5 mm• Required the compressor to operate at 150,000 rpm

Definition of Centrifugal Compressor

• Dynamic compressor– Closed volume of fluid does not exist– Achieves pressure rise by a dynamic transfer of energy

to a continuously flowing fluid stream

• Fundamentals of a dynamic machine– Geometry of Flow Path– Direction of Energy– Fluid State

• Liquid = Pump• Gas = Fan, Blower, Compressor

Characteristics of a Centrifugal Compressor

• 1 to 2 = actual compression process

• 1 to 2s = isentropic compression

= 14.7 psig

Unique Design Characteristics

• 150,000 rpm operating speed

• Axial flow exiting the diffuser

• 4.5 millimeter diameter impeller

Requirements for the Development of Compressor

• Material Selection– Weight (inertial forces at high speeds)– Strength (withstand centrifugal stress, hoop stress, and torque)– Feasibility for manufacturing– Cost

• Motor Selection– Design required 300 Watts of power at 150,000 rpm– Motor for testing purposes built by Koford Engineering

• Bearing Selection– Ball bearings for testing purposes (ceramic balls and cage)– Air foil/journal bearings implemented in future design

• Minimal losses in bearing necessary for minimizing power required

• Coupler Selection– Alignment (angular, axial, lateral)– Maximum speed– Ability to modify for speed– Availability

• Rapid Prototype models– Visual aid during design and

machining stages

• Manufacturing methods– Parts capable of machining– Manufacturing methods available

• Balance of rotating components– Internal parts of motor balanced– Impeller balanced– Balanced at American Hofmann

150,000 rpm no-load requires 33 Volts and 6.8 amps

150,000 rpm with 300 Watts output requires 27.89 amps

Therefore, to obtain the desirable speed and power…

Voltage Current Electrical_Power 33V 27.89 A 920.37W

Electrical_Power Shaft_Power Heat 920.37W 300W 620.37W

Heat 620.37W

A method to remove this heat had to be implemented…

•Torque required at maximum speed was 0.02Nm•3-phase DC motor via controller with AC supply

–Voltage controlled speed–Current drawn determined by torque

Motor Characteristics

Design Expectations:

Parallel Flow

Motor

Dispersed Parallel Flow

Motor

Chosen due to simplicity in manufacturing the cooling jacket.

Cooling Jacket

• Parallel Flow Concept– Pressure loss occurred

• one inlet line and six outlet lines

• Eliminated possibility of leaks

– Turning of fluid occurred in plate

• Sealed with Silicone sealant

• Few bearings available capable of 150,000 rpm– DN Limit (occurred when rotating components

involved)• Inner shaft diameter (millimeters) multiplied by the speed

(rpm)10 mm X 150,000 rpm =1,500,000

• Type of lubricant determined by the DN Limit• Maximum DN Limit of 2 million for ‘Barden’ ceramic ball

bearing– Expected power loss per bearing important to the

amount of power required by motor• Advantages of ceramic bearings

– Ceramic Balls versus Steel Balls

Bearing Selection

Ball Bearing Air Foil Bearing

•Very expensive to incorporate

•Integrated part of entire design

•Only a few manufacturers:

Mohawk Innovative Technology, Inc. (MITI)

R&D Dynamics

•Ceramic balls and cage allowed for higher speed capability

•Same bearing throughout assembly

•Expected 50 Watts of loss in each bearing

•Handled large radial loads compared to axial loads

Schematic of Foil Bearing courtesy of R&D Dynamics

Maximum Speed

30,000 rpm

Lateral Misalignment

±0.13mm

±0.005in

Axial Misalignment

±0.8mm

±0.032in

Angular Misalignment

± 1o

Mass 20 grams

Current Coupler Selection

Coupler Modifications• Modification of

coupler for increased speed capability– Stainless Steel Sleeve

Retainer

Compressor Modeling and Drawings

• Modeling of Parts– Pro Engineer– Finite Element Analysis on blade shape to determine

loading effect caused by maximum pressure• Assembly of Parts in Design Stage of Development

– Interference between components / clearances• Drawings of Parts for Machinist

– Complicated curvatures• required the coordinates of the curve

– Tolerance of bearing bores critical to the radial stress placed on the balls

– Prototyped parts developed • assisted the machine shop with visual aid of complicated

geometry

Part Modeling•Complex geometry

–Undercut of impeller blades

–Complicated manufacturing methods

Max Stress = 1ksi Yield Stress = 75ksi

Concluded that blade will not fail from fluid loading.

Load does not include the centrifugal forces.

Hub

Shroud

Compressor Assembly• Compressor Bearing

Placement– Bearing Jig Fixtures

• Bearing in Diffuser• Bearing in Top Cap

• Specific order of assembly– Collector to Diffuser (bearing)– Collector to Housing– Impeller to Diffuser Bearing– IGV to Housing– Top Cap (bearing) to IGV and

Impeller

Exploded View of Compressor

Bearing Jig Fixtures

•Interference fit (force)

–Arbor Press

–Fixtures to hold part

–Bushings to press bearings

•Expansion fit (heat)

–Used if excessive force required and for disassembly

•Jig fixtures designed for disassembly

Impeller bottomed out in bore

Rapid Prototyping of Parts• Allowed a visual aid

during design– Correction of

assembly issues / interferences

• Supplied the machinist with an aid– Blade shape and

complex geometry more understandable

– Provided a visual to CNC code

Diffuser

Comparison of prototype parts to manufactured parts

Impeller (cast)

Inlet Guide Vane

Manufacturing of Parts• Impeller cast in Aluminum A356

– Properties of A356• Used for aircraft and missile components requiring high

strength, ductility, and corrosion resistance.• Used for intricate castings such as cylinder blocks, cylinder

heads, fan blades, and pneumatic tools• Contains 7% Silicon and traces of Magnesium and Iron. These

alloying elements assist in the strength and corrosion properties.

• Tensile Strength37ksi• Yield Strength27ksi

Properties taken from “Structure and Properties of Engineering Alloys” by Smith

Straight Blade Impeller

•Similar blade shape except there was no undercut

•Only required 4-axis CNC

•More homogeneous material

•More naturally balanced

Balancing• American Hofmann (Lynchburg, VA)

– Balanced to a g-level (ANSI Standard) equivalent of 150,000 rpm

– Material was removed in order to re-skew the axis of the hub to the axis of the shaft

• Two of the three ‘Curved Blade’ Impellers• ‘Straight Blade’ Impeller

Photo courtesy of American Hofmann

Alignment of Assemblies• Motor shaft alignment to the compressor shaft

– Axial– Lateral– Angular

• Run-out restricted by the radial play in the bearings

• Accurate to the accuracy of the measuring tools– 0.0005 inch accurate dial indicators

Adjusting the Alignment• Adjust assembly

alignment with shims– By using shims of 0.0005”

thickness, shafts were adjustable

• More accurate the alignment, more rigid the coupler could act– Resulted in higher

operating speeds with less power consumption

The Testing Environment

Reinforced Cage Controller Case

Controller

Fan

Mass Flow Meter

Pressure Transducer

Curve

Voltage versus Speed Curve

Components Required for Testing• Motor and Compressor Assemblies

– Assembly support brackets– Common base

• Instrumentation– Temperature

• Thermocouples

– Pressure• Calibration curve for pressure range

– Flow (mass flow measurement)– Operating Speed

• Digital reader, Oscilloscope, Frequency counter

– Input Power

• Data Acquisition

Pressure and Temperature at Diffuser

Exit

Motor Case Temperature

Mass Flow Controller

Power InBearing Temperature

Power Out of Motor

Motor Bearing Temperature

Pressure and Temperature at Inlet

Motor Bearing Temperature

Bearing Temperature

Areas of Desired Measurements

Pressure and Temperature after Mixer

• Motor Test– Determine ‘Free-spin’ motor data– Compare the speed measurements for accuracy– Develop Voltage versus speed curve

• ‘Blank Shaft’ Test– Determine the efficiency of the motor– Determine the loss per bearing

• Compressor Test– Determine the efficiency of the compressor– Determine the work of the impeller on the fluid

Order of Testing

Motor Test• ‘Free-spin’ operation

– Motor shaft spun only– Input variables and shaft

speed recorded

‘Blank Shaft’ Test

• Motor efficiency = 40% to 70%– 90,000 rpm = 65% with load

• Loss per bearing = 105 Watts at 90,000 rpm

Purpose of Blank Shaft

• Blank shaft (no hub nor blades) machined for use in determining the power loss in the bearings– Run motor without any attachments and record power

supplied to motor.– Assemble entire unit with blank shaft and operate at

150,000 rpm and record power supplied to motor.

Psupply_with_shaft Psupply_no_attachments Psupply_2_bearings

P supply_1_bearing

P supply_2_bearings

2

View of Assembly with Blank Shaft

Compressor Test• Curved Blade Impeller

– 89,485 rpm, 3.13 g/sec, 2.70 psig

• Straight Blade Impeller– 93,984 rpm, 5.14 g/sec, 5.05 psig

• Video of Compressor Test

Power versus Speed

0.0

200.0

400.0

600.0

800.0

1000.0

1200.0

1400.0

0 20000 40000 60000 80000 100000 120000

Speed (rpm)

Po

wer

(W

atts

)Cast Impeller

StraightBladeImpeller Test1

StraightBladeImpeller Test2

StraightBladeImpeller w/DataAcquisition

Gage Pressure Versus Speed

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0 20000 40000 60000 80000 100000 120000 140000 160000

Speed (rpm)

Gag

e P

ress

ure

(p

sig

)

Cast Impeller

Straight BladeImpeller Test 1

Straight BladeImpeller Test 2

Straight BladeImpeller w/ DataAcquisition

Design Point

Compressor Efficiency• Power consumption curves

• Actual output conditions:– 93,984 rpm– 1.29 pressure ratio– 61.2% isentropic efficiency– 5.1 grams per second

mass flow rate

Dimensional Analysis Plots

Speed Impeller Work Mass Flow Raterpm Watts g/s50000 60 3.849 0.57160000 90 3.159 0.37670000 150 3.389 0.35580000 200 3.952 0.38190000 225 5.139 0.612

Compressor Efficiency

Conclusion

• Straight Blade Impeller more effective than Curved Blade Impeller

• Compressor was on way to design conditions– Pressure ratio of 1.7– Mass flow rate of 4-8 grams per second– Operating speed of 150,000 rpm

• Reduce losses– Improve alignment

• Implement laser aligning procedures• Introduce rigid coupler• Incorporate one shaft throughout the assembly

– Incorporate air foil bearing / air journal bearing• Only if power consumption remains high

ReferencesBarden, “Precision Bulletin-The Effects of High Speed on Ball

Bearings” MMG 2.5 5/94.

DellaCorte, C. “Performance and Durability of High Temperature Foil Air Bearings for Oil Free Turbomachinery” NASA/TM-2000-209187/REV1. Glenn Research Center, 2000.

Koford, Stuart. “MK-Koford Brushless and Brush Motors.” Website. 2003. http://www.koford.com

Rimtec, “Motion Control” A Couple of New Ideas. Vic Jha. January/February, 2000.

Smith, William F. Structures and Properties of Engineering Alloys, Second Edition. New York: McGraw-Hill, Inc., 1993.

Questions and Comments