Академический Документы
Профессиональный Документы
Культура Документы
DOI: 10.2316/P.2012.785-012 6
experimental boundary conditions close to that of a
clamped plate. Table 1: Stiffened plate specimen dimensions in mm
In the present study, static deflection of a clamped Sl Plate
L Lc tp Ls ts bs
stiffened plate subjected to uniform transverse pressure is No type
determined experimentally. Loaded natural frequencies of 1 US 480 40 2 - - -
the plate under uniformly distributed transverse load are 2 UASS 480 40 1.5 400 25 3
also experimentally found out. Uniform loading of the
3 UASS 480 40 2 400 25 3
specimen is imparted pneumatically by supplying
compressed air into a sealed box with the specimen 4 UASS 480 40 2 400 25 5
rigidly attached to it. 5 BACS 480 40 2 400 25 3
6 BACS 480 40 2 400 25 5
2. Experimentation Description
Figure 1 shows the schematic diagrams with
This section provides a detailed description of the notations for significant dimensions of the various types
specimens, the experimental set up on which the tests are of plates under consideration. Dimensions of the
performed and the test procedure followed. specimen and their nomenclatures are mentioned in Table
1. Photograph of these specimens are also presented in
2.1 Specimen Description Figure 2. Along the circumference of the plates specific
number of holes of a fixed dimension are drilled for
The specimens used in the experiment are made up of bolting the specimen to the set up. The dimensions and
square plates of uniform thickness and flat bars of the spacing of the holes correspond to those present in the
rectangular cross-section. Material for both the plates and upper flange of the set up. Certain amount of width (Lc)
the bars is mild steel. The bars are welded to one side of around the periphery of the plate on all sides is taken up
the square plates to prepare different types of stiffened by the frame used to bolt the plate securely. This implies
plates. The orientation of the stiffeners are parallel to the that the actual dimension of the clamped plate is Ls = (L -
sides of the plate specimen. Two types of stiffened plates 2Lc).
are analyzed in the present scenario and these are uni-
axially single stiffened (UASS) and bi-axially cross
stiffened (BACS) plates. In both the cases, stiffeners are
attached at the middle of the plate. Experiments are also
performed on an unstiffened (US) plate as well.
(a) (b)
(c)
Figure 2. Photographs of (a) unstiffened (b) Uni-axially
single stiffened and (d) Bi-axially cross stiffened plate.
7
plates to ensure that there is no leakage. The frame has an
inlet through which air is introduced into the enclosed box
for application of uniform transverse load. The
pressurized air is supplied by an air compressor. There is
provision for attaching a pressure gauge at the inlet
assembly, which also has a bypass valve to manipulate the
level of pressure inside the enclosed chamber. Two dial
gauges are set at predefined locations (A and B, as shown
in Figure 4) for measurement of deflection under
transverse loading. A fixed arm bolted at the two opposite
edges of the frame is utilized to position the dial gauges at
their desired locations.
A shear mode accelerometer is mounted on the
vibrating system at a suitable position using Petro-Wax
adhesive material. The location of the accelerometer is
chosen carefully to avoid any nodal points and it is
approximately placed near the point of maximum
displacement amplitude. However, the location of
maximum displacement varies depending on the type of
specimen. The accelerometer senses motion perpendicular
to the base and it has two components - a sensing
assembly and an impedence converter. The sensing
assembly consists of a seismic mass and a piezoelectric
crystal. The impedence converter is a miniature electronic
circuit which converts the high impedence charge signal
generated by the piezoelectric element into a low
impedence voltage output signal. This circuit is powered
by an external power source or coupler. Since the
dynamic mass of the accelerometer (1.6 grams) is
significantly less than the test specimen, it can be
assumed that the response is not affected by the mass
loading.
8
The accelerometer is connected to a coupler, which domain (frequency-amplitude plane) by the oscilloscope.
provides the constant current power supply to the The data is then transferred to the computer hard disk
impedence converter circuit of the accelerometer. The through the RS-232 ports and saved under an appropriate
coupler itself is externally powered by an AC-DC power file name and format. Using vertical cursors the frequency
adapter. The instrument also decouples the DC bias (in Hz) of the signal are read from the display and the data
voltage from the output signal. The connection between is tabulated in an excel sheet in the computer. Now the
the accelerometer and the coupler is through a two-wire dial gauge probes are lowered to touch the plate surface
cable, which carries both the signal and power. again and the pressure value is changed. At the new
A digital storage oscilloscope is used to display, pressure level the described procedure is repeated to
process and store the signal transmitted by the vibrating collect data corresponding to deflection and loaded
system. It has the capability to transform a time domain natural frequencies of the plate. The whole experimental
signal into frequency domain through a Fast Fourier procedure is replicated for different stiffened plate
Transform (FFT) extension module. The oscilloscope also specimens.
has RS-232 port compatibility through which it can be
connected to a desktop computer. The computer is used as 3. Theoretical Analysis
the storage and offline processing device for the data
acquired from the oscilloscope. The captured data is sent The theoretical static and free vibration analysis is
to the computer through RS-232 communication ports and performed using a semi-analytical method based on
saved into the computer hard disk using WSTRO variational principles.
Wavestar software in a standard format (.sht) provided by
the software. The software also has the capability to 3.1 Static Analysis
process the saved data in offline mode and convert it to
sequential data format for further processing. The static analysis is formulated through principle of
minimum potential energy which can be mathematically
2.3 Test Procedure expressed as,
(U V ) 0 (1)
The set up is readied by tightly bolting the specimen
where, U is total strain energy stored in the system, V is
under consideration to the frame so that the enclosed
work function or potential of the external forces and is
chamber is leak proof. Two dial gauges are positioned
the variational operator. The above mentioned energy
over their predefined locations with the help of a fixed
functionals (U and V) can be determined from the
arm supported over two posts bolted to the edge of the
assumed plate displacement fields (w, u and v), which are
frame. A flexible pipe is used to connect the compressor
depicted as linear combinations of orthogonal functions
outlet to the inlet of the air chamber and pressure gauge is
and unknown coefficients (di ) as shown below.
also fitted. Extreme care is taken in order to ensure that
nw nw nu
there are no leakages from the various joints and the pipe.
Electrical connections for compressor, computer,
w ,
i 1
d i i , , u , i nw1
di i nw ,
nw nu nv
oscilloscope, coupler and accelerometer are made to ready
the set up. The accelerometer is mounted on the test
and v , i nw nu 1
d i i nw nu , , where, ( , ) ,
specimen at a predefined location using adhesive and is ( , ) and ( , ) are sets of nw, nu and nv numbers of
connected to the coupler by a two-wire cable. Here, the orthogonal functions for w, u and v, respectively. The
signal from the accelerometer and power for its functions i ( , ) describe the displacements due to plate
impedence circuit share the same cable. Output from the bending and the base function for this comes from the
coupler is connected to one of the channels of the flexural boundary condition. The start functions for
oscilloscope. The oscilloscope is set to ‘math’ mode and stretching of the plate (u and v) come from the in-plane
‘auto’ trigger mode is kept on by pressing the appropriate boundary conditions. The higher-order functions are
buttons. It is then kept ready by pressing the ‘RUN’ generated from the selected start functions following a
button. A RS-232 port compatible cable is used to connect two-dimensional implementation of the Gram-Schmidt
the oscilloscope to the desktop computer. orthogonalization scheme. Substituting the expressions
The experiment for static deflection is carried out for strain energy and work function along with the
first by increasing the pressure gradually. A brief settling assumed displacement fields the set of governing
period is allowed before the readings of the two dial equations for the static problem is obtained in the form
gauges are noted. At this stage the probes of the dial given below,
gauges are carefully lifted and are kept at such a height K d f (2)
with the help of plastic stops so that they can not come in
where, [K] and {f} are stiffness matrix and load vector,
contact with the vibrating plate. This is done in order to
respectively. The set of nonlinear equations are solved
let the plate vibrate freely. At this deflected configuration,
through a direct substitution method with successive
the plate is struck with a rubber hammer to provide the
relaxation technique, which has been discussed in
disturbance to generate vibrating pulses. The signal from
previous publications presenting large deflection static
the vibrating system is captured and plotted in frequency
and free vibration analysis of stiffened plate at deflected
9
configuration [13,14]. The solution of the static analysis 4.1 Comparison with Theoretical Results: Static
provides the statically imposed large displacement fields Deflection
of the system, which serves as the starting point for the
dynamic analysis.
10
Comparison of experimentally obtained normalized load-
deflection curves with those generated numerically for
unstiffened, UASS and BACS plates are presented in
Figures 5, 6 and 7, respectively. In these figures the
ordinate represents normalized displacement (w/tp) at
point A or B, whichever is greater. The abscissa of the
figures is normalized pressure or load, given by
pa 4 E p t 4p corresponding to uniformly distributed
loading.
The figures show that the experimental results are
significantly higher than the numerical results in all the
cases. However, the nature of the curves is quite similar to
those predicted by the theoretical analysis. The variation
between the experimental and theoretical load-deflection
curve may be due to the fact that the membrane or in-
plane boundary conditions are not fully satisfied. As the (b)
plate is bolted to ensure clamped end condition, there can Figure 8. Comparison of experimentally obtained
be no transverse displacement. However, the same can not normalized load-deflection curve with theoretical analysis
be said about the axial displacement. The bolting pressure considering immovable and movable in-plane boundary
between the plate and the frame generate friction force, condition for (a) Unstiffened and (b) UASS plate.
which acts against axial movements of the plate. The
force may be sufficient to arrest in-plane displacement at 4.2 Comparison with Theoretical Results: Loaded
the edges at low pressure, but at high transverse pressures Natural Frequency
the edges become movable. It is noted that the difference
between the experimental and theoretical results increases
with increase in pressure, which also indicates that at
higher pressure the stretching forces come into effect.
Theoretical results for unstiffened and one of the
UASS plates (tp 0.002 m, bs = 0.005 m) is generated for
movable in-plane boundary condition and compared with
the experimental and numerical results for immovable
boundaries in Figures 8(a) and 8(b) respectively. It is
observed that the results for unstiffened plate with
movable membrane boundary condition is very close to
that of experimental results. However, there is little
difference between the two sets of data for UASS plate. In
this case (Figure 8(b)) the load-deflection curve for
movable boundaries is closer to that of immovable
boundaries than the experimental one.
Figure 9. Comparison of backbone curve for unstiffened
plate.
11
when the external pressure is low. But the two curves For all the backbone curves, the fundamental linear
move apart from each other as the pressure is increased. It frequency ( 1 ) is used as the normalizing parameter.
is also evident from the figures that the degree of Table 3 shows comparison between the experimentally
nonlinearity induced for the different types of stiffened obtained and programme generated natural frequency for
plate is different. the fundamental mode of various unstiffened and stiffened
plates. From the table it is clear that there is satisfactory
agreement between the two sets of results for unstiffened,
UASS and BACS plates.
5. Conclusion
In the present paper, experimental investigation of static
and free vibration behaviour of stiffened plates under
transverse pressure loading has been dealt with. A set up
simulating clamped boundary conditions at all edges of
the plate is used to perform the experiments. The set up
has provision for supplying pressurized air to an enclosed
chamber for applying uniform transverse loading to the
specimen. The objective of the static experiments is to
Figure 11. Comparison of backbone curve for BACS determine the deflection of the plate under external
plate. transverse loading, while the loaded natural frequencies
12
are found out from the free vibration experiments. The References
results reveal hardening type nonlinearity for the system.
The experimental results for both static and free vibration [1] F.J. Fahy, & R.B.S. Wee, Some Experiments with
studies are compared with those generated by theoretical Stiffened Plates under Acoustic Excitation, Journal of
analysis. Although the experimental results exhibit similar Sound and Vibration, 7(3), 1968, 431-436.
nature as those generated by numerical analyses, the [2] R.N. Yurkovich, J.H. Schmidt, & A.R. Zak, Dynamic
actual values are found to be higher than theoretical ones. Analysis of Stiffened Panel Structures, Journal of
This may be attributed to insufficiency in replicating Aircraft, 8, 1971, 149-155.
stretching boundary conditions of the system. [3] M.D. Olson, & C.R. Hazell, Vibration Studies on
Some Integral Rib-Stiffened Plates, Journal of Sound and
Vibration, 50(1), 1977, 43-61.
[4] M.N. Bapu Rao, P. Guruswamy, M.V. Rao, & S.
Pavithran, Studies on Vibration of Some Rib-Stiffened
Cantilever Plates, Journal of Sound and Vibration, 57(3),
1978, 389-402.
[5] K. Ghavami, Experimental Study of Stiffened Plates
in Compression up to Collapse, Journal of Constructional
Steel Research, 28(2), 1994, 197-222.
[6] N.K. Bau-Madsen, K.H. Svendsen, & A. Kildegaard,
Large Deflections of Sandwich plates - An Experimental
Investigation, Composite Structures, 23(1), 1993, 47-52.
[7] J.F. Wilson, J.K. Henry, & R.L. Clark, Measured Free
(a) Vibrations of Partially Clamped, Square Plates, Journal of
Sound and Vibration, 231(5), 2000, 1311-1320.
[8] L. Shuyu, Study on the Flexural Vibration of
Rectangular Thin Plates with Free Boundary Conditions,
Journal of Sound and Vibration, 239(5), 2001, 1063-
1071.
[9] N. Yamaki, K. Otomo, & M. Chiba, Nonlinear
Vibrations of A Clamped Rectangular Plate with Initial
Deflection and Initial Edge Displacement - Part II:
Experiment, Thin-Walled Structures, 1(2), 1983, 101-119.
[10] M. Amabili, Nonlinear Vibrations of Rectangular
Plates with Different Boundary Conditions: Theory and
Experiments, Computers and Structures, 82(31-32), 2004,
2587-2605.
[11] M. Amabili, Theory and Experiments for Large-
(b) Amplitude Vibrations of Rectangular Plates with
Geometric Imperfections, Journal of Sound and
Vibration, 291(3-5), 2006, 539-565.
[12] M. Amabili, & S. Carra, Experiments and
Simulations for Large-Amplitude Vibrations of
Rectangular Plates Carrying Concentrated Masses,
Journal of Sound and Vibration, 331(1), 2012, 155-166.
[13] A. Mitra, P. Sahoo, & K.N. Saha, Large Deflection
Static Analysis of Stiffened Plates through Energy
Method, Proc. of 3rd International Conf. on Advances in
Mechanical Engineering, SVNIT, Surat, India, 2010, 392-
397.
[14] A. Mitra, P. Sahoo, & K.N. Saha, Free Vibration
Analysis of Initially Deflected Stiffened Plates for
(c) Various Boundary Conditions, Journal of Vibration and
Figure 12. Variation of frequency-amplitude signal Control, 17(14), 2011, 2131-2157.
generated from the vibrating specimen during free
vibration experiment with external pressure for BACS
plate. (a) Waterfall diagram, (b) Mesh plot and (c)
Contour plot.
13