M E C H A N I C A L     E N G I N E E R I N G

 

DK LIEU

ELECTRO-MECHANICS

UC Berkeley Electro-mechanical Design Laboratory

The UC Berkeley Electro-mechanical Design Laboratory is a member of the Pacific Rim Electro-Mechanical (PREM) Laboratory Research Consortium. The Laboratory is dedicated to the design and analysis of novel electrical motors and actuators. Established in 1988, it is a 1800 sq. ft facility located in room 2163 Etcheverry Hall. In addition to student offices and laboratory areas, the laboratory has developed unique facilities and capabilities for its work. Modeling and simulation (ANSYS for finite-element magnetics modeling) is performed extensively on Pentium based workstations. Pentium based personal computers are used for less intensive computation. Test equipment include structural dynamics analyzers, sound level analyzers, magnetic flux and gauss meters, piezo-electric accelerometers, and a Vibrac motor test stand. A TSI laser Doppler vibrometer is used for dynamics measurements in the presence of large magnetic fields. Most of the research effort during the past several years has been concentrated on the design, analysis, and optimization of novel, highly stable DC brushless motors and actuators. Close industrial ties are maintained with IBM, Quantum, Adept, Encap, Samsung, and Daewoo.

     

Direct-drive devices using brushless DC motors are of increasing interest because of their simplicity, quick response time, and control linearity. Typical automotive applications include alternators and accessory motors, as well as possibly the main engine in electric automobiles. Other applications include precision spindles for robotics, machine tools, and computer disk files. The heart of these machines is a moving magnet or coil device, such as a D.C. brushless motor with a rotating complement of magnets. New high energy magnetic materials, such as neodymiun-iron-borons, will soon realize high performance drivers which are smaller and lighter than their predecessors. The development of new applications and products has shown that such devices should no longer be considered as an "off the shelf", purchased part around which the remainder of the system must be designed. The actuator for a robot arm, for example, may be designed to be contained within the arm itself, rather than attached externally through a gearbox. This approach would offer many advantages for both system dynamics (reduced inertia and backlash, increased control linearity) and packaging.

Custom actuators, especially those which use rare earth magnets, have proven to present some unique design problem. The quality of most electrically driven devices depends on the development of the structural stability of the devices. Instability due to the presence of excitation from the rotating magnetic and electro-magnetic fields within the motor results in transmitted vibration and acoustic noise. The magnets and windings on a DC brushless motor, for example, must be considered as part of the structure of the entire motor, and integrated by design into that structure. The difficulty is that electro-magnetic dynamics, and not windage nor mechanical imbalance, is often the largest source of noise and vibration in the motor. Elimination of the instability requires either modification of the structure or of the excitation force. In an automobile alternator, for example, the majority of the acoustic noise that is generated disappears when the rotor field is not energized. However, this is impractical since any modification must be done without altering the primary function of the device.

The high attractive force, proportional to the magnetic flux density squared, between a moving magnet and its surrounding structure creates an instability in that structure. This instability, usually ignored in low flux density devices, limits the design in many high flux machines. In disk file spindle motors, a design with neodymium magnets can be 50% smaller and lighter that an equivalent output ferrite magnet design which has 1/3 the flux density. However, the higher attractive force of the neodymium is an excellent driver of the motor structure. This problem would be most severe when any harmonic of the driving frequencies match a resonant frequency of the motor or surrounding structures. Vibration levels have been shown to be affected by shaping of the magnetic field through shaping of the magnets or magnetic structure. The precise source of acoustic noise generation in motors generally is not known, but has shown to be affected by flux density levels in a motor, the shape of its magnets, its terminal impedance, and method of speed or torque control. A clear understanding of the design of both existing and custom electro-mechanical devices, and their inherent problems, is essential for a successful mechatronics program.

Among the novel approaches to motor design that have been developed at the laboratory is geometric shaping of the magnetic poles to reduce vibration without altering the output torque. The magnetic poles can be optimally shaped to either reduce the vibration driving spectrum across a band of frequencies, or to totally eliminate a single driving frequency. Another approach that is under development is altering the level of and direction of magnetization at the pole transitions. Extensive modeling efforts are made in an attempt to characterize the distributed mechanical pressures induced by the magnetic fields, and the resultant structural reaction on the motor. Using this approach, the laboratory also has ongoing projects in reduction of motor reluctance torque and eddy current generated vibration, and acoustic noise and vibration reduction by mechanical damping. In an integrated mechatronics program, it would be possible to effect further improvement of motor performance using an active control through the motor driver.

TOP

OFFICE

 
LAB

Office Hours: TTH 10-12 and by appointment

dlieu[at]me.berkeley.edu

5128 Etcheverry Hall  phone: (510) 642-4014     fax: (510) 642-5599

2116 Etcheverry Hall Berkeley, CA, 94720-1740 (510) 643-5281

 

 
If you have any questions or comments, please e-mail the webmistress at mlieu[at]berkeley.edu