论文部分内容阅读
An efficient design optimization process was developed and applied to an automotive lock-up clutch system with damper springs and centrifugal pendulum absorbers.Simulated annealing and FE-based optimization technique were coupled to determine the locations and parameters of the compressive helical damper springs on the clutch and to find the topology and the thickness of the clutch itself satisfying the strength, equations of pendulum motions, and torsional natural frequency requirements.The determination of a number, location, a number of turns, and deflection of damper springs play an important role in reducing vibration and noise of the lock-up system.A recent need for a thinner torque converter requires the optimization of a lock-up clutch system to fit a slender torus and to satisfy given working conditions while maximizing its cost effectiveness.A robust optimization for seeking a global minimum must adopt a strategy where a higher order of a function is acceptable under some conditions.Simulated annealing provides such a strategy.The presence of high powers of wire and coil diameters, e.g., terms involving d4, etc., results in difficulty for gradient methods.The method of simulated annealing is especially attractive on the spring design.The objective is to increase the factor of safety.Since the spring can fail by fatigue or by yielding, our objective is to maximize the smaller of the two safety factors for fatigue and in yielding.To determine the optimum wire and coil diameters such that alternating shear stress, mean shear stress, endurance limit, strength, and torsional natural frequency are satisfied.Topology optimization was applied to find out conceptual configures of retaining plates of lock-up clutches and the minimization of its vibration with design variables of a pendulum mass and a radius of rotation.Several optimum design candidates which can avoid resonance reduce vibration.Vibration tests were performed to validate the accuracy of the proposed process.The results show a good correlation between analyses and tests.