Power Steering: Brushless DC or Switched-Reluctance?

The trend in the automotive industry is to replace mechanically actuated systems with systems based on electric motor technology. The traditional brushed dc motor is not suitable for many of these adjustable speed applications because of the wear associated with the brushes, lower overall power density, and EMI problems associated with commutator arcing. This arcing can become more serious as application requirements move from sub-fractional horsepower to transient integral horsepower. The low system voltage and the resulting high inductive motor currents that the commutator must interrupt can create substantial arcing.

The increase of power electronics components that target the automotive industry (namely, low voltage, high current semiconductors) has motivated automakers and OEMs to consider higher performance motor technologies with the expectation that high volume and improved fuel economy will bring value to their products. Few applications have seen broad acceptance at the 12V level, primarily because the load currents (which are often transient) can be quite high. While you can often size electric motors for their nominal usage, and overloaded in transient situations, the same is generally not true for the power electronics. The requirement that the electronics be sized for peak power requirements results in expensive systems, much to the dismay of the automakers.

There is a proposed move toward a 42V platform in an effort to find some balance between the projected electrical demands of a typical automobile and touch-potential safety. While this voltage does provide some relief from the 12V headaches, such as large conductors and inherent voltage drops across semiconductors, it is not a complete solution for the motor and power electronics designer. Although the challenges are likely not insurmountable, the first two production hybrid electric vehicles to be sold in this country, the Toyota Prius and the Honda Insight, are not examples of low-voltage technology.

Several problems are associated with selecting an appropriate technology for a given application. By way of example, let's consider the design of an electric motor drive for a power steering system.

Motor Technology

The brushless dc (BLDC) and switched-reluctance motors (SRM) have some barriers to market penetration. Unlike the traditional induction motor, the BLDC motor has permanent magnets that are susceptible to high temperature complications — a serious consideration in the automotive industry where ambient temperatures are often specified as -40 to 125°C. The BLDC motor also raises concerns about failure modes. Because the SRM has a reputation for being noisy and has high torque ripple, often the traditionally robust induction motor is force-fit into many applications. While the induction motor is an excellent performer in many industrial applications, the performance margins that the automotive industry demands require that other motor types (in addition to the two presented here) be pulled up the technology curve rather quickly.

The induction motor is ill-suited for most automotive applications because of the difficulties associated with extracting heat from the rotor, efficiency problems over wide speed and power ranges, long end-turns, and a more expensive manufacturing process due to distributed windings. The induction motor will find some automotive applications; however, lighter and more suitable candidates will ultimately replace its deployment. The BLDC^[1-2] and SRM^[1,3-4] technologies presented here represent two of those potential candidates.

Switched Reluctance

Power electronics technology has made the SRM an attractive choice for many applications, though to date it has seen few commercial applications. The SRM is a doubly salient, singly excited synchronous motor. The rotor and stator consist of stacked iron laminations with copper windings on the stator, as shown in Fig. 1, on page 24. To excite the motor, a power electronics inverter energizes appropriate phases based on shaft position. The excitation of a phase creates a magnetic field that attracts the nearest rotor pole to the excited stator pole, in an attempt to minimize the reluctance path through the rotor. Excitation is applied sequentially to the phase windings to produce continuous rotation.

The SRM structure is similar to the stepping motor, but with smooth rotation. Because no permanent magnets or windings are on the rotor, all the torque developed in the SRM is reluctance torque. While the SRM is simple in principle, it is rather difficult to design and develop performance predictions. This is due to the nonlinear magnetic characteristics of the motor under normally saturated operation. Modeling processes support design of the conventional SRM^[5,6]. Perhaps to a greater extent than in other types of motors, the structural design of the SRM requires careful attention. Inattention to structural details is often to blame for an acoustically noisy SRM.

The conventional SRM is relatively straightforward manufacturing, due to the concentrated windings around the stator poles in contrast with distributed windings in the induction motor. There are no windings or permanent magnets on the rotor. The simplicity of construction and lack of costly permanent magnets imply that the SRM would be much less expensive to produce in the quantities that dc or induction motors currently enjoy.

Brushless DC

The BLDC motor consists of stacked steel laminations on the stator. Traditionally, the stator resembles that of an induction motor; however, the windings can be configured so they're not distributed. The rotor on the BLDC motor can take many forms, but all have permanent magnets in some form. Depending on motor size, these magnets can be a full-ring magnet, spoke, embedded, etc. The magnet configuration is tied closely to the performance and manufacturing logistics of the permanent magnet material. Fig. 2, on page 26, shows the cross-section of the BLDC motor considered for the power steering application.

In motors of less than 1.5kW rating, the presence of permanent magnets can significantly increase the motor's power density. When space is tight, the energy density of permanent magnets can easily exceed that of a copper winding. This suggests that the BLDC motor should enjoy an advantage around this power level. However, as the speed range increases, the permanent magnets may become something of a liability unless provisions are made to weaken the field. Field weakening in the BLDC motor is sometimes attempted through phase advance of the current waveform, yet this tends to introduce substantial torque ripple. Field weakening is more natural in permanent magnet motors that have been wound for a sinusoidal back EMF.

Power Steering Specification

Based on required rack performance, steering wheel input, and a mechanical power balance between the motor output and the steering rack output, we can develop a motor requirement. The dc bus is 9V to 19V, 12V nominal. The motor drive system should not draw more than 80A from the dc bus when operated at 14.4V. Assume the inverter efficiency is 93%. The motor may be driven by a trapezoidal or sinusoidal drive.

Table 1 outlines the required performance points for the power steering motor. The 5 Nm points represent the predicted maximum torque for the application. The 7 Nm is a safety margin to account for losses in the mechanical interface between the electric motor and the tires. The required no load speed for the motor is 4,500 rpm. Because the duty is transient in nature, the time requirements are provided to help minimize the size of the motor by not designing to tolerate steady state performance under these conditions.

Switched Reluctance Evaluation

From Table 1, the maximum power output from the motor is about 1kW. The SRM operates in a constant torque region below base speed, and a constant power region above base speed. Thus, it's necessary to determine the shaft speed associated with 1kW at 7 Nm. This base speed point is a key design point and defines the torque-speed profile for the motor and is given by 143 rad/sec=1,364 rpm=1,000W/7 Nm. If the SRM is capable of producing 7 Nm at this speed, then it will satisfy all other points in Table 1. To accomplish this from 14.4V and 80A is quite challenging. The available battery power is 1,152W and the peak shaft power is 1kW. The required system efficiency to accomplish this power and satisfy the battery current requirement is 86.8%. This requires the motor and power electronics to have efficiencies in the range of 93%. Selecting a base speed significantly above the 1,364 rpm could result in excessive phase currents, adding cost to the inverter.

Table 2 summarizes the performance achieved at the specified points for a 4-phase switched-reluctance motor designed for this application. Base speed is between 1,600 rpm and 1,800 rpm. While certain points in the design exceed the battery current requirement, those points are well outside of the envelope of expected torque required by the system. The 5 Nm is the expected maximum torque. The two operating points at 5 Nm demonstrate battery currents below the specified level. In the event that there are static friction issues that require 7 Nm to break free, battery currents in excess of 80A should be expected transiently unless the duration is so short that inverter bus capacitors can supply the additional energy.

Figs. 3 through 5 show typical examples of motor phase current vs. phase inductance below base speed, the energy conversion cycle, and the achievable torque, respectively. While the total torque appears to be high in ripple, the ripple is a result of phase overlap. A certain amount of current shaping during phase overlap can achieve satisfactory torque performance.

Brushless DC Evaluation

A 3-phase BLDC motor was designed for this application as well. An important difference between the SRM and BLDC motors is their behavior at high speeds. The SRM is reasonably good at behaving as a constant power motor above base speed. The BLDC motor, on the other hand, often loses power output capability rapidly as the speed increases. As a result, the design must carefully evaluate points beyond base speed to make sure they fall within the actual torque speed envelope of the motor. Fig. 6 shows the torque speed curve for the BLDC motor designed here.

Notice that the base speed (point where torque demonstrates a rapid drop from 7 Nm) is extended out to more than 2,000 rpm. This is necessary to ensure the motor meets the no-load speed requirement and 3 Nm can be achieved at 3,170 rpm. The BLDC motor and SRMs considered here have the same active stack length and outer diameters. Performance was evaluated at the same points as demonstrated on the SRM to provide some basis for comparison. Table 3 summarizes the points.

The most noticeable difference between the two motors is the required phase current to produce the required torque. The back EMF created by the permanent magnet limits the ability to shape phase currents. To improve high-speed performance, this BLDC motor is designed with a full slot skew that produces back EMF waveforms that are nearly sinusoidal. As a result, the inverter will excite the motor with 3-phase sinusoidal currents. Fig. 7 shows a sample of these phase currents. Fig. 8 shows the resulting torque.

Motor Selection and Power Electronics

The previous sections suggest that the SRM and the BLDC motors can provide comparable electromagnetic performance over the intended duty of the motor. However, differences in motor control and the cost of the electronics must be considered in cost-sensitive applications. It would appear at the outset that the BLDC has an advantage over the SRM because of the lower number of phases. Consideration of the currents that must be supported by the switches shows that the BLDC inverter would be significantly more expensive than the SRM inverter: the BLDC motor has six switches that must support 325A while the SRM has eight switches that must support 140A. Even taking into account the penalty for discrete diodes in the SRM inverter, the SRM inverter would require less silicon than the BLDC. This is before consideration of any one of a number of SRM inverter topologies that support operation with five controllable switches (the number of phases plus one). Inverter costs favor the SRM.

Manufacturing costs favor the SRM. The BLDC design is based on using rare earth NdFeB permanent magnets that are expensive relative to steel. Use of a lower quality magnet will exacerbate the current requirements. Performance of any magnet is a concern with the BLDC motor, particularly in an automotive environment where extreme ambient temperatures cause variation in magnet strength. The SRM uses a smaller air gap than the BLDC motor; however, the cost of the magnet compensates for this.

Relative to motor and inverter cost considerations, the SRM appears to have an advantage over the BLDC motor in this application. However, the SRM is not better than the BLDC motor in all applications. Each application demands careful matching of the electromechanical requirements to the inherent characteristics of the electric motor. The best match between motor and load characteristics will generally yield the lowest cost drive.

References

Miller, T. J. E., Brushless Permanent-Magnet and Reluctance Motor Drives, Clarendon Press, Oxford, 1989.
Miller, T. J. E., Switched Reluctance Motors and Their Control, Clarendon Press, Oxford, 1992.
Hendershot, J. R., and T. J. E. Miller, Design of Brushless Permanent-Magnet Motors, Clarendon Press, Oxford, 1994.
Hanselman, D. C., Brushless Permanent-Magnet Motor Design, McGraw-Hill, 1994.
Miller, T. J. E., and M. McGilp, “Nonlinear Theory of the Switched-Reluctance Motor for Rapid Computer-Aided Design,” IEE Proceedings, Vol. 137, Pt. B, pp. 337-347, 1990.
Torrey, D. A., X.-M. Niu, and E. J. Unkauf, “Analytic Modeling of Variable-Reluctance Machine Magnetization Characteristics,” IEE Proceedings-Electric Power Applications, Vol. 142, pp. 14-22, 1995.