Page 59 - Glass Machinery Plants & Accessories no. 2-2020

Fig. 6 Simulated contact forces Fig. 7 Work-pieces a) before and b) after real
smooth finishing

the Cowper-Symond multiplier, introduced in the LS-DYNA com- simulation. This is an important
is shown in Fig. 3b for different mand, file and then the LS-DYNA topic in the work of material tests
effective plastic strain rates. The solver was started. The complete and constitutive equation formula-
selection of the best material mod- solution was obtained. The result tion.
els was made by the comparison of analysis accentuates the impor- The simulated working regimes
Brinell test simulation results with tance of numerical simulation on were performed with increased
experimental data. Fig. 4 shows smooth finishing processes. The longitudinal advancing. This fact
the partial model of the simu- surface quality may be interpret- generated some non-uniformity on
lation and effective plastic strain ed on the effective plastic strain the smoothed surfaces.
fields in the remanent stage for field as shown in Fig. 5. Field The numerical simulation of
both used materials. For Johnson- homogeneity and intensity indi- the smooth finishing process was
Cook materials model the values cate an improvement of the sur- validated experimentally using a
of Brinell simulated hardness (122 face quality by hardening. Surface special device. Fig. 7 gives the
and 140) are comparable with roughness is evaluated by means results of the experimental smooth
the experimental data values (118 of a special procedure using nodal finishing process applied on cylin-
and 140) averaged on each mate- coordinates in the final stage. The drical work-pieces. A new quality
rial lot. This method is very cor- residual stress field, represented in is visible on the work-piece after
rect for material model validation, the same figure, highlights one of processing.
especially in predominant com- the disadvantages of this process- The lamellar graphite in the
pressive plastic deformations. The ing. The contact forces, as time material structure has a negative
Johnson-Cook model is used in the functions, are shown in Fig. 6. The influence on the plastic deforma-
following simulations. final plastic strain field (Fig. 5) and tion, by limiting quality.
force variations (Fig. 6) highlight The numerical simulation offers
that a better control on the radial important possibilities for design,
SIMULATION
The finite element discretization advance is necessary for deforma- analysis and optimizing of work-
of the physical model (Fig. 1) was tion uniformity. ing regimes for smooth finishing
performed with the variable density processing. Q
mesh by using 3D-hexaedral ele- CONCLUSIONS
ments. The superficial layer of the In the work, a numerical
work-piece was meshed with high simulation methodology for the
density (Fig. 5). The various simu- smooth finishing processing of the
lation work regimes were estab- glass mould tools, based on the
lished within the following parame- LS-DYNA finite element codes,
ter limits: spindle speed: 384…920 was formulated. Special attention FABRICA DE STICLA
rpm; longitudinal advance: was conferred to model prepara- DIN CHISINAU
0.09…0.32 mm/rot; depth: tion, because if the physical model
0.05…0.2 mm. Random rough- is not correctly and completely 20 Transnistria street
ness was established in accordance designed, neither finite element MD-2037 Chisinau - Moldavia
with the middle turning opera- code thoroughness, nor computer Tel: +373 22 47 39 26
tion (Ra= 3.2…6.3 μm). For each capabilities vouch for the success E-mail:office@glass.md
established regime, input data was of solution accuracy and numerical www.glass.md

glass machinery plants & accessories 2/2020 57