Hydraulic Servo and Related Systems

ME4803 Motion Control

Wayne J. Book

HUSCO/Ramirez Chair in Fluid Power and Motion Control

G.W. Woodruff School of Mechanical Engineering

Georgia Institute of Technology

Hydraulics is Especially critical to the Mobile Equipment Industry

References1. Norvelle, F.D. Fluid Power Control Systems, Prentice

Hall, 2000.2. Fitch, E.C. and Hong I.T. Hydraulic Component

Design and Selection, BarDyne, Stillwater, OK, 2001.3. Cundiff, J.S. Fluid Power Circuits and Controls, CRC

Press, Boca Raton, FL, 2002.4. Merritt, H.E. Hydraulic Control Systems, John Wiley

and Sons, New York, 1967.5. Fluid Power Design Engineers Handbook, Parker

Hannifin Company (various editions).

The Strengths of Fluid Power(Hydraulic, to a lesser extent pneumatic)

• High force at moderate speed• High power density at point of action

– Fluid removes waste heat– Prime mover is removed from point of action– Conditioned power can be routed in flexible a fashion

• Potentially “Stiff” position control• Controllable either electrically or manually

– Resulting high bandwidth motion control at high forces

• NO SUBSTITUTE FOR MANY HEAVY APPLICATIONS

Difficulties with Fluid Power

• Possible leakage• Noise generated by pumps and transmitted by

lines• Energy loss due to fluid flows• Expensive in some applications• Susceptibility of working fluid to contamination• Lack of understanding of recently graduated

practicing engineers– Multidisciplinary– Cost of laboratories– Displaced in curriculum by more recent technologies

System Overview

• The system consists of a series of transformation of power variables• Power is either converted to another useful form or waste heat• Impedance is modified (unit force/unit flow)• Power is controlled• Function is achieved

Electric or IC prime mover

PumpTransmission line & valve

Motor or cylinder

Coupling mechanism

Load

Rpm-torque

Flow-press.

Flow-press.

Rpm-torque or force

Rpm-torque or force

Volts-amp

Simple open-loop open-center circuit

cylinder

4-way, 3 position valve

Pressure relief valve

Fixed displacement pump

filter

Fluid tank or reservoir

Actuating solenoid

Spring return

Simple open-loop closed-center circuit

Closed-loop (hydrostatic) system

Variable displacement

reversible pump

Drain or auxiliary line

Check valveMotor

Pilot operated valve

Proportional Valve

Basic Operation of the Servo Valve(single stage)

Torque motor moves spool left

Positive motor rotation

Torque motor moves spool right

Negative motor rotation

Flow enters

Flow exits

Orifice Model

xwA

C

pACQ

o

d

od

area flow orifice

coef. discharge flow orifice

2

4 Way Proportional Spool Valve Model• Spool assumptions

– No leakage, equal actuator areas

– Sharp edged, steady flow

– Opening area proportional to x

– Symmetrical

– Return pressure is zero

– Zero overlap

• Fluid assumptions

– Incompressible

– Mass density

21 qq

xppCq

CxppCq s

022

11 constant a ,

2021 ppppps

2;

2 :so

:pressure Load

21

21

ppp

ppp

ppp

ss

xpp

CxppCq ss 211

p0 p0ps

p1, q1p2, q2

x

Dynamic Equations (cont.)

s)order term(high )()(

linearize order tofirst toseriesTaylor ain Expand

1111

pp

p

qxx

x

qqq

ppxx

ppxx

xpp

C

p

qppC

x

q

sppxx

s

ppxx

22

;2

:sderivative partial Taking

11

0;2

0at which ,0;0

commonly point, operating Choose

21

11

1

Kp

qK

pC

x

q

qpx

ppxx

s

ppxx

p0 p0ps

p1, q1p2, q2

x

Dynamic Equations: the Actuator

If truly incompressible:

•Specification of flow without a response in pressure brings a causality problem

•For example, if the piston has mass, and flow can change instantaneously, infinite force is required for infinite acceleration

•Need to account for change of density and compliance of walls of cylinder and tubes

yAxKpKxKq 1211

Change in volume

Change in density

y

Net area Ap

q1

Compressibility of Fluids and Elasticity of Walls

2

:modulusBulk m

Np

dp

VMd

dp

d )/(111

For the pure definition, the volume is fixed.

dtqV

dpdtqdM

;

More useful here is an effective bulk modulus that includes expansion of the walls and compression of entrapped gasses

p

V

V

M

p

M

Vdp

VMd

eff2

11)/(11

Using this to solve for the change in pressure

dtqkdtMkdMV

dp eff

Choices for modeling the hydraulic actuator

With no compliance or compressibility we get actuator velocity v as

q

1/Adv/dt

With compliance and/or compressibility combined into a factor k

And with moving mass m

k dtq p

A /m

dv/dt

Manufacturer’s Data: BD15 Servovalve on HAL

Manufacturer’s Data: BD15 Servovalve on HAL

Two-stage Servo Valve

Torque motor rotates flapper, obstructs left nozzle

Flow gives negative rotation

Pressure increases Spool is driven right

With flapper centered the flow and pressure is balanced

Feedback spring balances torque motor force

Details of Force Feedback Design

Shown line to line; no overlap or underlap

2 Sharp edged orifices, symmetrical opening

Another valve design with direct feedback

Position Servo Block DiagramPosition measurement

Proportional control

May be negligible

Load torque

Net flow / displacement

Flow gain / motor displacement

Design of some components(with issues pertinent to this class)

• The conduit (tubing) is subject to requirements for – flow (pressure drop)

• 2 to 4 ft/sec for suction line bulk fluid velocity

• 7 to 20 ft/sec for pressure line bulk fluid velocity

– pressure (stress)

• The piston-cylinder is the most common actuator– Must withstand pressure

– Must not buckle

Design Equations for Fluid Power Systems

• Flow– Darcy’s formula– Orifice flow models

• Stress– Thin-walled tubes (t<0.1D)– Thick-walled tubes (t>0.1D)

• Guidelines– Fluid speed– Strengths– Factors of safety (light service: 2.5, general: 3.15,

heavy: 4-5 or more)

Darcy’s formula from Bernoulli’s Eq.

velocityfluid

areasection flow

rate flow

density mass fluid

perimeter)(section diameter)/section (flowx 4

diameter hydraulic

length tube

)Non (dependsfactor friction

tube thealong drop pressure

2

2

u

uDN

A

Q

D

L

f

p

A

Q

D

Lfp

hR

h

RD

hD

law) Poiseuille-(Hagen

viscosityabsolute

2000N ,128

or

R

4

p

L

DQ h

Friction factor for smooth pipes(empirical) from e.g. Fitch

Orifice Model

area flow orifice

coef. discharge flow orifice

2

o

d

od

A

C

pACQ

Buckling in the Piston Rod (Fitch)• Rod is constrained by cylinder at two points• Constrained by load at one point• Diameter must resist buckling• Theory of composite “swaged column” applies• Composite column fully extended is A-B-E shown below consisting of 2 segments

– A-B segment buckles as if loaded by force F on a column A-B-C– B-E segment buckles as if loaded by F on DBE– Require tangency at B

Cylinder construction (tie-rod design)

Resulting loading on cylinder walls

Applicable wall thickness stress formulas(conduits or cylinders)

• Thin walled cylinders (open, or where only circumferential hoop stress is significant) (Barlow)

• Thick walled cylinders– Brittle materials (based on max normal stress) use Lame’s

formula

– Ductile (based on max strain theory)• Open end (no axial stress) (Birnie)

• Closed end (cylinder bears axial stress) (Clavarino)

• Expansion of cylinder based on strain = stress/(Young’s modulus)

Stress formulas

ratio sPoisson'

ion)concentrat stress*safety offactor strength/(

stress allowableor design

diameter inside

2

open) (thin, formula sBarlow'

d

i

d

i

s

D

s

PDt

12

(brittle) formula s'Lame'

Ps

PsDt

d

di

1)1(

)21(

2

ductile) thick,(closed, formula sClavarino'

Ps

PsDt

d

di

1)1(

)1(

2

ductile) thick,(open, formula sBirnie'

Ps

PsDt

d

di

Results of Composite Column Model

r

br

bbb

aaa

b

a

D

ID

IE

FL

IE

FL

I

I

for solvethen

yiterativel solved bemay equation first The

rod ofdiameter 64

tantan

25.0

Composite column model matches manufacturer’s recommendations with factor of safety of 4

Equating the slope of the two column segments at B where they join yields:

Pressure Specifications

• Nominal pressure = expected operating

• Design pressure = Nominal

• Proof pressure (for test) = 2x Design

• Burst pressure (expect failure) = 4x Design

Pipes versus tubesTubes are preferred over pipes since fewer joints mean

•Lower resistance

•Less leakage

•Easier construction

Fittings between tube and other components require multiple seals

Flared tube design

New Approaches: Independent Metering

Independent Metering: Introduction

F

Kat

A

B

Ksa

Kbt

Ksb

Tank

Pump

Check Valve

x

Independent Metering Configuration

Advantages of Independent Metering: Metering Modes

• Energy saving potential: Regenerative flow.F

Kat

A

B

Ksa

Kbt

Ksb

Tank

Pump

Check Valve

x

Powered Extension Mode

F

Kat

A

B

Ksa

Kbt

Ksb

Tank

Pump

Check Valve

x

High Side Regeneration Extension

F

Kat

A

B

Ksa

Kbt

Ksb

Tank

Pump

Check Valve

x

Low Side Regeneration Extension

F

Kat

A

B

Ksa

Kbt

Ksb

Tank

Check Valve

Pump

x

Powered Retraction

F

Kat

A

B

Ksa

Kbt

Ksb

Tank

Pump

Check Valve

Low Side Regeneration Retraction

Regeneration flow can be defined as pumping the fluid from one chamber to the other to achieve motion control of the loadwith using no or minimum flow from the pump.

Power SavingsP

Q

R

P2

sP

1P

R

P2

1P

sP

Losses on Input Valve

Useful Power

Losses on Output Valve

Saved Power

Traditional Valve Independent Metering Valve Configuration

Saved Power

Regenerative Modes versus Powered Modes

• HSRE vs PEPs, Qs

a b x

High Side Regeneration Extension

a bx

Ps, Qs Pr, Qr

Powered Extension

AQ

x

APF

a

S

aS

max

max

AAQ

x

AAPF

ba

S

baS

max

max

• HSRE vs PE

Using Powered ExtensionWith High Pump Pressure

Using High Side Regeneration ExtensionSaves Pump Flow

Used Power

Lost Power

Saved Power

sQ sQ

sPsP

sP

sQQ

PP

Q

• LSRE vs PE

Using Powered ExtensionWith High Pump Pressure

Using Low Side Regeneration ExtensionSaves Pump Flow and Pressure

Used Power

Lost Power

Saved Power

sQ sQ

sPsP

sP

sQQ

PP

Q

Vibration Analysis• Effect of Mode Switching

0 10 20 30 40-0.1

-0.05

0

0.05

0.1

t

Z 1

Non-Dimensional Fundamental Modal Coordinate Z1

0 10 20 30 40-0.15

-0.1

-0.05

0

0.05

0.1

t

dZ1/d

T

Derivative of Non-Dimensional Fundamental Modal Coordinate dZ1/dT

0 10 20 30 40-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

t

Non-Dimensional Rigid Body Angular Speed

0 10 20 30 40-0.3

-0.2

-0.1

0

0.1

0.2

t

d/d

T

Non-Dimensional Rigid Body Angular Acceleration d/dT

KB1

= 2, KM1

= 0.5, KC = 0.1

1 = 0.6

Vibration Analysis• Telehandler Boom

0 1 2 3 4 5 6 7 80

200

400

600

800

1000Effect of Mode Switching on Boom Vibration

Time (sec)

Pos

ition

(m

m)

0 1 2 3 4 5 6 7 8

0

50

100

150

200

Time (sec)

Vel

ocity

(m

m/s

ec)

0 1 2 3 4 5 6 7 80

10

20

30

Time (sec)

Wor

kpor

t P

ress

ures

(M

Pa)

Head Chamber Pressure PA

Rod Chamber Pressure PA

Continuously Variable Modes (CVMs)

• Three-Valve Modulation Modes

• Use three valves to provide the fluid flow path instead of two valves

• Better force-speed capability and better velocity performance

Continuously Variable Modes (CVMs)

• PHSRE

Ps

Ps

Pr

Ksb

Kbt

Ksa

a

b

q1

qbq3

q2

Pump

Ksa Ksb

Cylinder

Ps

qb

qa

Pump

Ksa

Kbt

Tank

CylinderPs

Pr

qb

qa

Continuously Variable Modes (CVMs)

• PLSRE

Pump

Ksa

Kbt

Tank

CylinderPs

Pr

qb

qa

Kat

Kbt

Tank

Cylinder

Pr

qout

qin

Check Valve

Ps

Pr

Kat

Ksa

a

b

q1

q2

q3

KbtPr

Continuously Variable Modes (CVMs)

• PLSRR

Kat

Kbt

Tank

Cylinder

Pr

qout

qin

Pump

Kat

Ksb

Tank

Cylinder

Ps

Pr

qa

qb

Pr

Ps

Pr

Ksb

Kbt

Kat

a

b

q2

qb

q1

q3

Experimental Validation of CVM concept using PLSRE mode

Bucket

Boom

A typical Tractor Loader Backhoe (TLB)

Crowd

Experimental Validation

• Free Air Motion: Controller I: All in PE

0 1 2 3 4 5 6 70

0.1

0.2

0.3

0.4

0.5

Crowd Velocity with All Motion in PE Mode and in Free Air

Time (sec)

Cro

wd V

elo

city in (

Kph)

Velocity-kph

Energy Consumed:653.8 KJ

Experimental Validation• Free Air Motion: Controller IV: PLSRE CVM

Energy Consumed:190.23 KJ

0 1 2 3 4 5 6-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6Crowd Velocity with PLSRE CVM and in Free Air

Time (sec)

Cro

wd V

elo

city in (

Kph)

Velocity-kph

Experimental Validation• Free Air Motion: Summary

Controller Velocity Performance

Energy Consumed

All in PE Acceptable 653.8 KJAbrupt Transition Not- Acceptable 135.1 KJLinear Transition Not- Acceptable 129.1 KJ

PLSRE CVM Very Good 190.23 KJ