Semi-active control of torsional vibration can be realized through the use of variable friction brakes or clutches applied to a primary system acted on by oscillating torques. The performance of a given vibration control approach will depend greatly on the bandwidth of the actuators used to realize control. Three commercially available torsional actuators, a dry friction brake, a magneto-rheological fluid brake, and a magnetic particle brake have been tested and analyzed to assess their applicability for use in semi-active torsional vibration control. A test stand was constructed and used to run specific tests including step responses to determine "on" and "off" response times, open-loop bandwidth determination via swept-sine tests, and friction as a function of rotating speed. The data can be used to create general mathematical models to predict the behavior of the different actuators when excited with different control signals. The results indicate some of the limitations of the different actuators and will be used to provide a basis for determining the actuators' applicability to general torsional vibration control problems.
Analytical and experimental work has been performed on the use of a magneto-rheological (MR) fluid brake for controlling torsional vibrations in rotating systems. In this paper, two different strategies are examined for controlling the MR brake in such applications. First, implementation of the MR fluid brake as a passive friction damper with a variable friction torque is presented. In that application, fixed currents were applied to the electromagnets in the MR brake. As a result, the dominant behavior of the brake was as a dry friction damper whose friction was a function of applied current. The second approach was the implementation of the MR brake in a modified skyhook damping control approach. In that application, the friction in the MR brake was adjusted according to a comparison between the sign of absolute velocity of the primary system and the sign of the relative velocity between the MR brake and the primary system in order to add effective damping to the system. Characterization of the MR brake was an essential part of both control strategies. This work includes the results of system identification performed on the MR fluid brake, along with experimental performance results of the system under the different control strategies.
Crankshaft dampers are a common approach for controlling engine crankshaft vibration. The optimum damper parameters are relatively easy to determine for the case of single-mode systems and multi-mode systems with a dominant mode, provided that the primary system is undamped and the system response is linear. For nonlinear systems such as internal combustion engines that experience complex periodic inputs, the true optimum damper parameters may not be apparent. The crank kinematics introduce nonlinear torques acting on the crankshaft. In addition, the gas torque is, in some sense, a state-dependent input, as it is a function of not only the energy addition per cycle, but also of the crank angle. It is reasonable to expect that truly optimal damper parameters may not be obtained using classical approaches. As an alternative, genetic algorithms may be used to determine optimum crankshaft damper settings for this complex system. This paper will present the modeling of an internal combustion engine from the perspective of determining crankshaft vibrations. Optimum damper settings are then determined using a genetic algorithm. Simulation results are shown that compare the achievable vibration reduction in an engine equipped with a GA-tuned damper and the reduction achieved with a conventional passive damper.
Control of torsional vibrations in an automotive crankshaft is a classical vibration control problem. The most common solution is to mount a crankshaft damper at one end of the crankshaft. Typical crankshaft dampers are composed of parallel stiffness and damping elements connecting a rotational inertia to the crankshaft. Appropriate design of the damper elements may result in substantial crankshaft vibration. Conventional couplings include elastomeric spring-damper elements and purely viscous fluid couplings. While those approaches result in satisfactory reduction of crankshaft vibration, it may be that a semi-active approach can achieve improved performance. To that end, an investigation of a semi-active crankshaft damper using magneto-rheological (MR) fluid has been initiated. A torsional MR fluid brake was obtained and applied to a scale model of a crankshaft for a common eight-cylinder engine. Experiments were performed with the MR brake as a fixed-friction device. In addition, a simple stick-slip control algorithm was developed such that the MR brake became an on-line variable friction device. While a good deal of work remains to be performed in future efforts, the preliminary experimental results have demonstrated that a torsional damper composed of an MR fluid brake has potential application in the field of torsional vibration control.