Experimental Verification of a Numerical Model of a Vertical Machining Centre

摘要

The Machine Technology Department has developed an optimization method ofrnmachine tool bodies relying on the combined techniques of the finite elementrnmethod and genetic algorithms. The method was applied in the optimization processrnof selected bodies of a newly designed milling machine centre. We can say that arnnew concept of designing of machine tool frames has been proposed because up tillrnnow, frames are designed using an analogy or an experimental method, simplifiedrnanalytical models (for example, a column is modeled as a beam) or making use ofrnexperienced designers only. Because of complexity of typical machine tool framesrnanalytical models are used very rarely. The use of numerical models of frames needsrnnot only experience in the design process but in CAE systems (FEM, CAD) as well.rnIn the literature there is little information about results of the optimization of framesrnof machine tools and there is no information about the results of experimentalrnverification of optimized frames for machine tools. For that reason an experimentalrnprototype of a new milling machine centre was made of metal and then verified inrnthe course of the scientific investigation. The paper presents a numerical model ofrntwo key bodies of a milling machine centre together with simulation andrnexperimental investigation results. In the case of stiffness indicator values, for whichrnthe displacement is measured during the experimental investigation amounted up tornseveral or more μm, the conformity of results was satisfactory, i.e. the results variedrnno more than 10-15%. Yet, for cases where stiffness indicators amounted up tornseveral thousand N/ μm the discrepancies were 100% and more. The direct reasonrnfor such significant differences was the high uncertainty of measurements related tornextremely minute displacements. The measuring process in a workshop environmentrnat the level of 1 - 2 μm with uncertainty less than 1 μm is technically difficult tornconduct thus entailing significant measuring errors.