Document Sample
2004-01-1322v001 Powered By Docstoc


Review and Development of Electromechanical Actuators for Improved Transmission Control and Efficiency
A. J. Turner and K. Ramsay
Ricardo Driveline and Transmission Systems

Reprinted From: Transmission & Driveline Symposium 2004 (SP-1817)

2004 SAE World Congress Detroit, Michigan March 8-11, 2004
400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760 Web: www.sae.org

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. For permission and licensing requests contact: SAE Permissions 400 Commonwealth Drive Warrendale, PA 15096-0001-USA Email: permissions@sae.org Fax: 724-772-4891 Tel: 724-772-4028

For multiple print copies contact: SAE Customer Service Tel: 877-606-7323 (inside USA and Canada) Tel: 724-776-4970 (outside USA) Fax: 724-776-1615 Email: CustomerService@sae.org ISBN 0-7680-1319-4 Copyright © 2004 SAE International Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in SAE Transactions. Persons wishing to submit papers to be considered for presentation or publication by SAE should send the manuscript or a 300 word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE. Printed in USA


Review and Development of Electromechanical Actuators for Improved Transmission Control and Efficiency
A. J. Turner and K. Ramsay
Ricardo Driveline and Transmission Systems
Copyright © 2004 SAE International

With the drive toward more fuel efficient vehicles, the individual and combined efficiency of powertrain subsystems is increasingly important. Development of alternative transmission types (AMT, DCT, CVT) has brought significant benefits arising from an increase in the number of ratios, ratio span and unit efficiency. However, it is recognised that further potential lies in improving actuation technology in controlled transmissions, of all types. Since the economic benefit of further refinement of traditional hydraulic devices appears to be limited, it is suggested that alternative technology is required. This alternative would seek to further reduce parasitic energy consumption due to pumping and associated system losses and also improve poor duty cycle controllability, whilst enabling cooling of any slipping sub-elements. The reported research has focused upon the development of electromechanical systems that offer efficient intermittent actuation and also system simplification over hydraulic, and more recently electrohydraulic, systems. These improvements include overall energy conversion efficiency and idle power consumption, which support the trend towards the ‘more electric vehicle’ and 42 Volt architecture. However, there are a number of significant issues that need to be addressed in order that these devices can operate successfully within the given environment. This paper presents a review of current electromechanical actuation technology, a design example of an electromechanical actuator and identifies system simplification and integration with current technologies.

operate in the optimum range for best fuel economy. Currently however, hydraulic losses result in a significant parasitic power consumption that detracts from the improvement made by ratio selection. The recent developments in automated manual transmissions have combined the control of automatics with the high mechanical efficiencies of manual transmissions. However, it is recognised that further improvements can be made in the actuation technology used in these transmissions to further reduce losses. As hydraulic technology has been identified as a significant proportion of the loss in certain transmissions, this paper presents an alternative electromechanical actuation scheme, presenting a comprehensive review of current technology, stating merits and outlining their suitability for actuation tasks and including methods for improving performance. The paper also covers the consequences of 12 and 42 Volt systems on actuator design, performance and weight, and issues associated with operational environment, particularly temperature. A design evaluation of a motor driven actuator for clutch control is then presented. This includes system specification, technology selection, and rating of a suitable machine with regard to the mechanical system being driven, along with data for motor sizing and efficiency. Dynamic performance is evaluated with actuator models developed using MATLAB/SIMULINK and EASY5. The dynamic models were constructed with a view to establishing response times and to evaluate how a motor driven actuator effects clutch application.

Three types of transmission actuation technology are discussed briefly here; hydraulic, electrohydraulic, and electromechanical systems. Transmission data taken from tests undertaken by Ricardo shows hydraulic systems may represent 50% of the total transmission loss, but have the advantage in terms of actuator size and weight, with the hydraulic pressure source placed elsewhere in the transmission where space is at less of a premium. Electrohydraulic actuation (i.e. with an electric power source) attempts to provide a compromise between

The automated control of vehicle transmissions has traditionally been accomplished with the use of hydraulic actuation systems that offer high output forces from small actuators supplied with pressurised oil from a remote pump. Along with the convenience of automatic gear changes, automated transmissions may be controlled to select the appropriate gear ratio for the vehicle driving conditions, which allows the engine to

electromechanical systems and hydraulic technology. For example, the replacement of the mechanical drive to a hydraulic pump with an electric motor offers an improvement in duty cycle control, allowing better tracking of open circuit hydraulic pressure demand. However, losses associated with leakage and flow through components is still present. An alternative to this is a closed circuit hydraulic system [1] where an electric motor is able to vary hydraulic pressure by changing its speed and direction. Electromechanical devices can be highly efficient, >95% in some cases, and also offer excellent duty cycle control. However, their performance at high temperatures and in terms of force/torque density is generally poor. Because of this, electromechanical devices may need to be located remotely from the point of use, and unless some mechanical system is used to trade actuator speed, which is generally high in electrical devices, for force/torque capability, actuator dimensions may be unacceptable for a given specification.

move away from brushed DC technology in some areas because of efficiency and torque density issues. In general, the torque produced by an electrical machine is proportional to the rotor volume, the electric loading and the magnetic loading in the machine. The electric loading is limited by the thermal properties of the motor, with force cooled machines being able to stand higher electric loadings. Magnetic loading is governed by the magnetic circuit design and type of materials used in the machine construction. Brushed DC motors This machine is constructed using either a stator winding or magnets and a rotor winding that is supplied with current via a commutator and brush gear. For low power applications, the stator winding is usually replaced with ferrite magnets. The brushed DC motor is cheap, but compared to other designs, has many drawbacks. Although brushed motors account for the majority of electric motors in automotive applications, the commutator and brushes are the limiting feature of the motor as they limit both maximum armature winding current (and hence torque) and the maximum motor speed. Therefore, for high power, fast response applications, brushed motors may be too large and slow, and alternative technologies need to be employed. Brushless motors

Electromechanical technology can be separated in to two main classes; rotary devices such as conventional electric motors, and linear devices such as solenoids and linear motors. ROTARY ELECTRICAL MACHINES Rotary electrical machines can be further subdivided into two groups, limited angle actuators and motors. Motors can be precisely controlled in terms of torque, acceleration, speed and position, and there are a number of well established electric motor technologies each having different characteristics. Limited angle actuators may include devices such as rotary solenoids. Torque production in rotary devices is accomplished by two means; the alignment of magnetic fields (excitation torque) and the alignment of ferromagnetic materials (reluctance torque), and all machines use one or both of these mechanisms. This paper deals with three different motor topologies, brushed DC, brushless permanent magnet and switched reluctance (SR) motors. The induction motor is not considered here as traditionally this motor is used in high power applications where weight is less of an issue or the cost of using permanent magnet machines would be too great because of motor size. In automotive applications where power requirements are relatively low, there is little cost benefit in using induction motors, and their lower efficiency and torque density results in larger motors for a given torque requirement. Brushed DC motors have dominated motor technology used in the automotive environment, and this is largely due to cost. However, as motor abundance and power requirement has increased, it is becoming necessary to

The brushless motor class can be split into a number of different categories but only three main designs are considered here. These are the brushless permanent magnet AC and DC machine, and the Switched Reluctance (SR) motor. Permanent magnet brushless DC and AC machines are almost identical in construction but the main distinguishing feature between the designs is the back-EMF waveforms, with brushless AC (BLAC) machines having a sinusoidal back-EMF waveform and brushless DC (BLDC) having trapezoidal back-EMF waveform. Although the power electronic drives used to control these two machines are similar in layout, their control is somewhat different and each requires rotor position information of differing degrees. As the AC machine requires sinusoidal currents, accurate rotor position information is necessary and is usually measured using an encoder or resolver. The DC machine, which requires square wave current waveforms, needs to experience a change in coil currents every 60° (electrical) or rotation, therefore coarse rotor position information can be tolerated which may be achieved using Hall effect sensors. However, if the machines are to be used in position servo applications, then a position sensor such as that used in the BLAC case must also be used for rotor position feedback in the DC machine case, giving no significant advantage in using a DC machine over an AC

machine. If the application were speed regulated, then the DC machine would fulfill the requirements without the accurate position feedback needed by AC machines. There are also a number of subtle differences between the operation of the machines leading to slight differences in torque density, and hence torque/rotor inertia ratio, speed range and control drive electronics VA rating [2]. Brushless DC motors may also be supplied with unidirectional currents which has a number of advantages and disadvantages [3]. As cost reduction is of great importance, and as the silicon cost of a drive system can be a significant portion of the total material and production cost, the use of unipolar current drives to reduce the number of silicon devices needed is of interest. There are however, a number of significant drawbacks to this control scheme. The windings are poorly utilised as they may only be excited for half the maximum available interval for torque production. Some unipolar drive configurations do not allow four-quadrant operation (motoring and generating in forward and reverse directions). Inductive energy is dissipated in the windings of the motor and not returned to the supply as with conventional bipolar drives and they can exhibit greater levels of torque ripple than bipolar driven machines. However, bipolar drive problems include the risk of ‘shoot-through’ faults where both switches of a phase may conduct simultaneously, resulting in a short circuit across the DC link, a situation that does not effect unipolar drives. The performance of unipolar drives may be improved by increasing the number of phases but at the expense of increasing cost due to the increased silicon device count in the drives, the greater coil count in the motor and more complicated lamination. Permanent magnet motors can have surface mounted magnets or have magnets embedded in the rotor. A rotor with embedded magnets exhibits saliency effects that can add to the peak torque of the machine. The switched reluctance machine can be described as a doubly salient, singly excited motor. The salient rotor construction leads to the machine having a high torque/rotor inertia ratio as sections of the rotor are removed to create the rotor teeth. Torque production is brought about by the alignment of stator and rotor teeth and the phases are electronically commutated to produce a continuous torque with the sequence of the phase excitation determining the direction of rotation of the rotor. As the motor functions with unipolar currents, a simpler drive may be used to control the machine, pointing to a reduction in cost and making fault mitigation easier. The lack of permanent excitation field (magnets) gives the machine a high degree of fault tolerance with benign failure modes, and operation is still achievable when the machine is in a fault state. However, the pulsed nature of torque production leads to a number of undesirable properties such as a high

degree of torque ripple that causes speed fluctuations which is exacerbated by the low rotor inertia. SR machines may have a torque ripple of the order of 80% [4] and the machine can also be acoustically noisy. Motor Type Brushed DC BLDC/ BLAC SR Torque density Low High Speed range Low Medium Efficiency Low High Cost Low High

Med High Med Med Table 1 – Relative comparison of motor types

LINEAR DRIVES Linear drives create force in a similar way to that in which rotary machines create torque, and as such are subject to the same drawbacks, i.e. they have poor force capability relative to hydraulic devices. However, they do remove the need for mechanical linkages if the force requirement is not too severe. They can be classified into groups as follows: Solenoids Basic solenoids provide attractive forces only and the solenoid is reset to the off position by a spring. However, bi-directional motion is possible from these devices with correct design. The force profile of a solenoid is particularly non-linear, being highly dependent on the working airgap between the solenoid plunger and the coil housing which makes their control demanding. Due to the force/displacement characteristic, solenoids are mostly used in a bistable mode of operation. Some proportional controllability is possible, but the penalty in this is a reduction in maximum generated force. To a first order, solenoid forces can be described using equation (1).

( NI ) 2 Aµ 0 F= 2 2l g


Where N is the number of turns of wire in the solenoid coil, I is the coil current, A is the working cross sectional area of the solenoid, µ0 is the permeability of free space and lg is the airgap length. From equation (1) it can be seen that as the working airgap lg reduces, the output force becomes very large, making control difficult. Moving coil actuators Moving coil actuators include devices such as speaker coils, which feature very fast responses. However, output forces can be limited due to the coil being in the working airgap, which limits the ability to cool it, robustness is also an issue.

Moving magnet actuators In general, moving magnet linear actuators are very robust as they have no flying leads to the magnet shuttle, and the stationary winding arrangement on the stator allows efficient removal of any losses. The use of magnets could be problematic in hot environments because of their poor temperature capability. Linear motors Linear motors may be considered as multi phase moving magnet, coil or solenoid actuators, and are available in the same variants as rotary machines. As in rotary machines, phases are commutated to allow a greater degree of travel. POWER SUPPLIES AND 14/42 VOLT SYSTEMS FOR ELECTRIC ACTUATORS Ultimately, current levels drawn by electric devices are governed by the capability of the battery and the motor operating voltage. The nominal current demand and voltage also defines the type of power electronic devices required. In this respect, higher voltages and lower currents are more appropriate as they are easier to deal with from a power electronics point of view. There is also a greater degree of choice in terms of components compared to lower voltage / higher current rated devices. In motor design terms, once a satisfactory magnetic circuit has been designed and the device operating speed has been specified, the operating voltage dictates the form of the windings with regard to number of turns and wire diameter. For a given machine power the amount of current drawn by a 42 Volt machine will be 3 times less than for a 14 Volt machine. This dictates that the number of winding turns must be altered to satisfy the motor Amp-turns requirement, which is a constant set during the design of the magnetic circuit. For correctly designed 14 and 42 Volt machines, the winding losses should be approximately equal. The voltage levels cause differences in the drive losses and useful available voltage. The higher current demanded by 14 Volt machines leads to greater I 2R losses in the drive electronics. The voltage dropped across each power semiconductor device reduces the applied winding voltage, and in the case of brushed machines, the brush voltage drop is also a factor. As these voltage drops are essentially fixed for both 14 and 42 Volt cases, the 14 Volt system suffers a larger percentage loss of useful applied voltage, giving the 42 Volt system the upper hand in terms of both efficiency and performance. However, if the duty cycle is sufficiently short, these losses should be relatively negligible providing the device has stable states where power consumption is low or zero. Figure 1 shows a comparison of the loss components between 14 and 42 Volt motor drives.
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 14 Volts

14/42 Volt comparison

normaliesed losses



42 Volts

Figure 1 – Comparison of 14 and 42 Volt motor drive system losses

As stated earlier, electromechanical devices produce relatively low levels of specific torque and force relative to conventional hydraulic actuation, and to compete with hydraulics, a system is needed to amplify forces and torques. The selection of a mechanism is important to ensure correct system dynamics, system efficiency (losses in the mechanism and whether it provides a self locking feature) and system reliability (wear rate). All three criteria are linked and good design should lead to all the factors being well optimised. Some examples of mechanical systems for transmission control can be found in [5]. SIMPLE LEVERS An electromechanical device may be connected to the clutch release lever as is usually done with the cable from a manual transmission clutch pedal. However, this type of device offers no locking feature requiring a constant force from a linear device to operate it. Rotary devices must also include a motion conversion linkage, but the conversion stage may be designed to be self-locking. POWER SCREWS Power screws cover a number of different designs that convert rotary to linear motion. The main types of screw can be divided into sliding (square or Acme thread type) and rolling contact (ball and roller screw type) variations. A power screw alone may not be able to provide the necessary reduction gearing to multiply the motor torque to a usable level. A simple gear set or epicyclic may also be needed between the machine and power screw, determined by the space available for the machine and the lead of the power screw. WORM GEARS Although worm gears have a low efficiency, they provide a method of packaging a set of gears with a large reduction in a small volume. However, worm gears introduce

backlash, and the poor efficiency of the gears will increase the size of the motor due to increased torque requirement. A further linkage is required to convert the rotary output from the worm gear set to a linear force. ELECTROHYDRAULIC CYLINDER A closed circuit hydraulic system may be used so to allow electromechanical devices to be easily placed away from high temperatures or to areas where space is less limited.

• Low cost A balance of these attributes is required to optimise a particular system. Many factors influence the design considerations and use of permanent magnet machines, however, for simplicity the motor considered was a conventional surface mounted permanent magnet machine. Magnetic materials used in surface mounted rotors such as Neodymium Iron Boron (NdFeB), are susceptible to damage from high ambient temperatures and corrosive environments because of their high iron content. Ferrite magnets, which may also be used in brushless machines, are much cheaper, more robust and have a better temperature dependent characteristic than NdFeB, but are poorer in magnetic terms, leading to an increase in the amount of magnetic material required. Samarium Cobalt (SmCo) magnets produce fields of the same order as NdFeB and have a much better temperature characteristic, but are very expensive and are very brittle which makes handling them difficult. NdFeB magnets were selected because they offer the best performance/cost compromise, although temperature performance is still a significant issue. Figure 3 shows an example of how motor efficiency varies as a function of ambient temperature for two different types of NdFeB magnet. The higher grade, more expensive magnet exhibits greater efficiency through the entire temperature range.
Motor performance vs temperature
100 90 80 70 60 50 40 30 20 10 0 20 70 120 Temperature (degrees C) 170




Figure 2 – Basic design concept of for motor driven power screws showing (a) external and (b) integrated power screw arrangement The design and evaluation of a simple clutch actuator was conducted in terms of the rating the motor and conducting simulations to evaluate performance. The aim was to design an actuator such that the mechanical linkage used was integral to the torque producing mechanism to give an actuator that minimised volume, maximised output force and gave the required dynamic response. The actuator was based on integrating a power screw with a brushless DC motor, and figure 2 shows two examples of such designs. The brushless DC motor was chosen because it has a high torque density, low torque ripple and high efficiency. The power screw was chosen because it allows the multiplication of force and allows motion conversion in one unit. The basic requirements of the system should be: • • • • • • • • • High power/weight ratio giving the lowest machine mass High torque/inertia ratio giving the best acceleration possible Smooth production of torque particularly at low speeds to minimise speed variation and achieve good positional accuracy. Controlled torque at zero speed High maximum speed of operation. High efficiency and power factor to minimise drive VA requirement Compact integrated design with the application Good frequency response Low backlash

Motor efficiency

Low grade magnet High grade magnet

Figure 3 – Motor performance for two different types of magnet versus ambient temperatures

MECHANICAL DESIGN Initially, a figure for the required output force was needed to establish the torque requirement of the motor, which in turn leads to an approximate motor volume. The required torque is dependent on the load, system friction, the screw lead, the screw diameter and the system inertia. The torque requirement for a power screw having a lead L, a coefficient of friction µ and a diameter D that provides a force F is:


FD  L + πDµ  2  πD − µL   



πTRV Dr 2 L


To satisfy the criteria of low or zero power consumption whilst the screw is stationary, the screw was designed to lock. Power screws may become self-locking when πµD > L or µ > tan λ where λ is the lead angle of the power screw. The screw lead was selected to give high positional accuracy, repeatability, rigidity and the keep the torque requirement of the machine to a minimum. Small lead power screws can be achieved using a planetary roller type screw, where the inner and outer members are threaded and there is a planetary carrier supporting a number of threaded rollers. This type of power screw is better than a ball screw in that it can achieve smaller leads and can tolerate higher loads and greater speeds, however, they have a slightly increased coefficient of friction. Based on the force requirement to directly actuate a clutch, a nominal coefficient of friction of the screw and by specifying a lead, it is possible to obtain a figure for the torque requirement of an electric motor. As machine torque is proportional to machine volume, the space envelope of the machine can be estimated.

where TRV is torque per unit rotor volume. The total machine volume is found in the same manner as before, by defining an aspect and split ratio. After basic estimations of the required motor volume had been obtained and seen to be of an acceptable order of magnitude, specific motor design was undertaken using a software package developed by the University of Sheffield Electrical Machines and Drives group named ERINI (Electrically commutated, Radial field, Inner rotor, Non-overlapping, Iron cored). This package provides estimates for motor parameters such as motor efficiency, back EMF waveforms, phase resistances and inductances based on detailed motor input data.
Normalised Brushless DC motor efficiency figures
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 1 0.9 0.9 1 0.9-1.0 0.8-0.9 0.7-0.8 0.6-0.7 0.5-0.6 0.4-0.5 0.3-0.4 0.2-0.3 0.1-0.2 0.0-0.1



To a first order, the equation for motor torque from a given machine can be shown to be


πBQD 2 L


Figure 4 – Motor efficiency as a function of motor dimensions for a given torque and speed As motor volume and weight were to be minimised, the effect on motor efficiency was examined whilst motor dimensions were reduced from the original values calculated from equation 3. Figure 4 shows normalized efficiency variation versus motor dimensions for a given torque. As expected, it shows that the diameter of the machine has greatest impact on the motor efficiency due to motor torque output being related to rotor volume. Machines become less efficient as their size reduces as there is less area for windings and laminations, which requires the coils to carry more current, increasing copper losses, and the laminations have to carry more flux, increasing iron losses. Table 2 gives values for a BLDC motor used to directly drive a power screw. This data was used to construct a dynamic SIMULINK™ model of an actuator.

Where Β and Q are machine specific quantities, magnetic loading (T) and electric loading (Am-1) respectively, and D and L are the rotor diameter and axial length. B and Q may vary between 0.3 – 1T and 10 – 40Amm-1 respectively, dependent on machine design and cooling. As this equation describes the rotor volume, by defining an aspect ratio ( (

D ) and a split ratio L

Drotor ), motor dimensions can be estimated. Typical Dstator

split ratio values may range from 0.5 to 0.7 dependent on motor type. This equation applies to brushed DC and brushless permanent magnet motors with appropriate values of B and Q. Switched reluctance machines must be treated separately as they produce torque in a different manner and an alternative approach is required to design these machines. One such method is to have a previous knowledge of machine torque per rotor volume [6] generated for different machine types and applications. Torque output can then be found from

0.6 0.6

Motor length




Motor diameter

Normalised motor efficiency


Base speed 1000RPM Torque at base speed 2Nm Axial length 70mm Case diameter 80mm Weight 1.65kg Number of pole pairs 2 Rotor inertia 3.8×10-5kgm2 Table 2 – Direct drive brushless DC motor design figures



Figure 5 – Basic motor model circuit This represents the electrical input power (VI), the losses in the system (I2R) and the mechanical output power (EI). This shows output power is the product of the motor back EMF and the coil current. As brushless PM motor currents are bipolar and in the case of brushless DC, discontinuous, and because of the influence of saliency on the back EMF waveform, ripples on the average torque are produced as shown in figure 6. The ERINI software was used to give an estimate of the motor torque ripple magnitude. These torque variations will cause fluctuations in the motion of the screw, and cause variations in the force applied the clutch. This can have a direct impact on the launch and shift quality of a vehicle. Greater motor torque density (and hence a smaller machine) can be achieved by introducing a saliency torque contribution at the expense of introducing increase torque ripple. The use of switched reluctance and salient machines in this type of application is critically governed by the ability of any system to tolerate the increased level of torque ripple that these machines develop. BLDC ripple torque The ideal back EMF waveform for a brushless DC machine is trapezoidal with square wave coil currents, a more realistic back EMF wave form may be represented by sin θ + 1 3 sin(3θ ) which attempts to account for some variation in saliency and saturation effects. The combined result with 2 phases conducting at any one time is a periodic ripple. Electrical degrees may be converted into mechanical degrees by multiplying by the number of motor pole pairs, which gives the torque ripple per mechanical revolution of the rotor. Figure 6 shows a simplified example of the components that make up brushless DC motor output torque.

To analyse the dynamic characteristics after the completion of the development of the electromechanical system, a clutch actuator model was created in SIMULINK™. Contributions from the different system variables were combined to simplify the model. The motor rotor inertia, screw nut inertia and carrier inertia were combined, along with the referred inertial contribution from the linear motion parts, and was represented by

J tot = J motor + J nut + J carrier

 L  + m   2π 



where m is the mass of the linear components and L is the screw lead. This leads to

J totα = Tmotor − Tload − Tviscous − TCoulomb J tot d 2θ dθ dθ dt = Tmotor − Fχ − β − Fc 2 dt dt dθ dt

where F is the force on the clutch plates, χ represents the gear ratio and linear conversion, β represents the viscous damping coefficient and Fc represents the Coulomb force component. The stiffness of the clutch plates is represented by a clearance where the plates are brought to meet each other followed by a clutch plate compression stage. Given the small power screw lead, the mass of the linear components is negligible. To a first order, the motor model may be represented by the circuit in figure 5. Which gives the equation below, where E is the motor back EMF

V = IR + E


multiplying by I gives VI = I R + EI


Motor Voltage, current and torque waveforms for a single phase
1 0.8 0.6 0.4 Amplitude 0.2 0 -0.2 0 -0.4 -0.6 -0.8 -1 Angle (degrees eletrical) 100 200 300 Back EMF Coil current Resultant torque

This equation accounts for the winding time constant, which is omitted from equation 7. Figure 7 shows how the motor fits in the overall system.






Figure 7 – Basic system model of motor driven actuator A switched reluctance machine was also modelled having the same torque/speed characteristic as the brushless machine to provide an adequate comparison. The machine was chosen to be of 6/4 design (six stator teeth, four rotor teeth). This decision was taken on the basis that this configuration has the cheapest construction, having simple laminations leading to fewer coils and power electronic devices. However, large torque ripple is exhibited and the frequency of the ripples is low relative to designs with more stator and rotor teeth. The results from the analysis can be seen in figure 9. This shows the larger torque ripple on the SR machine compared to the DC machine. This ripple causes rotor speed fluctuations, but this does not significantly effect the linear position of the power screw.

Torque waveforms
1.4 1.2 1

0.8 0.6 0.4 0.2 0 -0.2 0 100 200 300

1st phase torque 2nd phase torque 3rd phase torque Resultant real torque

Angle (degrees electrical)

ERINI calculated torque variation
2.50 2.00 Torque (Nm) 1.50 1.00 0.50 0.00 0.00


2.00 3.00 4.00 Angle (Radians electrical)



Figure 6 – Torque production in permanent magnet DC machines

To investigate torque fluctuations on the system, a MATLAB/SIMULINK™ model was developed using the basic system dynamic equations. As switched reluctance motors are of interest a direct comparison was also made. The mechanical model was designed with a representative torque ripple generated by ERINI. The torque fluctuation was calculated as a function of the motor position and used to calculate power screw force. A plate stiffness look up table was used to represent the clutch plates. This motor model included a component for the motor winding inductance and hence can be described by the equation

Figure 8 – SIMULINK™ model of the clutch actuator and motor In an attempt to reduce the size and weight of the actuators, some reduction gearing was introduced between the rotor and the input to the power screw to reduce the torque requirement of the motor. The rotor dimensions were chosen so as to minimise inertia to maintain the required dynamic response. A gearset was used with a reduction ratio of 4:1. The motor torque and speed requirement was then adjusted so that the correct power output was obtained. The redesigned machines develop approximately the same percentage magnitude of torque ripple as the machine shown in table 2. Table 3 shows the parameters of the redesigned motors.

V = IR + L

dI +E dt


Base speed Torque at base speed Axial length 50mm 40mm Case diameter 50mm 120mm Weight 0.415kg 0.819kg Number of pole 2 2 pairs Rotor inertia 1.2×10-5kgm2 2.1×10-4kgm2 Table 3 – Low torque, high speed brushless DC motor design parameters From table 3, the lower torque requirement reduces the conventional motor volume and weight significantly, with both quantities reducing proportionally (~4:1). A large outer diameter (OD) machine is also presented for comparison. The large OD machines internal diameter was fixed to allow the power screw to fit inside it as shown in figure 2b. This constraint lead to the large OD machine being heavier.

Conventional machine 4000RPM 0.53Nm

Large OD machine 4000RPM 0.53Nm

Figure 10 – Direct drive brushless PM and SR motor performance figures The greater torque ripple of the SR machine shown in figure 10 causes slight speed fluctuations of the machine, but the linear position of the screw is not significantly effected by these variation in speed.

To develop the clutch actuator design further, the dynamic analysis and modeling was continued using the simulation package EASY5™. EASY5™ contains Ricardo developed models of engine and power train elements and a representative clutch model was created using the clutch model design blocks with typical parameters entered, and a simple spinning shaft representing the engine torque. All electrical elements were taken from the newly developed Ricardo electric drive library, which contains a number of different motor topologies along with associated power electronic drives, batteries and converters. The basis actuator model is shown in figure 11. Here a BLAC machine is used to replace the BLDC machine in the SIMULINK model.

Figure 9 – SIMULIMK™ mechanical system model In packaging terms, the integration of the motor and power screw may provide a better solution in terms of the volume available for the actuator. The simulations were performed to provide information on likely force outputs and step response clutch actuation times as well as to examine the effects of torque ripple produced by the machines on the inertia of the system and the application of the clutch. In developing the models, no control was employed to the motors. Clutch lock up times of the order of 0.3 to 0.5 seconds were achieved.

Figure 11 – Easy5™ clutch actuator model

Motor position, speed and torque outputs
Rotor position (rads) and speed (rads/s) Rotor position (rads) and speed (rads/s)

Motor position, speed and torque outputs
1000 800 600 400 200 0 -200
Motor speed Rotor position


Motor speed Rotor position



-100 0 2.4 2 0.2 0.4 TIME 0.6 0.8 1

0 0.8


0.4 TIME




Motor torque (Nm)

1.6 1.2 0.8 0.4 0 0 0.2 0.4 TIME Model: bldc_hydraulic, Runid: simulation, Case: 1, Display: 1. 24-NOV-2003, 16:05:54 0.6 0.8 1

Motor torque (Nm)



0 0 0.2 0.4 TIME Model: bldc_hydraulic, Runid: simulation, Case: 1, Display: 1. 24-NOV-2003, 16:15:13 0.6 0.8 1

Brushless AC motor currents and power
15 12.5
RMS current (Amps)

Brushless AC motor currents and power
15 12.5
RMS current (Amps)

10 7.5 5 2.5 0 0 20 0.2 0.4 TIME 0.6 0.8 1

10 7.5 5 2.5 0 0 20 0.2 0.4 TIME 0.6 0.8 1

Motor phase currents (Amps)


Motor phase currents (Amps)


A phase


A phase

B phase C phase


B phase C phase



-20 0 0.2 0.4 TIME Model: bldc_hydraulic, Runid: simulation, Case: 1, Display: 3. 24-NOV-2003, 16:05:54 0.6 0.8 1

-20 0 0.2 0.4 TIME Model: bldc_hydraulic, Runid: simulation, Case: 1, Display: 3. 24-NOV-2003, 16:15:13 0.6 0.8 1

Power screw outputs
Screw position (m) and linear velocity (m/s)
Screw position (m) and linear velocity (m/s)

Power screw outputs

0 -0.01 -0.02 -0.03 -0.04 -0.05 -0.06 0 500 0.2 0.4 TIME 0.6 0.8 1
Screw position Screw velocity


Screw position Screw velocity



-0.05 0 500 0.2 0.4 TIME 0.6 0.8 1


Clutch slip


Clutch slip




0 0 0.2 0.4 TIME 0.6 0.8 1

0 0 0.2 0.4 TIME Model: bldc_hydraulic, Runid: simulation, Case: 1, Display: 2. 24-NOV-2003, 16:05:54 0.6 0.8 1

Model: bldc_hydraulic, Runid: simulation, Case: 1, Display: 2. 24-NOV-2003, 16:15:13

Figure 12 – Easy5™ electromechanical clutch actuator model output for 2 Nm motor direct drive power screw

Figure 13 – Easy5™ electromechanical clutch actuator model output for 0.5 Nm motor with additional reduction gearing before the power screw stage

In the simulation, a torque command was applied at zero seconds and the model was used to predict likely actuation times with a BLAC drive machine running from a 42 Volt supply. The mechanical system was represented using an EASY5™ power screw model and clutch coupling elements and takes into account inertia and friction in each of the components. The presented results show the speed and position of the motor and how this translates to the linear motion and velocity of the output from the power screw. The power screw output was then applied to the clutch model to analyse how the clutch model behaved. The results in figures 12 and 13 show the clutch slip falls to zero in approximately 0.2-0.3 seconds. Each set of plots shows the screw position and speed, along with detailed motor data, and clutch cylinder motion profiles for an open loop step response input. As shown in figure 13, using a reduction gearing stage between the motor output and the power screw increases the motor speed and reduces the motor torque requirement, and so reduces the volume of the machine, without significantly effecting performance. In terms of power consumed in the presented model, peak current drawn in actuating the clutch is approximately 13 Amps, which causes the clutch slip to reduce to zero in 0.2 – 0.3 seconds. The resultant level of energy consumed over a driving cycle using this type of actuation would be low due to the duty cycle of the clutch.

sufficiently relied upon to give an appropriate representation of applied force to the clutch plates. Therefore, the inclusion of an appropriate force sensor may be necessary.

The main problems that are encountered when attempting to design electromechanical systems for automotive applications are the vehicle voltage and low actuator power density, which leads to large actuators for a given force compared to hydraulic systems. However, if the entire actuation system is considered, electromechanical devices become competitive against hydraulic systems that include pumps, accumulators, filters and other ancillary components needed for the system to operate. Electromechanical systems also hold a significant advantage in efficiency terms. Because electromechanical actuators can be designed to be very efficient, they can clearly make an impression on overall transmission efficiency. Simulations show that the dynamic performance of motor driven actuators falls within reasonable limits of performance. Although currently, brushed motors dominate actuation in automotive environments, given the level of performance expected by the driver for clutch actuation events, higher specification technology is required. As with the majority of applications that use electromechanical device, some means of gearing is needed to improve torque output from the machine. With the introduction of gearing, motor torque, weight and rotor inertia is reduced without the system dynamic response suffering. The models also show that the torque ripple produced by the motors, particularly SR machines, is of an order that is low enough to not significantly effect the mechanical system performance. For further electric actuation applications, a greater degree of integration of the electromechanical system is needed to assure that system expectations are met. The design of further electromechanical actuators is currently underway with the use of novel actuation methods being investigated. Great attention is being put on accurate duty cycle estimation as this has a direct impact on actuator size. As all presented work shows results from open loop control techniques, further work is being carried out to resolve the control issues of using these actuators in closed loop systems. Other current research and development work being undertaken by Ricardo DTS is aimed at removing the effects of backlash from mechanical systems using novel, low cost means.

All these simulations assume zero error for all signals of interest for the control of the actuators. Ultimately, force is usually the parameter of choice for control of this sort of application. Traditionally, motor torque is determined through measurement of the current supplied to the motor. If this method were to be applied to a power screw driven clutch actuator, the coefficient of friction of the mechanical system would need to be determined very accurately because the motor torque requirement is highly dependent on this factor. The coefficient of friction is dependent on the state of wear of the mechanical system, the level of lubrication and the temperature of the lubricant. Thus, estimating applied forces from the measurement of current is dependent on a significant number of changing factors that may be difficult to determine. Another motor parameter that is usually considered is rotor position, which is usually measured with a resolver or encoder attached to the rotor shaft. However, the measurement of position to determine force by knowing an appropriate force/displacement characteristic presents similar problems as with the measurement of motor current for the determination of torque. The force/stiffness characteristic of the materials to which the forces are applied are dependent on temperature and must be adjusted for wear and hence can not be

1. Design of a New High-Performance Electrohydraulic Actuator Habibi, S.; Goldenberg, A. IEEE/ASME Transactions on Mechatronics, Vol. 5, No. 2, June 2000 2. Application Characteristics of Permanent Magnet Synchronous and Brushless DC motors for Servo Drives Pillay, P.; Krishnan, R. IEEE Transactions on Industrial Applications, Vol. 27, No. 5, September/October 1991 3. Low cost multi phase brushless DC motors Arefeen, M.S.; Gopalarathnam, T.; Waikar, S.; Toliyat, H.; Moreira, J.C.; 4. A Unifies Theory of Torque Production in Switched Reluctance and Synchronous Reluctance Motors Staton, D.A.; Soong, W.L.; Miller, T.J.E. IEEE Transactions on Industrial Applications Vol.31, No2, March/April 1995 5. Integrated Shift-by-Wire for an Automatic Transmission Burgbacher, M; Martin, B.; Newberry, R.; Weber, M. Global Powertrain Congress 2001 - Advanced Transmission/Drive Line System Design & Performance, Jun 2001, Volume 19 6. Switched Reluctance Motors and their Control Miller, T.J.E. Clarendon Press, 1993 ISBN 0198593872

Degree electrical = degrees mechanical × number of pole pairs

The authors wish to thank Ricardo DTS and especially Jon Wheals, for their help and for the opportunity to prepare and publish this paper, and the University of Sheffield Electrical Machines and Drives Group for their help and assistance and the use of their machine design software.

Andrew Turner Electrical Engineer, Controls, Electronics and Calibration, Ricardo Driveline and Transmission Systems Radford Semele, Leamington Spa Warwickshire, CV31 1FQ, UK Direct dial: +0044 (0)1926 319379 Switchboard: +0044 (0)1926 319319 Facsimile: +0044 (0)1926 477130 Email: andy.turner@ricardo.com

AMT1 – Automated manual transmission, torque interrupt AMT2 – Dual clutch automated manual transmission AT – Automatic transmission B – Magnetic loading/flux density – Tesla BLAC – Brush-less Alternating Current BLDC – Brush-less Direct Current DCT – Dual clutch transmission E – Motor back EMF – Volts rpm-1 EMF – Electromotive force EMI – Electromagnetic interference L – Power screw lead N – Number of turns of wire in a winding NdFeB – Neodymium Iron Boron Q – Electric loading – Amps m-1 Saliency – The shape of actuator iron/steel parts that can aid force and torque production, but may introduce other effects such as torque ripple SmCo – Samarium Cobalt SR – Switched Reluctance VA – Volt Amps µ - Coefficient of power screw friction µ0 – Permeability of free space (4π×10-7) λ - Power screw lead angle

Shared By: