Jtam-A4.dvi


JOURNAL OF THEORETICAL

AND APPLIED MECHANICS

53, 4, pp. 1029-1039, Warsaw 2015
DOI: 10.15632/jtam-pl.53.4.1029

APPLICATION OF ANTHROPODYNAMIC DUMMIES FOR EVALUATING
THE IMPACT OF VEHICLE SEAT VIBRATIONS UPON HUMAN BODY

Tomasz Lech Stańczyk, Andrzej Zuska
Kielce University of Technology, Department of Automotive Vehicles and Transportation, Kielce, Poland

e-mail: a.zuska@tu.kielce.pl

The paper presents the anthropodynamic dummy constructed at the Department of Auto-
motive Vehicles and Transportation of the Kielce University of Technology, Kielce, Poland.
The dummy is shown at the concept and design stages, and as a final construction.The first
part of the study presents structural and dynamic requirements, and also the structure and
construction selected for the dummy. The vertical dynamics of the body of a seated human
is accounted for by conducting detailed analysis of the variation range of mass parameters
and of the dummy elastic and damping parameters. The second part of the study presents
the results of experimental simulations performed for the dummy.

Keywords: anthropodynamic dummy, vibratory comfort

1. Introduction

Initially, the measurements of transfer of vibrations by vehicle seats required participation of
human subjects exposed to vibrations. The research results obtained in this way were far from
satisfactory due to a significant variety of people’s individual body features. Moreover, such
measurements require input functions that must be safe for humans. Therefore, for ethical and
safety reasons, it is advisable to conduct evaluation of vehicle seats without people being subject
to vibrations. On the other hand, using stiff mass or a sand bag instead of a living person fails
to ensure results that are satisfactorily consistent with real conditions (Griffin, 1990; Łomako et
al., 1998; Mansfield and Griffin, 1996). Much better results can be obtained when the human
subject is replaced with a dummy imitating dynamic reactions of a seated person. An example
of such dummies is theMemosik. Information about impedance, measured under buttocks of a
seated person as well as about the head-seat acceleration transmittance are useful in building
the dummies. Memosik I was intended for tests at the 1-50Hz frequency range. An adequately
formed base and low centre of gravity ensure dummy stability during tests (Knoblauch et al.,
1995). Memosik 1 dummywas the basis forMemosik II model which is intended for tests at the
0.5-30Hz frequency range,which is typical for analyzingvibrational comfort inpassengervehicles
(Knoblauch, 1992).Memosik IIImodelwasbasedonMemosik II, inwhich thedegrees of freedom
were reduced from five to three (Cullmann andWölfel, 1998). Two subsequent dummymodels:
Memosik IV (Wölfel, 2006) andMemosik V (Mozaffarin et al., 2008) are active dummies, made
on the basis of a passive dummy. Memosik IV enables conduction of tests in vertical direction,
whereasMemosik V additionally enables tests in longitudinal and lateral directions.
Usingdummies is justifiedby the requirement of comparability and objectivity of the results.

In the course of comparative studies of several constructional solutions of seats, after some time,
people change their position to a more comfortable one, stiffen their bodies due to weariness,
etc. It distorts the research results. Dummies, on the other hand, manifest no such behaviour.



1030 T.L. Stańczyk, A. Zuska

2. Construction of anthropodynamic dummy

The construction of an anthropodynamic dummy presented in Fig. 1 is based upon the model
with four degrees of freedom, which imitates propagation of vertical vibrations to selected body
parts of a sitting person (Fig. 2).

Fig. 1. Anthropodynamic dummy on a passenger seat

By means of four masses, the anthropodynamic model imitates four selected parts of the
human body:
• mass m1 imitates mass of the head,
• mass m2 imitates mass of the pectoral girdle, chest and its internal organs and arms,
• massm3 imitatesmass of the internal organs, abdominal cavity, diaphragmandabdominal
walls,
• mass m4 imitates mass of the spine, pelvis, thighs, forearms and hands.

Fig. 2. Anthropodynamic model of a sitting human

The base of the dummy (imitating buttocks) and the support (imitating human back) con-
nected with the basis with a hinged grip is made of epoxy resin. The weight of those elements
accounts for 6 percent of the total dummyweight and it has not been taken into consideration
in the studies.
The dummy frame consists of guides, fixed andmoving discs, clamps and line bearings. The

frame is fixed to the base, in which six evenly spaced sockets have been made. Mounted in the
sockets are rubber springy-damping elements (Fig. 3), and the entire construction rests upon
the dummy base.



Application of anthropodynamic dummies for evaluating the impact... 1031

Fig. 3. The base of the dummy frame with rubber spring-damping elements mounted in six sockets:
1 – frame basis, 2 – rubber element

The guides are made of metal rollers. The accurate position of guides is ensured by fixed
metal discs connected with the guides by aluminium clamps. Three moving discs slide on the
guides, imitating respectively mass of the first (head), second (chest) and the third (abdomen)
part of the human body (Fig. 4). Mounting the suitable set of weights to vibrating elements
enables one to imitate individual anthropometric features of different body types.

Fig. 4. Moving discs imitating the respective body parts with a sample sets of weights: a – disc 1,
b – disc 2, c – disc 3

Themoving discs have holes where longitudinal bearings are mounted. The bearings ensure
minimal resistance to motion for the discs moving on the guides.
Each moving disc is connected to the dummy frame by means of a spring-damper system

consisting of a damper and a spring (Fig. 5).

Fig. 5. Spring-damper system: 1 – spring, 2 – damper

Properly selected sets of weights, dampers, springs and rubber elements enable the dummy
to imitate the vertical dynamics of various types of the human body.

2.1. Selection of spring-damping elements of the anthropodynamic dummy and
identification of their parameters

One of the fundamental requirements to be met by the anthropodynamic dummy is the
accurate imitation of the vertical dynamics of a seated humanbody.For that purpose, a number



1032 T.L. Stańczyk, A. Zuska

of tests has been carried out on a stand constructed on the basis of MTS components (Zuska,
2007). The objective of the tests was to determine the transmittance of acceleration for three
configurations (output-input): head-seat, chest-seat and abdomen-seat in the selected group of
people, Fig. 6 (Zuska, 2007).
Upon the transmittance courses for medians and quantiles 0.1 and 0.9 of the modules de-

signated in laboratory tests, the stiffness (k1, k2, k3, k4), damping (c1, c2, c3, c4) and mass
parameters of the model presented in Fig. 2 are identified.

Fig. 6. Courses of transmittance modules: (a) head – seat, (b) chest – seat, (c) abdomen – seat

Before the selection of the supple elements, a number of constructional requirements to be
met was adopted. The requirements have been established upon the assumed dimensions of
the dummy and the identified parameters of stiffness and damping elements of the mathematic
model (Table 1). For the obtained stiffness and damping coefficients, a 15% toleration has been
permitted.

Table 1. Parameters of the anthropodynamic model

Parameter Quantile 0.1 Median Quantile 0.9

Mass [kg]

m1 4.83 5.77 7.43
m2 13.75 16.43 21.17
m3 6.75 8.06 10.38
m4 33.05 39.40 50.90

Stiffness
coefficient
[N/m]

k1 3898 5098 7169
k2 14856 18124 24181
k3 11350 14761 20644
k4 508863 398800 279187

Damping
coefficient
[N·s/m]

c1 200 212 302
c2 254 257 349
c3 197 209 245
c4 13385 1113 17



Application of anthropodynamic dummies for evaluating the impact... 1033

The tests aimed at determining:

• stiffness coefficient of the springs,
• damping coefficient of the dampers,
• stiffness and damping coefficients of the rubber elements,

and have been carried out on the testing stand adopted for such tests (construction based on a
hydraulic actuator manufactured byMTS). They which enabled imitation of the working para-
meters of dampers, springs and rubber elements (Table 2) as well as recording of the following
signals:

• displacement of the plate with the fixed rubber element or spring and displacement of the
holder with the fixed cylinder of the damper,

• compressive force affecting the rubber element or spring and damping force of the shock
absorber.

Table 2.Working parameters of spring-damping systems

Parameter
System

No. 1 No. 2 No. 3 No. 4

Frequency [Hz] 1-20 1-20 1-20 1-20
Amplitude [mm] 0.00-2.73 0.00-4.49 0.00-2.15 0.05-0.62
Pre-load load [kg] 5.77 16.43 8.06 69.66

2.1.1. Selection of coil springs and identification of their stiffness coefficient

Taking into consideration the constructional assumptions and stiffness coefficients (k1, k2,
k3) identified for the mathematical model, a pre-selection of springs has been made.
For every pre-selected spring, a test has been carried out to identify their actual stiffness

coefficient (Table 3).

Table 3.Parameters of the 50-centile dummy springs

Stiffness coefficient [N/m] Deviation from

Designation
Assumed Catalogue Test-determined the assumed
value value value value [%]

k1m50 5098 5300 5495 7.8
k2m50 18124 17781 20249 11.7
k3m50 14761 15556 17230 16.7

The obtained values of stiffness coefficients k1m50 and k2m50 differ from the assumed values
by less than 15 percent and fall within the assumed range of toleration, whereas the value of the
k3m50 stiffness coefficient differs from the assumed one by 16.7 percent and does not meet the
15-percent toleration range. However, a decision has been made to keep the spring and not to
reject it until the entire dummy has been examined.

2.1.2. Selection of dampers and identification of their damping coefficient

Taking into consideration the constructional requirements:

• dampermaximum length 250mm,
• minimum stroke of the piston rod 10mm,



1034 T.L. Stańczyk, A. Zuska

as well as damping coefficients (c1, c2, c3) identified for themathematical model, a pre-selection
of dampers has been made.
The actual characteristics of the selected dampers differed from the required ones. Therefore,

constructional changes have beenmade, which included:

• using oil of various density and viscosity,
• replacing thevalves installed in thepistonbysuitablediscswithopenings (ducts,bypasses).

After each constructional change in the dampers, a test has been performed to determine
the work diagram, which served as a basis for identifying the damping coefficient of the tested
dampers (Fig. 7).
The obtained values of the damping coefficient (Table 4) differ from the assumed values by

less than 15 percent.

Table 4. Parameters of dampers of the 50-centile dummy

Damping coefficient [N·s/m] Deviation from the
Designation Assumed value Determined value assumed value [%]

c1m50 212 222.65 5.0
c2m50 257 264.36 2.9
c3m50 209 191.54 8.4

2.1.3. Selection of rubber elements and identification of their damping (c4) and stiffness (k4) coefficients

Therubberelement in themathematicalmodel ispresentedasaKelvin-Voigt spring-damping
element.
Taking into consideration the constructional requirements:

• rubber elements have cylindrical shape of a 50mm diameter,
• height of the rubber elements falls within the 20-100mm range,
• quantity of the rubber elements to ensure the dummy stability is either three, four or six
items,

as well as the damping (c4) and stiffness (k4) coefficients identified for themathematical models,
a pre-selection of the rubber elements has been made (Table 5) followed by identification of
actual values of those coefficients.

Table 5. Initial parameters of the selected rubber elements

Hardness Diameter Height
Shape

[◦Sh] [mm] [mm]

40, 50, 60, 70, 80, 90 50 20, 40, 60 cylinder

For every pre-selected rubber element, a test has been carried out to determine its static
characteristics (Fig. 8) uponwhich the stiffness and damping coefficients for the tested elements
have been identified.
An assumption has beenmade that the characteristics of the rubber element f(λ,λ̇)may be

presented in an additive form

f(λ,λ̇)= fs(λ)+ft(λ̇)

where fs(λ) is the stiffness component of the characteristics, ft(λ̇) – the damping component of
the characteristics.



Application of anthropodynamic dummies for evaluating the impact... 1035

Fig. 7.Work diagrams and the actual characteristics of dampers utilised in the 50-centile
anthropodynamic dummy: (a) work diagram of disc 1 support, (b) damping characteristics of disc 1
support, (c) work diagram of disc 2 support, (d) damping characteristics of disc 2 support, (e) work

diagram of disc 3 support, (f) damping characteristics of disc 3 support

Given a relatively small nonlinearity of the frame axis within the range of actual work
parameters (pre-load load – 23kg, deflection amplitude – 0.62mm) it has been assumed that it
has a linear shape (Fig. 8b, Fig. 9). The slope of the linear regression line of the frame axis is
the required stiffness coefficient of the rubber element (Fig. 11).

Since thevelocity of thedamperpiston rod is known for everypoint in thediagram(Fig. 10a),
it is possible to determine the damping characteristics ft(λ̇) (Fig. 10b).



1036 T.L. Stańczyk, A. Zuska

Fig. 8. Characteristics of selected rubber elements: (a) for wide range of loads, (b) made for actual work
parameters (actual load range)

Fig. 9. Stiffness characteristics of the rubber element

Fig. 10. Characteristics of the rubber element without stiffness component and the damping
characteristics based upon it: (a) work diagram of the rubber element, (b) damping characteristics

The slope of the linear regression line of the damping characteristics is the required damping
coefficient of a single rubber element (Table 6).

The obtained value of the stiffness coefficient differs from the assumed value by 18 percent
and meets the 20 percent tolerance range, whereas the value of the damping coefficient differs
from the assumed value by 98% and fails to meet the 20 percent tolerance range. It has been
decided to keep the rubber elements and not to reject it until the entire dummy has been
examined.



Application of anthropodynamic dummies for evaluating the impact... 1037

Table 6.Parameters of the selected rubber elements

Coefficient Designation
Assumed Determined Deviation from the
value value assumed value [%]

Damping [N·s/m] c4m50 1113 2207.94 98
Stiffness [N/m] k4m50 398800 470940 18

3. Anthropodynamic dummy – the results of verification tests

The value measured in the tests has been the acceleration of three moving discs of the dummy
(disc 1, disc 2 and disc 3) and the acceleration of the surface of the plate on which the dummy
is placed.
The recorded acceleration signals have been used for determining the transmittance of three

input-output systems: plate – disc 1, plate – disc 2, plate – disc 3.
The list of the test results conducted for the dummy and themedian transmittance courses,

determined for 80 tested people, are presented in the Fig. 11.

Fig. 11. Results of the experiment – testing of the dummy and the median of tests of 80 people:
(a) transmittance of disc 1 and body part 1, (b) transmittance of disc 2 and body part 2,

(c) transmittance of disc 3 and body part 3

Thehighest compatibility is displayedby the transmittance courses of disc 2 andbodypart 2.
As presented in the diagram of transmittance courses, the highest enhancement is observed for
frequencies of 5Hz and 5.5Hz, respectively. The difference between themaximum enhancement
values for these characteristics is 3.3 percent. The courses of the curves for other frequency
ranges presented in the diagrams are also very similar.
As indicatedby the transmittance courses of bodypart 1anddisc 1, thehighest enhancement

for both courses occurs for the same frequency of 4.5Hz. The difference of enhancement values
for this frequency is 10.6 percent. The characteristics have similar courses in the frequency range
between 3 and ca. 12Hz. Above the latter frequency, there are increasing discrepancies between
the characteristics.



1038 T.L. Stańczyk, A. Zuska

The transmittance courses for body part 3 and disc 3 are the least compatible. The highest
enhancement for both these courses exists in different resonant frequencies, with a 1.5Hz off-
set with respect to one another. The difference between the maximum enhancement values is
15percent. In this instance, the characteristics display thehighest compatibility in two frequency
ranges: 1-6Hz and 9-12.5Hz.

4. Conclusion

Testing vibrational comfort in vehicles may involve either the entire body of a sitting person or
its selected parts. For ethical reasons, aswell as for ensuring safety and repeatability of obtained
results, people are increasingly often replaced with anthropodynamic dummies. The structure
of such dummies, used for testing vibrational comfort, should reflect the structure of a sitting
human being while taking into consideration the analysed body parts for which test are to be
carried out.
The results of the conducted experimental tests indicate that at least three areas of the

human body should be distinguished in a dummy, as each one is characterised by a different
resonance frequencyandvarious levels of acceleration enhancement.The structure of thedummy
presented in this paper comprises these parts. The proposed amount of degrees of freedom is a
compromise between obtaining the most accurate representation of the human body structure
and the constructional limitations which occurred as the dummywas being built.
The symmetric construction of the dummy prevents the occurrence of forces producing

unwanted longitudinal or lateral rocking which, in consequence, increases the dummy stabi-
lity. Properly formed buttocks and the back ensure the adequate imitation of the distribution of
individual pressures produced by a vehicle driver or passenger. The construction of the dummy
enables one to imitate mass structures of people of various weight (slim, normal weight, obese).
A significant qualitative and quantitative similarity between the characteristics obtained

in laboratory tests of the dummy transmittance and the respective characteristics of human
subjects proves that dummies (with a construction based upon the results of the experiment)
may be successfully applied to evaluate the impact of vibrations on vehicle passengers.
The assumed solution enables a fairly easy reconfiguration of the dummy to imitate the 10th

and the 90th percentile of the weights of tested people.

References

1. CullmannA.,WölfelH.P., 1998,Komfortbewertung und Sitzprüfungmittels des Schwingungs-
dummys MEMOSIK, Fachtagung, Haus der Technik, Essen

2. Griffin M.J., 1990,Handbook of Human Vibration, Academic Pres, London

3. KnoblauchJ., 1992,Entwicklung undBau eines physikalischen Schwingungsmodells des sitzenden
Menschen, Dissertation, Shaker Verlag

4. KnoblauchJ.,WölfelH.P.,BuckP., 1995,EinSchwingungsdummydes sitzendenMenschen,
ATZ – Automobiltechnische Zeitschrift, 97, 10

5. Łomako D.M., Stańczyk T. L., Wyrębkiewicz A., 1998, Zagadnienie komfortu jazdy w sa-
mochodach z pasywnym zawieszeniem fotela kierowcy, VI Międzynarodowa Konferencji ”Pojazdy
Samochodowe. Problemy Rozwoju, Jakości, Eksploatacji AUTOPROGRES”, 1

6. MansfieldN.J.,GriffinM.J., 1996,Vehicle seat dynamicsmeasuredwith an anthropodynamic
dummy and human subjects,Proceedings of Inter-Noise 96, Liverpool UK



Application of anthropodynamic dummies for evaluating the impact... 1039

7. Mozaffari A., Pankoke S., Wölfel H.P., 2008,MEMOSIKV –An active dummy for deter-
mining three-directional transfer functions of vehicle seats and vibration exposure ratings for the
seated occupant, International Journal of Industrial Ergonomics, 38

8. Nélisse H., Boileau P.É., Rakheja S., Patra S., Boutin J., 2006, Évaluation de manne-
quins anthropodynamiques destines aux essais de vibrations des sièges en laboratoires; Études et
Recherches, Raport R479

9. WölfelH.P., 2006,Numericalmodels andhardwaredummies for simulatingwholebodyvibration
of human – an overview,Proceedings of the First American Conference on HumanVibration,West
Virginia, U.S.A., DHHS (NIOSH) Publication No. 2006-140 June

10. Zuska A., 2005, Identyfikacja rozprzestrzeniania się pionowych drgań w organizmie siedzącego
człowieka,Teka Komisji Motoryzacji PAN, Oddz. Kraków, 29/30

11. ZuskaA., 2007,Analizaporównawczarozprzestrzeniania się pionowychdrgańworganizmachosób
o różnej budowie ciała,Logistyka, 3

12. ZuskaA., Stańczyk T.L., 2005, Stanowisko do symulacyjnych badań podukładu fotel-kierowca,
Prace Naukowe TRANSPORT, Wydawnictwo Politechniki Radomskiej, 21, 1, 149-158

Manuscript received November 29, 2013; accepted for print June 26, 2015