A Magnetorheological Hydraulic Actuator Driven by aPiezopump

Jin-Hyeong Yoo Jayant Sirohi Norman M. Wereley Inderjit Chopra

Alfred Gessow Rotorcraft Center,Department of Aerospace Engineering

University of Maryland, College Park, MD, 20742

ABSTRACT

Magnetorheological (MR) fluids can be used in a variety of smart semi-active systems. The MR damper shows anespecially great potential to mitigate environmentally induced vibration and shocks. Another aspect of MR fluidsis the construction of MR valve networks in conjunction with a hydraulic pump resulting in a fully active actuator.Conventional hydraulic pumps, however, are bulky, contain many moving parts, and do not scale favorably withdecreasing size motivating development of high energy density piezohydraulic pumps. These devices are simple, havefew moving parts and can be easily miniaturized to provide a compact, high energy density pressure source. Thepresent study describes a prototype MR-piezo hybrid actuator that combines the piezopump and MR valve actuatorconcepts, resulting in a self-contained hydraulic actuation device without active electro-mechanical valves. Durabilityand miniaturization of the hybrid device are major advantages due to its low part count and few moving parts. Anadditional advantage is the ability to use the MR valve network in the acuator to achieve controllable damping.The design, construction and testing of a prototype MR-piezo hybrid actuator is described. The performance andefficiency of the device is derived using ideal, biviscous and Bingham-plastic representations of MR fluid behavior,and is evaluated with experimental measurements. This will provide a design tool to develop an actuator for aspecific application. The prototype actuator achieved an output velocity of 5.34mm/sec against a mass load of5.15kg with a piezopump weighing 300gm.

Keywords: ER(Electrorheological) and MR(Magnetorheological) valve, Piezo electric driven pump. Hybrid actu-ator, ER/MR damper, Hydraulic actuator.

1. INTRODUCTION

Magnetorheological (MR) fluids can be used in a variety of smart semi-active systems, such as in optical polishing1

and fluid clutches,2 as well as in aerospace, automotive,3,4 and civil damping applications.5 In active systems, theER/MR fluid can be used as a fully active actuator in conjunction with a conventional hydraulic pump.6,7 In such asystem, MR fluid is used as the hydraulic fluid, and a network of MR valves functions as a directional control valve.However, the hydraulic pump assembly is complex and needs to be externally powered by a motor or some othersimilar power source. Many industrial and aerospace applications need highly reliable, precisely controllable andhigh energy density actuators. In order to address this need, there has been some interest recently in developing highenergy density piezohydraulic actuators. These hybrid devices are self-contained, electrically-driven linear actuators,consisting of a hydraulic pump driven by piezoelectric stacks, acting in conjunction with a conventional hydraulicoutput cylinder and a fast-acting set of valves. The pump-valve-cylinder system is used to rectify and convert the highfrequency, low amplitude motion of the piezostacks to lower frequency, higher displacement motion of the hydrauliccylinder. Piezohydraulic hybrid devices have been proposed for a variety of aerospace8 and automotive applications.9

Several prototype piezohydraulic actuators have been designed and tested over the past few years,8,10,9,11 clearlydemonstrating proof of the concept. Due to their mechanical simplicity, and operation off an electric power supply,piezopumps show great promise in applications requiring a miniature hydraulic power source.

Further author information: (Send correspondence to N.M.W.)J.H.Y.: Research Scientist, E-mail: jhyoo@eng.umd.edu, Tel: 301 405 1988J.S.: Research Associate, E-mail: sirohij@eng.umd.edu, Tel: 301 405 7286N.M.W.: Associate Professor, E-mail: wereley@eng.umd.edu, Tel: 301 405 1927I.C.: Alfred Gessow Professor and Director, E-mail: chopra@eng.umd.edu, Tel: 301 405 1122

The present study describes a novel actuator that combines the advantages of the piezopump and the MR valveconcept, resulting in a self-contained actuator with few moving parts. In this hybrid MR-piezo device, the mechanicaldirectional control valves typically present in conventional hydraulic actuation systems are replaced by a Wheatstonebridge of MR valves and the hydraulic pump/driving motor assembly is replaced by a piezopump. The MR fluidfunctions as a hydraulic fluid and a conventional hydraulic cylinder is coupled to the system to extract useful work.The hybrid MR-piezo actuator would have a low number of moving parts, a simple design, and can be scaled downfavorably. The moving parts inside of the system are the actuator shaft with two rod seals, two reed valves andthe piston in the piezo pump. An additional advantage of such a system is the high compliance of the device whileoperating as a semi-active MR damper. The actuator can adapt to large changes in disturbace mass loading andcan achieve reliable and precise control in severe environments.

The objective of this study is to develop a prototype MR-piezo position (or force) control actuator with acontrollable MR damping effect. Hence, this system behaves not only as a controllable MR damper, but also as afully activated actuator that can be used for position or force control. The MR valve and piezopump are the keycomponents of the actuation system. Driving force, stroke, cut-off frequency and efficiency are the main evaluationparameters for this actuator.12 The performance of a prototype MR-piezo actuator is determined experimentallyby suspending deadweights from the end of the actuator and measuring its output velocity. This test is performedat various pumping frequencies and applied magnetic field in the MR valve. The experimental results are comparedwith performance trends predicted using biviscous and Bingham-plastic MR fluid constitutive models as well as toan ideal mechanical valve case (infinite blocking pressure in the MR valve).

There are two main challenges in this system. Firstly, the piezoelectric pump has a low flow rate. Secondly, theMR valve has a low blocked force. However, through previous studies, we have already optimized the piezoelectricpump11 and MR valve performance13,14 with given volume. In this study, we will develop a prototype of MR-piezohybrid actuator, and evaluate its performance. Some experimental results will be presented, along with simulations.This will provide a design tool to develop an actuator for a specific application.

2. MR VALVE NETWORK CONCEPT AND CONSTRUCTION

The MR-piezo hybrid actuator is a combination of an MR valve network and a piezopump. A schematic of thedevice is shown in Fig. 1. The device consists of 4 MR valves arranged in a Wheatstone bridge configuration, anaccumulator, a piezopump and a conventional hydraulic cylinder. The accumulator is used to apply a bias pressure tothe hydraulic circuit. The device can function in two modes: an active mode and a semi-active mode. A descriptionof operation of the MR valve network as well as the MR valve is given below.

2.1. MR valve network operation

In the active mode, shown in Fig. 1(a), the piezopump functions as a pressure source to the system and the outputdisplacement of the device can be controlled by activating the MR-valve network. In the semi-active mode, shownin Fig. 1(b), the MR-valve network operates independently of the pump and the 4 valves can provide a more precisecontrol over a broad range as a semi-active damper.

In Fig. 1(a), the load attached to the output cylinder generates a force F . The piezopump forces fluid throughthe accumulator and MR valves configured as a Wheatstone bridge. Applying current to valves 1 and 4 activatesthese valves and the fluid flows predominantly from the output port of the pump at a pressure PS to the highpressure arm at a pressure PH , through valve 3. The flow into the lower chamber of the hydraulic actuator causesthe piston to move up and the fluid in the upper chamber flows through the low pressure arm, at a pressure PL,to the reservoir through valve 3. Under ideal conditions, or infinite blocking pressure, valves 1 and 4 permit noflow. However, in a real system, valves 1 and 4 permit a relatively low volume flux as compared to valves 2 and3. Assuming well balanced symmetric conditions in the Wheatstone bridge configuration, the flow rates in valves 2and 3 are defined as Qa and the flow rates in valves 1 and 4 have Qb, where Qa À Qb.

The performance of the hydraulic actuator with MR valves will be evaluated using three models: 1) an idealizedvalve in which infinite blocking pressure is assumed, 2) a Bingham-plastic model with finite blocking pressure, and3) a biviscous model, also with finite blocking pressure. With these assumptions, system efficiency can be derivedbased on knowledge of the field dependent yield stress of the MR fluid.

2

43

1

Accumulator

Piezopump

MR Valve network

Output cylinder PL

PH

PS

(a) Active operation

2

43

1

Accumulator

Piezopump

MR Valve network

Output cylinder

(b) Semi-active operation

Figure 1. Schematic of hybrid MR-piezo actuator

2.2. MR Valves

The MR valves used in this study consist of a core, flux return, and an annulus through which the MR fluid flows,as shown in Fig. 2(a). The core is wound with insulated wire. A current applied through the wire coiled aroundthe bobbin creates a magnetic field in the gap between the flange and the flux return. The magnetic field increasesthe yield stress of the MR fluid in this gap. This increase in yield stress alters the velocity profile of the fluid in thegap and raises the pressure difference required for a given flow rate. For Bingham-plastic flow, the typical velocityprofile is illustrated in Fig. 2(b). The primary parts of the MR valve design are pictured in Fig. 3.

33.2

16.4

2.3 10.8

R5.5

CW CCW R6.0 CW CCW

7.0

(a) Cross-section of MR valve

Region 1 (Post - yield)

Region 2 (Pre - yield)

Region 3 (Post - yield)

d

y -

-

-

δ

(b) Flow profile in the annulus

Figure 2. MR Valve Concept

BobbinHydraulic Cylinder/

Flux Return

End Cap

Figure 3. Parts of the MR valve

A 1-D axisymmetric analysis is given by Kamath et. al.,15 and an approximate rectangular duct analysis wasprovided by Wereley and Pang.16 Gavin17 provided an analysis for annular valves with more appropriate radialfield dependence. In our simplified analysis, we assume a uniform field across the valve gap.16 Following this latterstudy, we consider the approximate rectangular duct analysis of Poiseuille flow through a valve system containingMR fluid. For Newtonian flow, the volume flux Q through the annulus is a function of the area moment of inertiaI(= bd3/12) of the valve cross-section, the fluid viscosity, and the pressure drop over the valve length, P/La in thecase of the rectangular duct model. The dimensional volume flux through the valve can be determined.16,3

QN =bd3∆P

12µpoLa

QBP =bd3∆P

12µpoLa(1− δ̄)2(1 + δ̄/2)

QBV =bd3∆P

12µpoLa

[(1− δ̄)2(1 + δ̄/2) +

32µ̄

(1− δ̄2

3

)δ̄

](1)

where QN denotes for Newtonian flow, QBP denotes Bingham-plastic flow and QBV denotes biviscous flow. Here, thenon-dimensional plug thickness, δ̄ = δ/d and non-dimensional viscosity ratio, µ̄, which is defined as the ratio of thepost-yield differential viscosity (µpo) to the pre-yield differential viscosity (µpr), have been introduced. Normalizingeach volume flux by the Newtonian (field off) value of volume flux yields the non-dimensional volume flux for eachof the flow models

Q̄N = 1

Q̄BP = (1− δ̄)2(1 + δ̄/2) = QBP /QN

Q̄BV =[(1− δ̄)2(1 + δ̄/2) +

32µ̄

(1− δ̄2

3

)δ̄

]= QBV /QN

(2)

Figure 4 shows the trends of the non-dimensional volume flux as a function of the plug thickness δ̄, for Bingham-plastic and biviscous models, for the case of rectangular duct. In this figure, Q̄ = 1 implies Newtonian flow andQ̄ = 0 implies that the valve has blocked the flow.

0 0.2 0.4 0.6 0.8 1 0

0.2

0.4

0.6

0.8

1

Bingham-plastic model (µ = 0)

Plug thickness, δ [1]

µ = 0.2

µ = 0.4µ = 0.05

µ = 0.1

Biviscous model

Vol

ume

flux,

Q [1

]

Figure 4. Non-dimensional volume flux as a function of plug thickness

Note that the MR valve behavior based on a biviscous MR fluid constitutive model, is not capable of blocking theflow completely since Q̄BV for all 0 ≤ δ̄ ≤ 1. This implies that the two activated valves in the hydraulic circuit willexperience leakage, which is a key source of efficiency loss in the actuator system. This efficiency loss will occur eventhough the fluid will tend to predominantly flow through the inactive valves. Yield stress characteristics of a MRfluid change as a function of the applied magnetic field. Therefore, the magnetic field applied to the MR fluid is very

important to the performance of the valve and actuator. A high efficiency design was explored for meso-scale MRvalves. The magnetic circuit consisted of a bobbin, which a coil wound about its shaft. Surrounding the bobbin wasa tubular magnetic flux return. Key geometric properties were the bobbin shaft diameter, bobbin flange thickness,and gap between bobbin flange outer diameter and the flux return. A lower limit to miniaturizing the MR valveswas that the bobbin shaft saturates magnetically at lower field strengths as the shaft diameter decreases. Table 1summarizes the valve parameters.

Table 1. Valve dimensions

Outer diameter 16.4 mmBobbin diameter 11.0 mm

Flange length 11.6 mmAir gap 0.5 mm

No. of windings 114 turnsMaterial HIPERCO-50A

3. PIEZOPUMP CONCEPT AND CONSTRUCTION

The piezopump concept is illustrated in Fig. 5. An exploded view of the prototype pump used in the present studyis shown in Fig. 5(a). A description of the design and testing of this piezopump coupled to an output hydraulicactuator can be found in Ref. 18.

The main components of the piezopump are the piezostack assembly, piston assembly, pump body, pumpinghead and preload assembly (Fig. 5(b)). The piezostack assembly consists of two commercially available low voltagepiezostacks (model P-804.10, Physik Instrumente19), that are bonded together, end to end. The overall size of thisassembly is 36mm × 10mm × 10mm. One end of the piezostack assembly is bonded to a preload mechanism andthe other end is pushed up against a piston-diaphragm assembly. The preload assembly serves to adjust the positionof the piezostack assembly relative to the pump body as well as to provide a compressive preload to the piezostacks.The piston-diaphragm assembly consists of a steel piston of diameter 1 inch, which has a tight running fit with abore in the pump body, bonded to a 0.002” thick C-1095 spring steel diaphragm. The diaphragm seals the pumpbody from the hydraulic fluid in the pumping chamber, and the piston serves to constrain the deflected shape of thediaphragm to remain flat over most of its surface, thus maximizing the swept volume of the pump per cycle. Whileone face of the pumping chamber is formed by the movable piston, the other face is formed by the pumping head,that contains two oppositely oriented passive check valves. The piezostacks are actuated by a sinusoidal voltagefrom 0-100 V, resulting in an oscillatory flow of fluid in the pumping chamber. The check valves rectify this flowand provide a uni-directional output flow from the pump.

The piezopump has an outer diameter of 1.25”, a length of 3.5” and weighs 300gm. The major parameters ofthe complete device are given in Table. 2.

Table 2. Piezopump parametersPiezostack – Model P-804.10

Number of piezostacks 2Length 0.3937 inWidth 0.3937 inHeight 0.7087 in

Blocked Force (0-100 V) 1133 lbsFree displacement (0-100 V) u 0.5 mil

Maximum voltage 120 VMinimum Voltage -24 V

Capacitance 7 µFPumping Chamber

Diameter 1 inHeight 0.050 in

Connecting ports

Valve assembly

Pumping head

Spacer

Piston assembly

Pump body

Piezostack assembly

Preload base

(a) Exploded view ofpiezopump assembly

Preload

Piezostacks

Diaphragm

Piston

Check Valves

Hydraulic Fluid

(b) Schematic of the piezopump

Figure 5. Piezopump Concept

4. MR-PIEZO HYBRID ACTUATOR

The volume flux through each valve in Fig. 1(a) can be defined as

Qa =bd3∆P

12µLaQ̄a(PS − PH)

Qb =bd3∆P

12µLaQ̄b(PS − PL)

(3)

The total flow rate QS from the pump and the flow rate for moving the actuator QW are defined as

QS = Qa + Qb

QW = Qa −Qb = Apu(4)

The force equilibrium equation at the hydraulic actuator is

(PH − PL)Ap = F (5)

The force F includes friction force and output force of the cylinder. It follows that the steady-state forceequilibrium of equation (5) and the velocity of the actuator will be

u =bd3PS

12µLaAP

[(Q̄a − Q̄b)− (Q̄a + Q̄b)

F

ApPs

](6)

The maximum velocity of the actuator shaft and maximum force of the actuator can be expressed as

u|max =bd3(Q̄a − Q̄b)

24µLaApPS

F |max =Q̄a − Q̄b

Q̄a + Q̄bApPS

(7)

The maximum velocity and force are functions of the pressure source, PS . In the case of an ideal valve, thevelocity and force will increase as the source pressure increases. However, in the case of an MR valve, the non-dimensional volume flux, Q̄ is also a function of the supply pressure, so that the maximum velocity and force isdependent on the non-dimensional volume flux. In the piezopump, PS and the flow rates are functions of drivingfrequency and external loads.

From equation (6), the non-dimensional actuator performance equation can be stated

Q̄W =12µQW La

bd3PS=

12

[(Q̄a − Q̄b − (Q̄a + Q̄b)F̄

](8)

where, F̄ = F/ApPS . The maximum value of F̄ is 1 and Q̄W is 0.5. If a current is applied to the shaded valves 1and 4 in Fig. 1(a), we can define for the rectangular duct

Q̄a = 1 For Newtonian flow

Q̄b = (1− δ̄)2(1 + δ̄/2) For Bingham-plastic flow

Q̄b = (1− δ̄)2(1 + δ̄/2) +32

(1− δ̄2

3

)δ̄ For Biviscous flow

(9)

Corresponding to the MR fluid model used, the trends of Q̄b will follow the simulation results of Fig. 4. Fig. 6shows the actuator performance predicted by the Bingham-plastic model as a function of the non-dimensional plugthickness, δ̄. On increasing the current to the valve, the magnetic flux density at the gap will be increased. Thiscauses an increase in the plug thickness of the MR fluid flowing through the gap. The performance of the actuatorwill approach the ideal case, as the plug thickness increases. In the case of biviscous model, the performance asa function of the viscosity ratio is shown in Fig. 7. As can be seen in Fig. 7, the biviscous model cannot reachthe maximum performance of the actuator. The maximum performance with δ̄ = 1 is dictated by the value of theviscosity ratio, µ̄. On decreasing the current to the valve, the performance of the actuator also decreases, followingthe trends of Bingham-plastic model, in Fig. 6. The pressure from the piezo pump is a function of driving frequencyand external loads.

0 0.1 0.2 0.3 0.4 0.5 Working Flowrate, QW [1]

0

0.2

0.4

0.6

0.8

1

For

ce, F

[1]

δ = 0.2

δ = 1, ideal

δ = 0.8

δ = 0.6

δ = 0.4

Figure 6. Force and flow rate as a function of plug thickness, Bingham-plastic model

0 0.1 0.2 0.3 0.4 0.5 Working Flowrate, QW [1]

0

0.2

0.4

0.6

0.8

1

For

ce, F

[1]

µ = 1, idealµ = 0.2µ = 0.4

µ = 0.1µ = 0.05

Figure 7. Force and flow rate as a function of viscosity ratio, biviscous model, δ̄ = 1

4.1. System EfficiencyThe system efficiency, defined as the power transferred to the load divided by the power supplied to the MR valves,is given by

η =Power delivered to loadPower supply to system

=PaQa − PbQb

PaQa + PbQb=

Q̄a − Q̄b

Q̄a + Q̄b(10)

From the above, system efficiency of the actuator model can be derived as follow

η(δ̄, µ̄) = 1− 2Q̄b

1 + Q̄b(11)

Figure 8 shows the efficiency for the actuator when the working MR fluid behaves as a biviscous fluid. In thiscase, the maximum efficiency at δ̄ = 1 can be derived as

η(δ̄, µ̄)|δ̄ = 1− 2µ̄

1 + µ̄(12)

0 0.2 0.4 0.6 0.8 1 Non-dimensional plug thickness, δ [1]

0

0.2

0.4

0.6

0.8

1

Effi

cien

cy, η

[1]

µ = 0.4

µ = 0.2

µ = 0.1µ = 0.05

µ = 0, Bingham-plasti

c

Figure 8. Efficiency as a function of plug thickness, biviscous model

Thus the system efficiency is a function of both δ̄ of the valve and µ̄ of the fluid.

5. EXPERIMENTTo validate the nonlinear hybrid actuator performance, a set of four MR valves was implemented within a Wheatstonebridge hydraulic power circuit to drive a hydraulic actuator using a piezopump. The configuration of the hybridactuator is shown in Fig. 9. The actuator consists of three main part: a hydraulic cylinder, a set of four MR valveswith Wheatstone bridge configuration and a compact designed piezopump as a hydraulic source.

Piezopump

LVDT

MR valves

Hydraulic cylinder

Accumulator

Figure 9. MR-piezo hybrid actuator configuration

5.1. Experimental Setup

Experiments were performed to measure the output power of the actuator as a function of driving current to theMR valves and driving frequency to the piezopump. The hydraulic cylinder in this system as shown in Fig. 9, had7/16” bore diameter with 0.187” shaft diameter. The maximum stroke of the cylinder is 2-1/2”. To measure thedisplacement response, a potential meter(LVDT, TR50 Novotecknik) was attached with a rigid bar. The accumulatorregulated the source pressure from the piezopump. Figure 10 shows the schematic diagram for the experimentalsetup. The piezopump was driven by a high voltage amplifier and the MR valve was driven by a power supply,directly. Deadweights were hung off the end of the output hydraulic cylinder and the displacement was measuredusing the LVDT. The bias pressure was set to about 300psi. Figure 11 shows simulation results of yield stress of acommercially available MR fluid, namely MRF-132AD (Lord Corporation) with the MR valve design in this study.For the simulation the magnetic analysis was performed with ANSYS/Emag 2-D.

Actuator Assembly

Loads

LVDT

High Voltage

Amplifier

Power Supply

Oscilloscope

Figure 10. MR-piezo hybrid actuator test setup

0 1 2 3 Current, i [A]

0

10

20

30

40

50

Data Sheet Curve Fit

Yie

ld s

tres

s, τ

y [kP

a]

Figure 11. Yield stress as a function of applied current, MRF-132Ad(Lord Corporation)

5.2. Experimental Results

In Fig. 12, the experimental velocity response of the actuator shaft are plotted. On increasing the deadweight, thevelocity of the actuator tended to decrease. On increasing the applied current, the velocity of the shaft increased.

For the case of high loads condition has more enhancement of velocity response according to the applied current.It shows the possiblity of performance enhancement of this actuator. The performance of this actuator mainlydepends on the flow rate from the piezopump and blocking pressure of the MR valve. These parameters can beoptimized according to the application. The external force and velocity are the key design parameters for eachapplication. The actuator in this study can move 5.15 Kg with 5.34mm/sec velocity. The conditions of driving is140Vp−p at 250Hz for the piezopump and 1.8 Ampere for MR valves.

1 1.2 1.4 1.6 1.8 Applied Current, i [A]

0

4

8

12

16

1.73 Kg 2.88 Kg 4.00 Kg 5.15 Kg

Pzt pump: 250 Hz

Vel

ocity

, u [m

m/s

ec]

Figure 12. Output velocity as a function of applied current and external load

In Figs. 13(a), 13(b) and 13(c), the non-dimensional actuator performance test with MR valve is compared withthe test of the ideal valve. The driving frequency of the piezo pump is 200, 250 and 300 Hz, respectively. Thedriving current for the MR valve is 1.8 Ampere.

In the low loads region shows fluctuating data trend than the high loads condition. On increasing the drivingfrequency of the piezo pump, the fluctuation region on low loads site is diminishing. However, the scattering inhigh loads site is emerged for the ideal valve. In general, on increasing the driving frequency of piezopump and onincreasing the deadweight, the performance of the actuator is stabilizes.

Figure 13(d) shows the efficiency of the system as a function of pump frequencies and load. The efficiency at300 Hz driving frequency for the piezopump was highest among those frequencies tested. We are continuing testingof this actuator to better characterize its efficiency.

0 0.1 0.2 0.3 0.4 0.5 0

0.2

0.4

0.6

0.8

1

Ideal Valve Test with Ideal Valve Test with MR Valve

Working flow rate, Qw [1]

Pzt pump: 200 Hz, MR vavle 1.8 AF

orce

, F [1

]

(a) Actuator performance correlation, 200Hz pumping frequency, 1.8 A current

0 0.1 0.2 0.3 0.4 0.5 0

0.2

0.4

0.6

0.8

1 Ideal Valve Test with Ideal Valve Test with MR Valve

Working flow rate, Qw [1]

Pzt pump: 250 Hz, MR vavle 1.8 A

For

ce, F

[1]

(b) Actuator performance correlation,250 Hz pumping frequency, 1.8 A current

0 0.1 0.2 0.3 0.4 0.5

Working flow rate, Qw [1]

0

0.2

0.4

0.6

0.8

1 Pzt pump: 300 Hz, MR vavle 1.8 A

Ideal Valve Test with Ideal Valve Test with MR Valve

For

ce, F

[1]

(c) Actuator performance correlation, 300Hz pumping frequency, 1.8 A current

0 0.2 0.4 0.6 0.8 1

Force, F [1]

50

60

70

80

90 200 Hz 250 Hz 300 Hz

Effi

cien

cy, η

[%]

(d) System efficiency

Figure 13. Performance of the MR-piezo actuator

6. CONCLUSIONS AND FUTURE WORK

A MR-Piezo hybrid actuation system was analyzed and experimentally evaluated by testing a prototype. Theactuation system was constructed with four MR valves with Wheatstone bridge configuration and piezo pump. Thepiezopump forces the MR fluid through the MR valves and MR valves control the direction of flow through to thehydraulic cylinder. The controlled fluid flow makes the piston move. A non-dimensional volume flux was defined andthe performance of the actuation systme was evaluated with the non-dimensional equation. The system efficiencydefined as the power transferred to the load divided by the MR valve supply power. Through this study we canconclude that

1. With 250 Hz driving frequency for piezopump and 1.8 Ampere driving current for MR valve, the hybridactuator moves 5.15 Kg mass with 5.34 mm/sec velocity.

2. On increasing the driving frequency of piezopump, the performance of the actuator system stabilized and theefficiency increased.

The behavior of this MR-hybrid actuation system is determined by two separate non-linear systems: the MRvalve network and the piezopump. In the MR valves, the yield stress of the MR fluid has a non-linear dependenceon applied current. The plug thickness also depends on the pressure difference and flow rate. Furthermore, the flowrate generated by the piezopump is a complex function of the driving frequency and external force. Future work willinvolve the development of a non-linear model to predict the exact performance of this hybrid system. With thismodel, an optimized design can be generated depending on the application. Additionally, the damping effectivenessof the system in the semi-active mode will be evaluated.

REFERENCES1. W. I. Kordonski and D. Golini, “Fundamentals of magnetorheological fluid utilization in high precision finish-

ing,” Journal of Ineglligent Material Systems and Structures 10(9), pp. 83–689, 2000.2. U. Lee, D. Kim, N. Hur, and D. Jeon, “Design analysis and experimental evaluation of an mr fluid clutch,”

Journal of Ineglligent Material Systems and Structures 10, pp. 701–707, September 2000.3. J. Lindler and N. M. Wereley, “Analysis and testing of electrorheological bypass dampers,” Journal of Ineglligent

Material Systems and Structures 10(5), pp. 363–376, 2000.4. F. Gordaninejad and S. P. Kelso, “Fail-safe magneto-rheological fluid dampers for off-highway, high-payload

vehicles,” Journal of Ineglligent Material Systems and Structures 11, pp. 395–406, May 2000.5. S. J. Dyke, B. F. Spencer, M. K. Sain, and J. D. Carlson, “Experimental study of mr dampers for seismic

protection,” Smart Materials and Structures 7, pp. 693–703, October 1998.6. S. B. Choi, C. C. Cheong, J. M. Jung, and Y. T. Choi, “Position control of an er valve-cylinder system via

neural network controller,” Mechatronics 7, pp. 37–52, February 1997.7. Z. Lou, R. D. Ervin, and F. E. Filisko, “Behaviors of electrorheological valves and bridges,” in Proceedings of

the International Conference on ER fluid, pp. 389–423, 15-16th October 1991.8. L. D. Mauck and C. S. Lynch, “Piezoelectric hydraulic pump development,” Journal of Intelligent Material

Systems and Structures 11, pp. 758–764, October 2000.9. K. Konishi, H. Ukida, and K. Sawada, “Hydraulic pumps driven by multilayered piezoelectric elements -

mathematical model and application to brake device,” Proceedings of the 13th Korean Automatic ControlConference , 1998.

10. K. Nasser, D. J. Leo, and H. H. Cudney, “Compact piezohydraulic actuation system,” Proceedings of the 7thSPIE Conference on Smart Structures and Integrated Systems, Newport Beach, CA 3991, pp. 312–322, March2000.

11. J. Sirohi and I. Chopra, “Design and testing of a high pumping frequency piezoelectric-hydraulic hybrid ac-tuator,” Proceedings of the 9th SPIE Conference on Smart Structures and Integrated Systems, San Diego, CA4701, pp. 552–569, 17-21 March 2002.

12. J. Watton, Fluid Power Systems, Prentice Hall, 1989.13. J.-H. Yoo and N. M. Wereley, “Performance of a mr hydraulic power actuation system,” in Proceedings of the

9th SPIE Conference on Smart Structures and Integrated Systems, 4701, pp. 9–19, (San Diego, CA), 17-21March 2002.

14. J.-H. Yoo and N. M. Wereley, “Design of a high-efficiency magnetorheological valve,” in 8th InternationalConference on Electrorheological Fluid (ER) and Magnetorheological (MR) Suspensions, (Nice, France), 9-13July 2001.

15. G. M. Kamath, M. K. Hurt, and N. M. Wereley, “Analysis and testing of bingham plastic behavior in semi-activeelectrorheological fluid dampers,” Smart Materials and Structures 5, pp. 576–590, 1996.

16. N. M. Wereley and L. Pang, “Nondimensional analysis of semi-active electrorheological and magnetorheologicaldampers using an approximate parallel plate model,” Smart Materials and Structures 7, pp. 732–743, 1998.

17. H. P. Gavin, R. D. Hanson, and F. E. Filisko, “Electrorheological dampers, part i: analysis and design,” Journalof Applied Mechanics ASME 63, pp. 669–675, 1996.

18. J. Sirohi and I. Chopra, “Design and testing of a high pumping frequency piezoelectric-hydraulic hybrid actua-tor,” Proceedings of the 13th International Conference on Adaptive Structures Technology, Potsdam, Germany, October 2002.

19. Physik Instrumente (PI), Products for Micropositioning, US-edition, 1997.