Modeling and Identification of Joint Dynamics Using a Frequency-Based Method.

摘要

There is an ever increasing demand for more productivity along with improved accuracy of goods produced by manufacturing technologies. Traditionally, physical prototypes were tested and changed in order to improve productivity and obtain the optimal operating conditions, which imposed a great cost on manufacturers. Nowadays, virtual prototyping technology is being employed to aid in eliminating costs associated with iterative testing and development processes. Virtual prototypes facilitate the implementation of simulations, predictions and optimizations based on the kinematics and dynamics of a machine tool structure, all within a virtual environment. The creation of such an environment, however, is not a simple endeavor.;Building an accurate virtual model requires thorough knowledge of all constituent elements of the physical structure, including the joints. Joints play an important role in the overall dynamics of assembled structures; as much of flexibility and damping in the structures are originated at the joints. Ignoring joint effects and modeling the joints as rigid connections result in deviations between the physical structure dynamics and model dynamics. In order to improve accuracy of model predictions, joint dynamic properties need to be identified and incorporated into the virtual model. This will allow for a higher fidelity representation of the real physical system.;Joints are usually complex in geometry and often inaccessible in the assembled structure, making it difficult for their direct measurements and mathematical modeling. In order to accurately identify joint dynamics, this study aims at identification of joint dynamics using a frequency-based method. The overall essence of joint identifications through the frequency-based approach is the determination of the difference between the measured overall dynamics and the rigidly coupled substructure dynamics.;The inverse receptance coupling (IRC) method is introduced as the primary identification technique used in this study. Applications of the IRC method in 2-dimensional (2D) structures is examined on two physical structures: a lathe machine and a vertical computer numerical control (CNC) machining centre. On the lathe machine, the joint dynamics of a modular tool are obtained; and, on the CNC machine, the joint dynamics at the tool / tool-holder / spindle interfaces are obtained. The joint dynamics at these locations have shown significant effects on the overall dynamics of the assembled structure. An extension to the IRC method is also proposed to account for the effects of multiple joints in structures.;The IRC method is also extended to 3-dimensional (3D) structures. A complete joint model which accounts for the effects of joint's inertial properties is developed and validated through finite element (FE) simulation. Experimental tests on a mock test setup of a vertical CNC machine are performed to assess applicability of the proposed identification method in actual 3D structures.;The results of this study can be used in constructing a database for various types of joints in machine tool centers as a function of influential factors on the joint dynamics such as preload, material and surface contact. Such a database can then be used in the design stage to improve the correlation between predictions made by the virtual model and the behaviour of the physical structure.