PM Motor Drives for Automotive Steer-by-Wire
Nicola Bianchi(1), Silverio Bolognani(1), Michele Dai Pr´e(1), Matteo Tomasini(1), Luca Peretti(2), and Mauro Zigliotto(2)
(1) Department of Electrical Engineering, University of Padova, Italy,
(2) Department of Engineering and Management, University of Padova, Italy
Abstract— This paper presents the analysis and design of a steer-by-wire drive system, developed in the frame of a two- years inter-university research project. The application requires fault-tolerance capability, effective steering control, and torque feedback control. The study focuses on the details of the PM motor used in the steering mechanism, which features innovative design concepts. Experimental results are included in the paper.
Index Terms— AC motor drive, PM motor drive, Interior PM motor drive, Fault-tolerant machines
I. INTRODUCTION
THE electric power steering presents several advantages over the conventional hydraulic power steering, as en- hanced fuel economy, assist action given as function of vehicle speed and still available with the engine off, reduced compo- nents, modular assembly, customised steering characteristics, and no need of hydraulic fluid [1].
The steer-by-wire system requires the use of two motor drives. The first motor is for the mechanical steering sys- tem, while the second delivers the torque feedback to the steering wheel, as shown in Fig. 1. An Interior Permanent Magnet (IPM) synchronous motor [2] is used for the steering actuator while a Surface-mounted Permanent Magnet (SPM) synchronous motor, with fractional–slot winding [3], is used for the torque feedback.
Current sensor
Steering wheel
Inverter
Controller
Inverter
IPM Motor and ball-screw
Front wheels Rack
Tie rod Position sensors
SPM Motor
Fig. 1. Scheme of the steer-by-wire system.
Since the IPM motor is mechanically linked to the steering rack of the vehicle, some peculiar features are expected:
• Smooth, ripple-free torque, to avoid mechanical vibra- tions and audible noise [4].
• High efficiency, with the steering system powered only when needed.
• Fault-tolerance, as a stringent requirement in electric motors for steering application. In the case of a phase short–circuit, the inherent motor braking torque [5], [6]
has to be kept as small as possible, to let alive the steering capability of the driver. [7], [8].
• Minimised size and weight, which often make a rotating motor with gearbox preferable to a direct drive system.
Details of the design and test measurements on the IPM motor prototype are included in the paper.
As regards the control aspects, they have been mainly devoted to the design of the algorithms for the combined control of both steering and force feedback motors. The IPM motor requires a smooth position control, not particularly fast.
The SPM motor drive has to implement an appropriate and customable steering feel, through a suitable torque feedback control.
The reaction torque on the steering wheel consists of two components. The main component is a function of the contact patch reaction torque. It is estimated by an observer since neither force transducer nor torque transducer are used in the control scheme. A side component of the reaction torque is added to customise the steering feel and response. In the pro- posed work, different steering feels have been implemented, depending on the vehicle velocity.
A complete laboratory test bench has been developed for the pre-compliance tests. The mechanical steering load, bulky and expensive, has been emulated by a special servo drive, which has been used to verify the effectiveness of the control algorithms.
Implementation details, simulations and experimental re- sults on the test bench are included.
II. THE MECHANICAL SYSTEM
The main characteristic of the proposed steer-by-wire sys- tem is the absence of any mechanical link between the steering wheel and front tyres. The steering torque is provided to the tyres by means of the traditional mechanical coupling, driven by two IPM motors as schematically shown in Fig. 1. A proper choice of the geometrical parameters of the mechanism is required to minimize the steering error.
Fig. 2. Electric power steering rack. In evidence, the two motors and the ball-screw.
For the sake of safety, the system features an hot–
redundancy, in which all of the electric components are dou- bled: double batteries, inverters, feeding circuits and electric motors, both insisting on the same shaft and driving the steering rack.
In a traditional steering system, the motion conversion (from rotating to linear) is obtained by means of a pinion–rack gear–
set, which is bulky and not always safe in case of sudden car crash. In the present application, the rotatory motion is accomplished by connecting the rotors of the two IPM motors to a screw-nut mechanism that has rollers or balls as interposed elements. The ball-screw mechanism satisfies all the mechanical requirements imposed by car manufacturer, with improved efficiency and lower cost and size. The selected screw-nut mechanism also yields higher dynamics and posi- tioning precision even though heavily loaded.
A schematic draw of the connection mechanism between the two motors and the main shaft of the mechanical linkage scheme is shown in Fig. 2. The stators of the two IPM motors are directly connected to the chassis, while the rotors are linked to a hollow shaft hold on by two rolling bearings, able to sustain the axial force. The hollow shaft forces the nut of the ball-screw to rotate, and the motion is transmitted to the screw. An anti-rotational pin is provided at both the ends of the screw in order to avoid the rotation. A metal ring is screwed on the left hand side of the shaft. It joints tightly together all the components. The angular position is measured by a resolver, which is the component that best bears the high temperature inside the engine compartment. The design of the whole system has been carried out by containing its overall size below 130 mm diameter and 350 mm length. The distance between the two connection points of the shaft is 875 mm. The operational range of the screw is ± 75 mm, with a maximum linear speed of 250 mm/s (for load up to 3000 N) and a maximum acceleration of 8 m/s2. The ball-screw mechanism has a nominal diameter of 20 mm and a lead of 5 mm, and it can operate with a dynamic load of 11.7 kN and a static load of 18.3 kN.
The proposed system is addressed to a EuroNCAP small family car (Compact, according to North America classifica- tion, or C segment). A static peak force (in stall conditions) of 10 kN is forecast during a parking operation. The force is required for at least 5 s, with a duty cycle of 5 s over 1 minute. The steering rack is required to sustain a continuous force of 7.5 kN (at a speed of 40 mm/s) with a duty cycle of
180 seconds over 15 minutes. At last, a maximum dynamic force is required to be of 3 kN (at a speed of 200 mm/s) for at least 60 seconds over 4 minutes.
The steering action has to be guaranteed even in case of failure of one of the two drives or motors, although with reduced torque, and according to the failure type. The worst case is the complete loss of an inverter or motor, so that only one healthy motor remains. The corresponding steering system performances (forces, acceleration and speed) are halved, but temporary operations are still guaranteed.
The hot-redundancy concept is worth a deepening. In case of a phase short-circuit, the faulty IPM machine brakes the healthy one, possibly causing the jamming of the steer mech- anism. The design has faced this delicate issue, so that even a single healthy motor is able to carry on the commanded steering action.
III. CURRENT AND BRAKING TORQUE IN CASE OF FAULT
The worst motor fault could be the three-phase short-circuit.
In that case, because of the PM flux linkage, the IPM motor acts as a brake, hindering the steering movement [9], [10].
The analysis of the steady-state three-phase short-circuit and the braking torque is carried out in the synchronous d − q reference frame. By posing the voltage components equal to zero, i.e. vd = vq = 0, the currents id and iq are computed [11]. The resulting braking torque is
Tbrk = −3
2pRΛ2mω R2+ ω2L2q
(R2+ ω2LdLq)2 (1) The maximum value of Tbrk is found by posing the deriva- tive of (1) with respect to the speed ω equal to zero. It results a maximum braking torque given by
Tbrk∗ = −3 2pΛ2m
Lq f (ξ) (2)
where the function f (ξ) is a function of the saliency ratio ξ = Lq/Ld. In the range between ξ = 2 and ξ = 6, such a function can be approximated by the straight line f (ξ) = ξ −1 [12].
A. IPM motor design hints
To match the required mechanical performance, described in section II, the IPM motor shall feature an extended speed range, up to the flux-weakening region. For the sake of generality, this paper will adopt the p.u. values, from now on denoted by lower case characters. With a base motor torque τb = 1 p.u., the base speed ωb = 1 p.u. and the nominal voltage vN = 1 p.u., it is possible to express the motor parameters λm, ld, lq and the nominal current iN as a function of the required flux-weakening performance [13]–[15].
The IPM motor drive parameters are chosen to fulfil the system specifications with the minimum short–circuit braking torque τbrk∗ , e.g. lower than 0.1 p.u.. This is obtained by a peculiar combination of the IPM motor magnetic parameters, as detailed in the following.
The saliency ratio ξ is determined by the rotor configuration (e.g. with one, two, or more flux–barriers per pole). Then, by
assuming f (ξ) ≈ ξ − 1, a precise ratio between the PM flux linkage λmand the q-axis inductance lq is obtained from (2).
For example, let us fix τbrk∗ = 0.1 p.u.. Assuming a unity p.u. q-axis inductance, the PM flux linkage results λm ≈ 0.18 p.u. and λm≈ 0.14 p.u. for a rotor saliency ξ = 4 and ξ = 6 respectively. Then, to get the required flux-weakening performance in a speed range equals to 3, the reduction of PM flux linkage causes an increase of the nominal current, which becomes iN = 1.56 p.u. and iN = 1.47 p.u. with ξ = 4 and ξ = 6 respectively [13].
B. Magnetic energy considerations
Expressing the PM flux linkage Λm and the q-axis induc- tance Lq by means of classical equations, and introducing them into (2), the maximum braking torque becomes
Tbrk∗ = −2pWmf (ξ) (3) where p is the number of pole pairs and Wm is the magnetic energy in the air-gap volume due to the PMs only.
Eqn. (3) gives a simple relationship among the given maxi- mum braking torque, the main geometrical dimensions of the machine, and the air-gap flux density due to PMs. The next section will show how to determine the maximum air–gap flux-density ˆBg0. In particular, ˆBg0 has to decrease
• when the number of motor poles increases;
• when the saliency ratio ξ increases;
• when the motor size increases.
IV. IPMMOTOR DESIGN
A. Performance requirement to each motor
The ball–screw lead is 5 mm and its mechanical efficiency is about 80%.
By the mechanical requirements, the maximum speed of the motor is fixed to 3000 rpm (i.e. 314.2 rad/s).
In order to guarantee the peak force, the maximum torque required to the couple of motors is 10 Nm (considering the ball–screw efficiency). Therefore, the maximum torque of each motor is 5 Nm, with a duty cycle of 8.5%.
The maximum angular acceleration is about 10000 rad/s2 (corresponding to the linear acceleration of 8 m/s2). This is reasonable value, for PM machines.
In order to guarantee the other operating conditions, each motor has to produce a torque of 3.8 Nm at a speed 480 rpm (duty–cycle 20%), and of 1.5 Nm at a speed 2400 rpm (duty–
cycle 25%).
B. Geometrical constraints
A commercial lamination has been selected for the stator.
The external diameter De = 120 mm is imposed by the application (see Section II). The corresponding inner diameter is D = 70 mm for a 4–pole motor. For this size, commercial laminations have been found with a number of slots Q = 24.
As far as the harmonic content due of the air–gap M.M.F.
distribution is concerned, a higher number of slots should be preferred, e.g. Q = 36.
The air–gap thickness is fixed equal to g = 0.5 mm, and the stack length equals Lstk = 40 mm. The total length is composed by two modules, of 20 mm each, skewed by the appropriate angle.
In order to facilitate the experimental tests in the laboratory with industrial inverters, the motor winding was designed for 300 V D.C. bus voltage, higher than 42 V, that is becoming a standard in automotive applications [16].
As mentioned, the entire steering system, including the drive, is placed inside the engine compartment. The high temperature reduces the allowable temperature rise in the motor windings, and therefore the current density has to be adequately low.
C. Design considerations
From the required flux–weakening capability, the mechan- ical characteristic of the IPM motor has a constant–torque region, followed by a constant–power region [17], [18]. The base speed is fixed to 900 rpm so that the flux–weakening speed–range is 3.33.
The aforementioned considerations, together with that of a low braking torque in case of fault, suggest to adopt an IPM motor characterized by a limited amount of PMs. As a consequence, the IPM motor has to be characterized by a high saliency ratio, and this justifies the choice of an IPM rotor with three flux–barriers per pole.
As regards the maximum braking torque, a value Tbrk∗ = 0.30 Nm has been fixed. The number of poles is 2p = 4, and the saliency ratio is estimated to be ξ = 6, then f (ξ) ≈ 5.
Thus, from (3), the magnetic energy in the air–gap at no–
load results Wm= 17 mJ. Since the air–gap volume is fixed (D = 70 mm, Lstk= 40 mm, and g = 0.5 mm), the maximum air–gap flux density (the peak value of the ideal sinusoidally distributed waveform) results ˆBg0= 0.1 T.
D. Finite element simulations
A finite element analysis is used to refine the IPM motor design. The selected IPM motor configuration is reported in Fig. 3 (a). The flux plot at no–load and the computed flux linkages of the motor are shown in the part (b) of the same figure, in which a synchronous d-q reference frame, with the direct axis d fixed to the PM flux linkage, has been selected.
The computed unsaturated q–axis inductance results Lqu= 0.317 H, while the d–axis inductance is Ld = 0.055 H. The unsaturated saliency ratio results ξ = 5.7.
E. Torque ripple minimization
A finite element analysis has been carried to design the IPM motor for a minimal torque ripple [19].
The rotor length Lstk= 40 mm is split in two modules of 20 mm. These two modules have been skewed each other to reduce the torque harmonic of 24–th order. This is reported in Fig. 4, which shows that the torque harmonic of 12–th order remains almost the same with two skewed modules, while that of 24–th order is almost cancelled, and those of upper orders decrease.
(a) Structure and flux plot at no–load
−6 −4 −2 0 2 4 6
−0.1 0 0.1 0.2 0.3 0.4 0.5 0.6
currents (A)
flux linkages (Vs)
λq(iq)
λd(id) id, iq
id
−6.25 A 0 A
iq 0 A 7 A
(b) d − q axis flux linkages vs. currents Fig. 3. Results of the finite element analysis of the IPM motor
Moreover, since the hot–redundancy feature requires two motors on the same shaft, it is possible to skew one motor on the hollow shaft with respect to the other, to have the torque harmonic of 12–th order of the two motors 180 degrees out of phase. The resulting harmonic content is shown in the third plot of Fig. 4, which refers to the average torque of both motors.
The resulting torque ripple is slightly higher than 2%, referring to a rated torque of 2.6 Nm.
6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 0
0.01 0.02 0.03 0.04 0.05 0.06
No skewing
Amplitude (Nm)
Harmonic order 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 0.01
0.02 0.03 0.04 0.05 0.06
Two skewed modules
Amplitude (Nm)
Harmonic order 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 0.01
0.02 0.03 0.04 0.05 0.06
Two skewed motors
Amplitude (Nm)
Harmonic order
Fig. 4. Torque harmonics in different motor solutions
The final assembled rotor is shown in Fig. 5.
F. Measurements on the prototype
At first, the PM flux linkage has been measured at no–load.
The back E.M.F. has been measured with the motor running at 1000 rpm. By integrating the back E.M.F., the flux linkage due to the PMs are obtained. Fig. 6 shows the measured flux linkages due to the PMs. The peak value is Λm= 75 mVs.
The electromechanical torque has been tested at different d– and q–axis currents. The dots in Fig. 7 correspond to the
Fig. 5. The final rotor assembled with two skewed modules
0 10 20 30 40 50 60
−0.1
−0.05 0 0.05 0.1
Time (ms)
(Vs)
PM Flux linkages
Fig. 6. Measured flux linkages due to the PMs
measured torque in the d– q–axis current plane. The solid lines are the constant torque curves obtained from the finite element simulation. The fair agreement between measurements and simulations is evident.
−5 −4 −3 −2 −1 00
1 2 3 4 5 6
d−axis current (A)
q−axis current (A)
Motor torque (Nm) τ=7 τ=6 τ=5 τ=4 τ=3 τ=2
1.06
1.03 2.01
0.92 2.12
2.15 2.59 3.01
2.22 3.01 4.07
0.97 4.03
2.21 3.12 4.19 4.82 5.06
5.08 5.6
3.96 4.8
3.07 1.7 5.1
Fig. 7. Measured torque of the IPM motor in the (id, iq) plane: simulated values (solid lines), and measured values (dots)
In order to measure the short–circuit current and the corre- sponding braking torque, the three terminals have been short–
circuited while the IPM motor was running at different speeds.
The three–phase current waveforms are almost sinusoidal, as shown in Fig. 8 referred to a motor speed of 1000 rpm.
The short–circuit current and the corresponding braking torque at different motor speed are reported in Fig. 9. With R = 7.3 Ω (referring to a temperature rise of 60 K), a saliency ratio ξ = 5.7, and a PM flux linkage Λm = 75 mVs, the maximum braking torque is found from (1) to be Tbrk∗ = 0.25 Nm at the mechanical speed ω∗ = 43.5 rad/s [12]. The measurements reported in Fig. 9 confirm this analytical result.
0 10 20 30 40 50
−2
−1 0 1 2
Time (ms)
Currents (A)
Short circuit current at n = 1000 rpm
Fig. 8. Measured short–circuit currents
0 50 100 150 200
0 0.5 1 1.5
(A)
Short−circuit current
0 50 100 150 200
−0.25
−0.2
−0.15
−0.1
−0.05
Motor speed (rad/s)
(Nm)
Braking torque
Fig. 9. Measured short–circuit current (peak value) and braking torque at different motor speed
V. CONTROL ISSUES
The control of the steer–by–wire system requires a coordi- nate action on both PM motors, i.e. the position control of the IPM motor is tighly linked to the torque feedback control of the SPM motor.
The position control has to be precise and robust against any load variation, which may arise from road conditions, vehicle velocity and so on. The SPM motor drive gives a feedback torque to the driver, reflecting the impact of the vehicle on the road, with the aim to get a more sensitive and safer car driving than that of a conventional steering system.
The control scheme is drawn in Fig. 10. The reaction torque has two components: the main one is a function of the contact patch reaction force, which is estimated by a force observer.
The second one is added to give the driver a good steering feel.
Auxiliary stabilisation could also be included in the block.
For the sake of reliability and cost reduction, torque sensors are avoided. The torque information is got from two observers
Fig. 10. Control scheme.
Fig. 11. Wheel position vs. IPM torque transfer function, Bode diagram.
instead. One of them (the block contact patch reaction esti- mator, shown in Fig. 10) estimates the torque delivered by the IPM motor, using the conventional expression that links torque with direct and quadrature currents [9]. The estimated torque is then devoid of the contributions due to the ball–screw friction and system inertia, so to get the net contact patch reaction.
The use of the estimate torque is twofold. Firstly, it is transferred as a torque reference contribution to the SPM motor drive through a non–linear block, in order to optimize the steering feel. Secondly, it can be used as feed–forward (FF) signal to improve the accuracy of the position control of the IPM motor drive itself.
It is worth to highlight that the position control system of the IPM motor is merged with the torque feedback control loop of the SPM motor. This cross-relation implies that the speed loop cannot feature the usual integral action, to avert instability in the torque control loop. At the most, the speed loop controller could also be removed, as in Fig. 10, where the PD position regulator directly issues the torque reference to the IPM motor drive.
As mentioned, the specifications for the position loop were quite relaxed, with a required bandwidth of about 4 Hz. The PD regulator has been tuned on the basis of the existing phase and amplitude relations between the IPM torque and the wheel angle, respectively input and output of the steering system transfer function. The relative Bode diagram is reported in Fig. 11.
It derives from a proper combination of two subsystems, which account for both the contact patch reaction and the cinematic model of the steering chain. The details can be found in [20]. The system gain, as function of input frequency, is slightly affected by the vehicle speed, and this further eases the PD tuning.
It is a shared opinion that the steering feel is greatly improved when the reaction torque sensed by the driver is an image of the contact patch reaction only. That is why the reaction torque control includes an estimator of friction and inertia torques.
The feedback torque loop has to meet the specification of a 4 Hz bandwidth for a steering wheel rotation of ±15deg, which lowers down to 0.5 Hz for a complete steering wheel
rotation (±180deg). The design of the feedback torque loop has to guarantee stable and accurate operation at any vehicle speed and running condition. For instance, it can be found that the stability of the control at zero speed of the car (parking manoeuvre) decreases with the tyre stiffness. Instability can occur both at the highest stiffness and when the gain between IPM motor torque and feedback torque on the steering wheel is too high.
The problem can be solved by adding a lead-lag compensa- tion to the aforementioned gain. In the present design, oriented to a C-segment commercial car, the zero and the pole of the lead-lag action were respectively of 30 rad/s and 120 rad/s.
These values ensured good stability for any tyre stiffness, allowing the reaction torque gain to be chosen for the best steering feel.
A different operating condition is the free mode behaviour, when the steering wheel is released and the driver applies no torque. In this case the control has to be stable and the initial angular position of the steering wheel has to tends to zero if the vehicle is running, to track the effect of the inherent self-aligning torque. The proposed control system meets this requirement. In particular, it has been found that the lead-lag compensation eliminates the overshoot in the position response.
VI. EXPERIMENTAL SETUP AND RESULTS
A mechanical model of the steering system has been imple- mented in a laboratory test bench, to emulate the mechanical behaviour of the whole steering system. The resulting fast control prototyping (FCP) facility has been exploited to verify the effectiveness of the control algorithms.
Figs. 12 and 13 show the scheme and the reality of the laboratory test–bench. The left–hand side of Fig. 12 represents
IPM Synchronous
Motor
SPM Synchronous
Motor
Customized Inverter
dSPACE ds1102
Industrial Inverter
Industrial Inverter dSPACE
ds1003
SPM Synchronous
Motor
Position Position
Rotor position, actual torque
Rotor position, actual torque
Torque reference
Torque reference 6 PWM
switch pattern Rotor
position
Mechanical joint
Position
Steering wheel
Data exchange
Wheel model
Fig. 12. Scheme of the experimental test–bench setup.
the IPM motor drive. A stiff and low–inertia joint links the IPM and SPM motor shafts (shown inside the dot–dashed
Fig. 13. Photo of the laboratory test–bench.
rectangle in Fig. 12). The SPM plays the role of mechanical load, and it is not a part of the actual steer–by–wire system: as said, it emulates the steering mechanism, including the road feedback from the contact patch [20]. The control of SPM motor reproduces the dynamics of the steering system of the vehicle, delivering a load torque to the IPM motor shaft as a function of the assumed vehicle speed, steering angle, and inertia, speed and elasticity of the steering components. To this purpose, a detailed model of the whole steering system has been derived and implemented on a DSP board with different levels of approximations. The bottom of the right side of the Fig. 12 represents the SPM motor drive for the steering wheel torque feedback.
The SPM motors are fed by two voltage inverters, controlled by a PC plug-in board, featuring a TI TMS320C40 Floating–
Point 32–bit DSP, running at 50 MHz clock rate and 40 ns cycle time. Another control board, based on a TI TMS320C31 Floating–Point DSP, performs the whole control of the IPM drive.
A. Simulations
An exhaustive set of simulations has been performed in order to confirm the control design approach and to compare different solutions. Fig. 14 shows the Bode diagram of the torque feedback control. The input is the load torque distur- bance due to the contact patch (amplitude equal to 2/3 of the IPM motor rated torque), while the output is the steering wheel reaction torque under the condition of locked steering wheel.
The figure shows that the specification on the bandwidth given by the carmaker involved in the project is fairly obtained.
B. Experimental results
Fig. 15 and Fig. 16 show some experimental results carried out on the test–bench: a sinusoidal motion has been imposed to the steering wheel. The first subplot of Fig. 15 shows the reference and actual IPM motor position (the two traces are in practice overlapped); the second subplot represents the position error and the third one reports the torque delivered by the IPM motor. The position error is negligible and torque and position traces are almost proportional (in phase). The latter evidence derives from the dynamic equations of mechanical
10−1 100 101 102
−25
−20
−15
−10
Gain [dB]
10−1 100 101 102
−150
−100
−50 0
Phase [deg]
frequency [Hz]
Fig. 14. Bode diagram of the torque feedback control (simulation).
steering system, which point out that in the low frequency range the mechanical load behaves as a spring element.
Fig. 16 shows an experimental measurement to document the contact patch reaction estimator performance. The ob- server is based on Luenberger algorithm. The measurement has been carried out while the steering wheel was moved manually.
In the figure, the output of the estimator (line B) is compared with the load torque delivered by the virtual mechanical load (line A), showing a rather good match. The vehicle speed is set to 72 km/h during the test. The largest curve in Fig. 16 (line C) is the torque reference of the IPM drive, corresponding to the delivered torque with a good approximation. The differences between the torque delivered by the IPM motor and the road reaction torque are mainly due to the rotor inertia.
0 1 2 3 4 5 6 7
−20 0 20
position [%]
0 1 2 3 4 5 6 7
−1 0 1
pos. error [%]
0 1 2 3 4 5 6 7
−50 0 50
time [s]
IPM motor torque [%]
Fig. 15. Position control response (experimental result).
VII. CONCLUSIONS
This paper has presented the complete electrical drive sys- tem for a C-segment car steer–by–wire application, including design hints of the electric motors and control techniques for precise and stable operations.
The electrical motors have been designed to exhibit an inherent fault–tolerant capability, in order to enhance driver safety. In particular, reduced braking torque and torque ripple have been accomplished.
Fig. 16. Contact patch reaction estimator response.
The control algorithms takes into account the interaction between the IPM motor drive for the steering and the SPM motor drive devoted to the torque feedback. A torque estimator has been implemented and tested as well.
The design was oriented to the application to a commer- cial car, already mass-produced with the conventional power- assisted steering. All of the technical specifications and param- eters used in the design and test of the proposed system have derived from a tight collaboration with the car manufacturer, which will look after the details for the final transposition on the vehicle.
ACKNOWLEDGMENT
This work was financed by the Italian National Ministry of Education, University and Research (MIUR), PRIN 2003 (”Innovative electrical motor drive for power–steering”). Au- thors thank Dr. Diego Bon for his help during the optimization process by means of the finite element analysis and Dr. Luca Tubiana for the virtual load analysis. The authors are very grateful to Dr. G. Teruzzi, Saimag. S.p.a., Pogliano Milanese, Italy, that supplied permanent magnets, and to Magnetic S.p.a., Montebello Vicentino, Italy, for assembling the motor prototype. Finally, the Authors gratefully acknowledge the contribution of ELITE Programme of Texas Instruments.
REFERENCES
[1] S. Mir, M. Islam, and T.Sebastian, “Role of electronics and controls in steering system,” Records of the 29th IEEE Annual Conference of the IEEE Industrial Electronics Society, IECON, November 2–6 Roanoke, 2003.
[2] T. Jahns, G. Kliman, and T. Neumann, “Interior pm synchronous motors for adjustable speed drives,” IEEE Trans. on Industry Applications, vol.
IA–22, no. 4, pp. 738–747, July/Aug 1986.
[3] N. Bianchi, M. D. Pr´e, G. Grezzani, and S.Bolognani, “Design consid- erations on fractional-slot fault-tolerant synchronous motors,” in Proc.
of IEEE International Electrical Machines and Drives Conference, IEMDC’05, 14–18 May 2005.
[4] N. Bianchi and S. Bolognani, “Design techniques for reducing the cog- ging torque in surface–mounted PM motors,” IEEE Trans. on Industry Applications, vol. 38, no. 5, pp. 1259–1265, 2002.
[5] A. Jack, B. Mecrow, and J. Haylock, “A comparative study of permanent magnet and switched reluctance motors for high–performance fault–
tolerant applications,” IEEE Trans. on Industry Applications, vol. IA–32, no. 4, pp. 889–895, July/Aug 1996.
[6] T. M. Jahns, “Improved reliability in solid state a.c. drives by means of multiple independent phase–drive units,” IEEE Trans. on Industry Applications, vol. IA–16, no. 3, pp. 321–331, May 1980.
[7] B. Welchko, T. Jahns, and S. Hiti, “Ipm synchronous machine drive response to a single–phase open circuit fault,” IEEE Trans. on Power Electronics, vol. 17, Sept 2002.
[8] B. Welchko, T. Lipo, T.M.Jahns, and S. Schulz, “Fault tolerant three–
phase ac motor drive topologies: A comparison of features, cost, and limitations,” IEEE Trans. on Power Electronics, vol. PE–19, no. 4, pp.
1108–1116, July 2004.
[9] N. Bianchi, S. Bolognani, and M. Zigliotto, “Analysis of pm syn- chronous motor drive failures during flux weakening operation,” Proc.
Power Electronics Specialists Conference, PESC, vol. 2, June Baveno, 1996.
[10] B. Welchko, T. Jahns, W. Soong, and J. Nagashima, “IPM synchronous machine drive response to symmetrical and asymmetrical short circuit faults,” IEEE Transactions on Energy Conversion, vol. EC–18, Jun.
2003.
[11] T. Jahns and V. Caliskan, “Uncontrolled generator operation of interior pm synchronous machines following high–speed inverter shutdown,”
IEEE Trans. on Industry Applications, vol. 35, Nov/Dec 1999.
[12] N. Bianchi, S.Bolognani, and M. D. Pr´e, “Design of a fault–tolerant ipm motor for electric power steering,” in Proc. of IEEE Power Electronics Specialists Conference, PESC’05, 13–17 June 2005.
[13] N. Bianchi and S. Bolognani, “Parameters and volt–ampere ratings of synchronous motor drive for flux–weakening applications,” IEEE Transactions on Power Electronics, vol. 12, Dec 1997.
[14] W. Soong and T. Miller, “Field–weakening performance of brushless synchronous ac motor drives,” IEE Proc.– Electr. Power Appl., 1994.
[15] W. Soong, N. Ertugrul, E. Lovelace, and T. Jahns, “Investigation of interior permanent magnet offset–coupled automotive integrated starter/alternator,” Rec. IEEE Ind. Appl. Soc. Annual Meeting, vol. 1, Oct Chicago, 2001.
[16] H. Murakami, H.Kataoka, Y.Honda, S.Morimoto, and Y.Takeda, “Highly efficient brushless motor design for an air–conditioner of the next generation 42v vehicle,” Proc. of IEEE Industry Applications Annual Meeting, vol. I, October 2001 Chicago,.
[17] A. Fratta, A.Vagati, and F. Villata, “Design criteria of an ipm machine suitable for field–weakened operation,” Proc. of International Confer- ence on Electric Machines, ICEM, Sept. 1990.
[18] T. Jahns, “Flux–weakening regime operation of an interior permanent magnet synchronous motor drive,” IEEE Trans on Industry Applications, vol. IA–23, no. 3, pp. 681–689, May 1987.
[19] M. Islam, S. Mir, and T.Sebastian, “Issues in reducing the cogging torque of mass–produced permanent magnet brushless dc motor,” Records of the 38th IEEE Ind. App. Annu. Meeting, October 12–16 Salt Lake City, 2003.
[20] S.Bolognani, D.Ciscato, M.Tomasini, and M.Zigliotto, “Virtual me- chanical load setup for steer-by-wire - a case study,” Proceedings of IEEE International Symposium on Industrial Electronics, vol. ISIE’05, Dubrovnik, Croatia, no. ISBN:0-7803-8739-2, pp. 1795–1801, 12–16 October 2005.
