Multi-criteria analysis of elastic couplings with metallic flexible membranes.
Dobre, Daniel ; Simion, Ionel
1. INTRODUCTION
These couplings use clamping packs of spoked membrane to combine an
adequate level of torque transmission with a reasonable axial and
angular misalignment capability.
In order to transmit the torque between two shafts of a kinematical
chain, in conditions of compensation of some important misalignments, it
is necessary for coupling producers to handle some multi-criteria
requirements regarding size, safety in exploitation (by guaranteeing a
superior mechanical strength), dynamic behaviour and constructive
simplicity.
The paper task is analysis of stress and buckling state of the
metallic membrane as component of the flexible couplings. This task is
carried out by a Finite Element Analysis (FEA).
2. ON STRESS STATE FOR MEMBRANE
The couplings having spoke form membranes transmit the torque from
the inner diameter to the outer diameter and reciprocally, by shear
loading of the membrane. The flexible membranes are made from spring
steel and have thin spokes by making radial indents in the central
portion (figure 1), the deformation of the spokes giving to the coupling
its flexibility and thus its ability to handle installation
misalignments. The couplings use these membranes arranged in a packet
and assembled rigidly at their inner and outer diameters to other
components. An example of this coupling is given in the fig. 2.
The mechanical strength analysis is made only for a single
membrane, for the critical situation of existing axial deviations,
taking into consideration that the stress as greater as in the case of
deviation absence.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
The stress state of the membrane spoke is determined by:
* torque transmission causing shear, bending and tensile stress;
* the angular and axial misalignments that causes bending and
tensile stress;
* centrifugal loading.
For the membrane study by FEA the meshing is realised with finite
elements of type SHELL, because the membrane depth is very small (0.3
mm), resulting 9054 nodes and 8316 elements. The meshing with the
maximum size of 0.5 mm on the side resulted after a number of elements
to ensure a sufficient convergence and accuracy of the FEA. The figure 1
shows the finite element mesh of the flexible membrane with trapezoidal
holes using a uniform distribution of the elements.
The analysis was made for the nominal torque T=50 Nm and the
rotation speed of 4500 rpm. A stress analysis was made for the cases
without deviation and with axial deviation of 0.6 mm in the absence and
with consideration of centrifugal load.
The tables 1 and 2 indicate the values for the principal stresses
[[sigma].sub.1] and [[sigma].sub.3] and the maximum equivalent stress
calculated using the von Mises theory [[sigma].sub.eq max].
One would see that the introduction of axial misalignment ([DELTA]a
= 0.6 mm) increases almost three times the principal stress
[[sigma].sub.1] and the maximal equivalent stress given the situation
without axial deviation.
The equivalent (von Mises) stress distribution in the nodes of the
structure is presented in figure 3 (with axial deviation of 0.6 mm) in
grey colours. One would see that the maximum stress is located to the
filleted root of the spokes. Other analysis at different torque
(non-presented in this paper) shows that the maximal stress appears also
on a large membrane area, that is the membrane solicitation is
increasing with the load.
The maximum stress, [[sigma].sub.eq max], serves for the calculus
of the safety factor, using the general expression:
c = [[sigma].sub.0.2]/[[sigma].sub.eq max], (1)
where [[sigma].sub.0.2] represent the yield point of the material.
The Romanian steel symbol used in calculus is OLC 65A, having
[[sigma].sub.0.2] = 800 N/[mm.sup.2]. The admissible safety factor
indicated in literature for pieces subjected at fatigue (Niemann, 1986)
is [c.sub.a] = 1.5 ... 3. The values of safety factor for the two cases
of stress analysis and the discussion regarding the checking at stress
solicitation are indicated in the table 3.
[FIGURE 3 OMITTED]
3. BUCKLING ANALYSIS BY FEA
Buckling analysis determines the loading level at which the
membrane becomes unstable. This analysis takes into account the effects
resulting from the compressive stresses, which tend to lessen the
capacity of the membrane to resist lateral loads. As the compressive
stresses increase, the resistance to lateral forces decreases, so that,
at some load level, this negative stress stiffening overcomes the linear
structural stiffness, causing the structure to buckle (Sorohan &
Sandu 1997).
The linear buckling analysis determines the scaling factors,
[c.sub.f], for the stress stiffness matrix, which offset the structural
stiffness matrix. The equation that describes linear buckling is:
([K] - [c.sub.f] x [S]){u} = 0 (2)
where: [K] is the structural stiffness matrix; [S]-the stress
stiffness matrix; [c.sub.f]-the buckling load factor; {u}-the
eigenvector representing the buckled shape.
The buckling load factor is a safety factor, defined as:
[c.sub.f] = Limit buckling load/Applied load (3)
For the buckling analysis the membrane was blocked both on its
inner and outer holes as displacement on the direction of the membrane
axis with considering the centrifugal loading.
The buckling load factors for 10 shape modes has values in the
interval [c.sub.f] [member of] [11.767, 17.625], being greater than the
ones admissible conforming to Niemann (1986), [c.sub.a] = 3 ... 5. This
means that the membrane is stable at buckling.
The figure 4 shows an image of the second buckling shape. The
buckling load factor, [c.sub.f], is noted FREQ, a denomination used in
the ANSYS 6.1 program.
[FIGURE 4 OMITTED]
4. MODAL ANALYSIS OF THE MEMBRANE
The modal analysis is necessary to extract the natural frequencies
and mode shapes of a flexible membrane. A membrane should be designed to
produce natural frequencies that will prevent the coupling component
from vibrating at one of its fundamental modes under operating
conditions.
The dynamic torsional stiffness determines the torsional vibration
behaviour of the drive system. This gives rice to very large natural
frequencies and the operating speeds lie below the resonant speeds. The
natural frequencies for 10 vibration modes have values in the interval f
[member of] [2013, 2577]. This means that the membrane is stable at
vibrations.
It is relevant to state that the increased of dynamic torsional
stiffness causes an increase in the resonant speeds (Birsan &
Jascanu 1998). This is important because the resonant speeds should be
present above the operating speed range.
5. CONCLUSION
The finite element analysis (FEA) is basic to evaluate stress and
buckling state of flexible membrane.
The areas of maximum equivalent stress are placed on fillet root of
the membrane spokes. The maximum equivalent stress values increase with
the axial deviation at the limit of the recommended admissible safety
factor values.
The centrifugal loading has a significant stiffness effect which
causes law increasing of the buckling safety factors, leading to the
increased resistance to torsional deformations and to deformations
caused by the misalignments of the two shafts.
The results are important for other researches for the case of the
entire membrane packet and are in concordance with the experimental
dates presented by Dobre (2004).
6. REFERENCES
Birsan, J.G. & Jascanu, M. (1998). Elastic couplings'
dynamic, Tehnica Publishing House, ISBN 973-31-1232-1, Bucharest
Dobre, D. (2004). Researches on multi-criteria optimization of
elastic couplings with metallic flexible membranes, Ph.D. Thesis,
University "Politehnica" from Bucharest
Niemann, G. (1986) Maschineelemente, Springer Verlag, Berlin
Sorohan, S. & Sandu, M.A. (1997). Nonlinear FEM-Analysis of a
Diaphragm Spring, ELFIN 4, Iasi, oct. 9-11, pp. 125-128.
Tab. 1. The results of stress analysis for the case without
deviation ([DELTA]a=0), by considering the inertial loading.
Parameter
Principal stress [[sigma].sub.1] [N/[mm.sup.2]] 71.8
Principal stress [[sigma].sub.3] [N/[mm.sup.2]] -65.2
Maximum equivalent stress [[sigma].sub.eq max] [N/[mm.sup.2]] 69.7
Tab. 2. The results of stress analysis for the case with axial
deviation ([DELTA]a = 0.6 mm).
Parameter
Principal stress [[sigma].sub.1] [N/[mm.sup.2]] 286
Maximum equivalent stress [[sigma].sub.eq max] [N/[mm.sup.2]] 255
Tab. 3. The values of membrane stress safety factors.
Analysis case
Size Without With axial
deviation deviation
of 0.6 mm
Maximum equivalent 69.7 255
stress
[[sigma].sub.eq max]
[N/[mm.sup.2]]
Safety factor, c 11.48 3.14
Checking discussion c > [c.sub.a] c [congruent to] [c.sub.a]