Project Details
Abstract
With the development of production technique, grinding plays the increasingly
important role in mechanical processing. Among advanced industrial countries, the
annual yields of grinders has trended exceed the annual yields of lathe machines. At
present, the research focus of high speed grinding with intelligent functions has
inclined to the development of static and dynamic pressure with components adopted
from static hydraulic pressure bearings, air bearings and magnetic bearings, together
with other components like static-and-dynamic guides, linear guides and linear motors,
etc. To reach the requirement for spindles, guides and feed mechanisms with high
speed and excellent precision, other machine designs shall required for excellent
performances like anti-vibration and stability under various cutting parameters.
Therefore, if theoretically, it is available to create the whole-machine dynamic model
and predict dynamic performance, and the try-and-error effort can be reduced. It is
significantly important to enhance whole-machine dynamic performance, computing
efficiency and cost-down.
This research project is prepared in response to the 2-year plan prompted by
National Science Council focusing on precision external grinders. It is mainly aimed
to create the practical whole-machine dynamic model with lumped mass parameter
units and interface units, together with experimental validation and comparison. The
model compliance and energy distribution are adopted as theory foundation of
structure dynamic design and analysis is implemented to improve the weakness of
technique researches like the whole-machine structure and vibration control. Within
this research, there are two different surrogate model analysis skills used to find the
system design parameters and the optimal design parameters after additional
installation of an absorber. Therefore, the practically applicable analysis software can
be developed; meanwhile, the test is practically implemented or the features of
stiffness and damping of interface parts of whole-machine are gathered. The
applicable analysis database is created thereafter. After the plan is done, the data can
be served to grinding process practitioners with the important reference of analysis
and design evaluation available for preliminary stages.
The 1st year of this plan: the Lagrange energy method is firstly adopted to
deduce lumped parameter dynamic models of precision external grinders. It is aimed
to determine the inertia energy distribution rate for every sub-structure, the elastic
energy distribution rate for sub-structure interface and the model compliance between
the tool and work piece under various resonance frequencies to reach sufficiently
representative factors with significant influence on system dynamic performance. The
Response Surface Methodology (RSM) and Back Propagation Neural Network
(BPNN) are also separately used to construct two types of surrogate models for
structural dynamic performance to make comparison of predicting capabilities and
their variance. Finally, optimization analysis can be implemented to reach the optimal
parameters of system structure and the testing technique potentials for initial
components are built simultaneously. The 2nd year of this plan: It is aimed to explore the issue of vibration control
under the whole-machine cutting. Firstly, in view of energy, the locations most
suitably to install with an absorber can be found and the system dynamic equations for
the whole-machine additionally equipped with an absorber can be also well deduced.
It is aimed to determine optimal design parameters and locations of an absorber.
Finally, the analysis and experimental comparison are implemented on similar
machine tools and overall research results are well summarized thereafter.
Project IDs
Project ID:PB9611-0536
External Project ID:NSC96-2221-E182-063
External Project ID:NSC96-2221-E182-063
Status | Finished |
---|---|
Effective start/end date | 01/08/07 → 31/07/08 |
Keywords
- Precision External Grinder, Response Surface Methodology, BackPropagation Neural Network, Surrogate Model, Modal Compliance,Dynamic Absorber, Optimization Analysis
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.