External Six-Component Strain Gauge Balance For Low Speed Wind Tunnels
External Six-Component Strain Gauge Balance For Low Speed Wind Tunnels
40-46
UDK: 533.6.07(047)=20
COSATI: 01-01
Marija Samardžić1)
Zoran Anastasijević1)
Dragan Marinkovski1)
Dušan Ćurčić1)
Jovan Isaković2)
The external six-component wind tunnel balance designed and manufactured in the Military Technical Institute is
described in this paper. The balance is of a platform type. Aerodynamic load which acts on a model in the wind tunnel
test section is transferred into the set of six load cells. The relations between the aerodynamic load and the forces
measured using the balance load cells are presented. All load cells are gauged with foil strain gauges. The results of the
balance calibration are shown. The balance calibration was done using dead weights.
Key words: wind tunnel, strain gauge balance, load cell, calibration.
1)
Military Technical Institute (VTI), Ratka Resanovića 1, 11132 Belgrade, SERBIA
2)
College of Applied Engineering Studies, Tehnikum Taurunum, Nade Dimić 4, 11080 Belgrade, SERBIA
SAMARDŽIĆ,M. etc.: EXTERNAL SIX-COMPONENT STRAIN GAUGE BALANCE FOR LOW SPEED WIND TUNNELS 41
resistance. The experience in designing strain gauge consists of the manual application of dead weights.
balances in the VTI showed that this stainless steel is one of Calibration increments of dead weights are applied to the
the best choices of material. The selected material has a balance at precise locations relative to the balance reference
strong influence on the hysteresis, creep and repeatability of centre. This process is used for both positive and negative
the force transducers. loadings at various stations. The selection of calibration
Bending stress and strain at the strain gauges location loads is one of the most important elements in the
are: calibration process. Ideally, a balance should be calibrated
over only those loads expected during a specific wind
1⋅F ⋅L tunnel test program. For the most calibration programs,
t
σb = 2 combined load effects must also be developed.
b ⋅ h2 , (2) The balance calibration and the wind tunnel test balance
6 data reduction algorithms used in the VTI are a
σb
ε= generalization developed upon the concept presented by
E Galway [23]. This method was further developed and lately
Where: σ b is the bending stress, ε is the strain and E is the it has become known as the "Single-Vector Force
modulus of elasticity. Calibration Method" [24]. The generalized calibration
The value of the strain in the horizontal load cells is: method developed in the VTI permits more than one
composite load vector to be applied [25,26].
μm Fig.6 shows the set up for the calibration of the six-
ε1 = 966.7 ⋅10−6 , (3) component balance. The calibration body was mounted on
m
the supporting struts, replacing a wind tunnel model. The
while in the vertical load cells it is: calibration body was made from two crossed beams on
μm which the measured load application points were positioned
ε 2 = 920.5 ⋅10−6 . (4) very accurately. During the calibration, each component
m
was loaded at least 5 increments from zero to the maximum
The selection of strain gauges was done in accordance load. The coefficients of the calibration matrix were
with the values of the measured strain and the dimensions calculated for each component by the least-squares method
of the measuring sections. The gauge selection process using the data from the complete representative load subset.
consists of determining the particular available combination
of parameters which is most compatible with the
environmental and other operating conditions and at the
same time best satisfies the balance and operating
constrains. The process of gauge selection generally
involves compromises. The selection of the optimum strain
gauges is an iterative process, moving between the
requirements of the load cell and the characteristics of the
available strain gauges. The selected strain gauges for all
load cells are TK-06-S082R-350 [21].
The optimum strain gauge excitation level depends on
the size and type of the strain gauge. Excitation voltage
should be as high as possible to get a high output signal, but
high excitation voltages heat up the gauge and can produce
associated errors. The selected strain gauges and load cells
material allow maximum excitation voltages of 20 V [22].
The excitation voltage of 6 V was selected for all load cells.
The nominal values of the output signals from the load cells
are:
e1 = k ⋅ U ⋅ ε1 = 11.6 mV (5)
e2 = k ⋅ U ⋅ ε 2 = 11.97 mV (6)
where e1 and e2 are the output signals from the horizontal
and vertical load cells, respectively, and k is the strain
gauges sensitivity factor.
Before the external balance was assembled, the load
cells had been calibrated. The achieved accuracy, for each
load cell, was 0.05 % of the cell full scale.
In the mathematical model selected for the calculation of vector of the output signals from the load cells, [C ] is the
the calibration matrix for the new balance, the output
balance calibration matrix and
signals are expressed as functions of the applied loads:
{F } = {ZPD, ZPL, ZZ , XL, XD, YY }T is the vector of
{e} = [C ]{F } (7)
the balance load cells.
The matrix below shows the obtained calibration matrix:
where {e} = {eZPD , e ZPL , eZZ , eXL , eXD , eYY } T
is the
⎡ 7.84296 ⋅10−6 2.67539 ⋅10−8 2.60566 ⋅10−8 −1.11741 ⋅10−7 2.20897 ⋅10−7 −5.18028 ⋅10−8 ⎤
⎢ 4.39524 ⋅10−8 7.98568 ⋅10−6 4.55817 ⋅10−8 −2.51020 ⋅10−7 −2.60434 ⋅10−8 −1.45441 ⋅10−7 ⎥
⎢ ⎥
−5.28450 ⋅10−9 −2.74084 ⋅10−8 7.89606 ⋅10−6 6.19845 ⋅10−7 5.08105 ⋅10−8 −2.20494 ⋅10−8 ⎥
[C ] = ⎢⎢ (8)
−6.60848 ⋅10−8 1.55969 ⋅10−10 9.13932 ⋅10−8 1.84165 ⋅10−5 5.92386 ⋅10−7 2.30471 ⋅10−8 ⎥
⎢ −8 ⎥
⎢ −7.00291 ⋅10 1.56519 ⋅10−8 −1.34279 ⋅10−7 −5.71505 ⋅10−8 1.88196 ⋅10−5 −8.88533 ⋅10−8 ⎥
⎣⎢ 1.99289 ⋅10 1.86023 ⋅10−5 ⎦⎥
−7
1.36100 ⋅10−7 2.32718 ⋅10−7 −1.56258 ⋅10 −7 −2.09407 ⋅10−7
The accuracy achieved in the balance checkout is The accuracy achieved for the composite loads
displayed on the graphs in Fig.7 where Err is the maximum (calculated components of the aerodynamic load) is listed
difference between the applied and measured load and FS is in Table 3.
the component full scale. The nominal component
accuracies (displayed with lines on the graphs) are better Table 3. Accuracy of the composite loads
than 0.5% FS. Component X Y Z L M N
Range (N/Nm) 100 75 150 7 7 7
Fmax (N/Nm) 98.06 78.45 173.92 3.43 4.12 7.08
Fmin (N/Nm) -98.06 -78.45 -154.09 -3.43 -4.12 -7.08
Err (N/Nm) -0.338 -0.281 0.327 -0.062 0.061 -0.029
Err (% FS) -0.338 -0.374 0.218 -0.886 0.871 -0.414
Std.d (% FS) 0.103 0.112 0.085 0.213 0.246 0.154
Conclusion
The six-component external balance described in this article
is the first platform type balance designed and manufactured in
the VTI. It is shown that the balance has provided very good
accuracy. Accurate measurement of the aerodynamic forces
and moments is a major task in wind tunnels testing. Usually,
design and manufacturing of wind tunnel balances are based
on very contrasting requirements. The balance flexure
elements should be optimized such that the magnitude of the
strain response is approximately the same for the individual
application of each component of the load. However, the
magnitudes of the loads generated on a wind tunnel model in
each balance component are not equivalent; therefore, the
flexure elements do not have the same deflection in all
directions. Ideally, each balance response signal should
respond to its respective component of load and it should have
no response to other components of the load. This is not
entirely possible. The design of the described balance was
optimized to minimize interaction effects.
Figure 7. Accuracy achieved in the balance checkout The result of the calibration is a calibration matrix with six
Tables 2 lists a summary of the achieved accuracy of each terms per load cells. This was the first calibration of the new
load cell as obtained in a checkout after the calculation of the balance. The achieved accuracy of all balance load cells in the
calibration matrix, where Fmax is the maximum applied load, set is approximately 0.25% within a designed load range
Fmin is the minimum applied load and Std.d is the standard (based on three standard deviations), or even better. The
deviation of errors. A checkout was performed on the same accuracy achieved for the composite loads is very satisfactory
load subset that was used to calculate the matrix. considering the balance type and its operational requirements.
Combined load effects have to be included in the next
Table 2. Accuracy of the load cells calibration of the balance. It can be expected that the
introduction of load cases with combined load effects will
Load cell ZPD ZPL ZZ XL XD YY improve the accuracy of balance calibration.
Range (N) 200 200 200 100 100 100
Fmax (N) 87.00 87.00 121.09 49.03 49.03 78.4
Fmin (N) -91.67 -91.67 -125.76 -49.03 -49.03 -78.45 ACKNOWLEDGEMENT
Err (N) -0.396 0.390 -0.444 -0.226 -0.162 -0.281 This paper was supported by the Ministry of Education
Err (% FS) -0.198 0.195 -0.222 -0.226 -0.162 -0.281 and Science of Serbia [project number TR 36050, 2011-
Std.d (%FS) 0.056 0.064 0.080 0.082 0.039 0.084 2015].
SAMARDŽIĆ,M. etc.: EXTERNAL SIX-COMPONENT STRAIN GAUGE BALANCE FOR LOW SPEED WIND TUNNELS 45
Mots clés: soufflerie, balance de soufflerie à jauges à ruban, cellule de mesurage, étalonnage.