Dynamic Characteristics of Memorial

Columns of Trajan and Marcus Aurelius

using Microtremor

Yutaka Nakamura, Jun Saita, Tsutomu Sato

System and Data Research, Tokyo JAPAN

 Columns of Trajan and Marcus Aurelius
Image Model of Trajan’s Column Comparison
Spiral Staircase in
Columns: Trajan Marcus Aurelius
Columns (TC) (AC)
Height: 185 steps 190 steps
(entrance to top floors)
Foundation: Hard Soft
Ground
Construction: AD 113 AD 193
Massive Repair: none at 16c.
Common:
Marcus Stacking of the marble brock with 8 steps,
Aurelius’s
Column one turn of spiral staircase consist of 14 steps

Two columns have same size and

structure on foundation with
different hardness build before about 19c.

Plaster Cast Miniature of I would like to know the physical situation

the Marble Brock of columns in present using microtremor.
 Object

Object of this research is to know the following items;

1) Natural Frequency in present

2) Characteristics of stacking structure of marble
blocks; Bending or Shearing?
3) Necessity of simultaneous measurement: if could
not conduct it, how to correct?
4) Vulnerability of the Columns
Microtremor Measurement
The measurements were done at 28th March of 2012,
TC in the morning and AC in the afternoon.

At each point, microtremor of three components EW, NS and

UD were measured simultaneously for three minutes with 100
Hz sampling.

Frequency analysis has been carried out by Fourier transform

for the measured waveform. The response spectra have been
estimated using simultaneous measurement for TC, and
estimated using individual measurement for AC. The analysis
used 40.96 seconds (4096 data) length record and the final
result was derived from spectra averaged three or four times.

Also the spectral ratio of target point against base point or

H/V spectral ratio is derived.

These results of individual or simultaneous measurement are

compared and investigated.
 Measurement
MC

MTS
MTE MTC MTW
TTE TTC
TTN

Plan North for TC TTS MTN ①

TTW ①

Points for Columns

②
②
③
③
MC1 ④
TC1
④
⑤
⑤

⑥
⑥

⑦
Trajan’s column has 17 measurement points.
TC2 ⑦ MC2

⑧
⑧

slippage joint
Simultaneous measurement for two points with
⑨
⑨
⑩
three components were done:
⑩
MC3 ⑪
TC3
⑪
⑫
(1)N and S or E and W on the top terrace
⑫
⑬ for torsional and/or vertical vibration
⑬

⑭ MC4
⑭
MC4
(2)on the top terrace and the other steps of
TC4 ⑮

staircase for shear or bending vibration

TC4
Plan North for TC
Trajan's ⑮
Column ⑯ Marcus
point steps ⑯
TC45
Aurelius's
⑰ Column
TTN 185
TTE 185 ⑰ point steps
MC5 ⑱
TTS 185 MTN 190
Marcus Aurelius’ column has 13 measurement
TC5 う
TTW 185 ⑱
⑲ MTE 190
TTC 185 う
⑲ MTS 190
TC1 157
TC2 129
TC6 MTW 190
MTC 190 points.
TC3 101 MC1 162
TC4 73 MC6 MC2 134
Individual measurement for each points with
TC7
TC45 59 MC3 106
TC5 45 MC4 78
TC6 31
TC7 26
TBE
MC5 50
MC6 22 three components were done at a similar points
TBN 0 MC7 0
TBE
TBS
0
1
TBN
TBW
TBS MC7
MC8 -14 on the top terrace and the shaft of Trajan’s
column.
MC8
TBW 0 Ｃ８
MC8
*MC8 より下には螺旋階段はなく、狭
く複雑な地下空間が拡がっているよ
(1) Trajan’s column (2)
うにみえた。
Marcus Aurelius’ column
*ブロック⑩と⑪の継目部分（赤）の
表面レリーフは、8.2cm ずれている。
 Amplification spectra of TC estimated
from simultaneous measurement
100
EW component
185 These figures show the
157
10
129
101
amplification spectra of the points
73
in the shaft and on the terrace
Amplification Factor (Reference Point TC7)

59
1 45
31
26
against the point TC7, 26 steps
0.1
0.1 1 10 100
from the lowest level in a pedestal
100
NS component
185 as a rigid block.
157
129
10
101 It is possible to draw the vibration
73

1
59
45
mode for the entire frequency
31
26
mode because the phase can be
0.1
0.1 1 10 100 grasped by the simultaneous
100
UD component 185
157
measurement.
129
10
101
73
I have tried to visualize the
1
59
45
vibration mode sweeping
31
26 frequency with animation.
0.1
0.1 1 10 100
Frequency in Hz Here I introduce the movie for TC.
 Amplification spectra of TC estimated
from individual measurement
1000
185
100 157
129
101
Averaged spectra for each
10 73
59
point are calculated. And then
1 45 the amplification spectral
31
0.1 ratio against reference point
0.1 1 10 100
Amplification Factor

1000 185 TC-7 measured different time

157
100 129 for each points are calculated
101 as individual measurement.
10 73
59
45
1 31 Extremely large amplifications
0.1 are recognized because of
0.1 1 10 100
1000 different vibration source at
185
100 157 each measurement.
129
101
10 73
59
1 45
31
0.1
0.1 1 10 100
Frequency in Hz
 H/V spectral ratio of TC
100
185
EW/UD 157
129
These figures are lapped
10 101
73
H/V spectral ratio of points
59
in the TC shaft.
Amplification Factor

45
1
31
26

0.1 1.514
0.1 1 10 100

100
185
NS/UD 157
129
10 101
73
59
45
1 31
26

0.1 1.440
0.1 1 10 100

Frequency in Hz
Results of Microtremor
Measurements for TC
Table 1. Results of Microtremor Measuements for TC
TC Simultaneous. F Hz maxAF maxKb/Kg This is summary of measurement.
EW 1st 1.59 18.8 83.1
According to this table, comparing
2nd 8.99 7.8 5.0
3rd 19.17 6.3 1.2 simultaneous and individual
NS 1st 1.59 15.5 77.2 measurements, they are almost
2nd 8.99 5.2 3.4
equal for each other on natural
3rd 20.29 9.3 1.0
Longitudinal 1st 25.64 7.8 0.2 frequencies, but there is large
Torsional 1st 15.53 difference for 1st mode shape. On
TC Individual F Hz maxAF maxKb/Kg
the other hand, from H/V spectral
EW 1st 1.59 188.8 2271
2nd 9.03 8.3 4.8 ratio, estimated natural
3rd 19.04 6.0 1.2 frequencies are little lower but
NS 1st 1.59 98.6 1293 amplification factor is close to the
2nd 9.03 7.3 4.4
3rd 20.26 9.6 1.3 results estimated by simultaneous
TC H/V F Hz maxAF maxKb/Kg measurement.
EW 1st 1.51 22.0 181
NS 1st 1.44 11.2 111
Ground: BN 1st 5.81 1.9 0.6
BE 1st 5.00 2.0 0.8
BS 1st 5.27 2.3 1.0
BW 1st 5.00 2.3 1.0
Estimation of Wave Propagation Velocity
Table 1. Results of Microtremor Measuements for TC
TC Simultaneous. F Hz maxAF maxKb/Kg
Vertical or Torsional
EW 1st 1.59 18.8 83.1 Longitudinal
2nd 8.99 7.8 5.0
3rd 19.17 6.3 1.2
NS 1st 1.59 15.5 77.2
2nd 8.99 5.2 3.4
3rd 20.29 9.3 1.0
Longitudinal Fv 1st 25.64 7.8 0.2
h
Torsional Ft 1st 15.53
TC Individual F Hz maxAF maxKb/Kg
EW 1st 1.59 188.8 2271
2nd 9.03 8.3 4.8
3rd 19.04 6.0 1.2
NS 1st 1.59 98.6 1293
2nd 9.03 7.3 4.4
3rd 20.26 9.6 1.3
TC H/V F Hz maxAF maxKb/Kg
EW 1st 1.51 22.0 181
Because theNS propagation
1st velocity
1.44 of longitudinal
11.2 vibration
111 of ordinary marble would
Ground: BN 1st 5.81 1.9 0.6
be 4800m/s,BE
the marble
1st
of Trajan’s
5.00
column2.0
might 0.8
be deteriorated to 65% of initial
value of propagation
BS 1stvelocity. Estimated2.3
5.27 Poisson's1.0
ratio is agreeable value as marble.
BW 1st 5.00 2.3 1.0
 Mode diagram of TC derived from
simultaneous measurement
200
180 This figure shows mode
160 diagram for 1st to 3rd natural
Trajan's frequencies of each direction,
Step Numbers

140

120
Column EW and NS,
from simultaneous
100
Mode Shape measurement.
3rd 19.165Hz_EW
80 The shape of 1st mode
2nd 8.984 Hz_EW
1st 1.587 Hz_EW
suggests that bending
60
3rd. 20.288Hz_NS vibration predominates.
40 From the ratio higher mode
2nd 8.984 Hz_NS
20 1st. 1.587 Hz_NS frequency to 1st mode
0 frequency, bending vibration
-10 -5 0 5 10 15 20 25 might predominate.
Amplification Factor
1stF : 2ndF : 3rdF = 1 : 6.2669∙∙∙ : 17.5449∙∙∙ (Bending), 1 : 3 : 5 (Shearing)
= 1 : 5.65 : 12.06 (NS), 1 : 5.65: 12.76 (EW) for TC, almost Bending
= 1 : 5.11 : 11.64 (NS), 1 : 5.72: 12.41 (EW) for AC, almost Bending
 200

180
Comparison of the
160

140
mode diagram of TC
Trajan's Column
Step Numbers

120

100
1st Mode with various
80

60
Indiv.
Indiv.
H/V
H/V
Simul.
Simul.
1.587Hz_EW
1.587
1.514
Hz_EW
1.514Hz_EW
Hz_EW
1.587Hz_EW
1.587 Hz_EW
measurement
40
Indiv.
Indiv. 1.587Hz_NS
1.587 Hz_NS These figures show estimated mode diagram
H/V
H/V 1.440Hz_NS
1.440 Hz_NS
20 Simul.
Simul. 1.587
1.587 Hz_NS
Hz_NS
using spectral ratio to point TC7.
0
0 10 50 20 30
100 40 150 50 60
200
Predominant frequencies are almost equal to
Amplification Factor that of simultaneous measurement.
200
Trajan's Column 200 However, the amplification factor of 1st mode
Step Numbers
Step Numbers

180
2nd Mode
180 from individual measurement becomes larger
160 160
close to the terrace, although higher mode
Step Numbers
Step Numbers

140 140
shapes are almost similar.
120 120 Trajan's Column
100 100 3rd Mode It is necessary for accurate estimation to
80 80 restrict surrounding motion during
Indiv. 19.043Hz_EW
Indiv. 9.033 Hz_EW 60 60
Simul.
measurement or conducting simultaneous
Simul. 8.984 Hz_EW 19.165 Hz_EW
Indiv. 9.033 Hz_NS
40 40
Indiv. 20.264 Hz_NS measurement.
Simul. 8.984 Hz_NS 20 20 Simul. 20.288 Hz_NS In case of individual measurement, it is
0 0
-10 -5 0 5 10 -10 -5 0 5 10 15
expected that the amplification factor and
Amplification Factor Amplification Factor mode shape can be estimated by the H/V.
 On the Kb-value
a Destructive Index
dj: Displacement at jth Floor This is explanation of
nth Floor
A gj: Drift Angle at jth Layer Kb-value. It can be
used for estimation
dj = Ajag/wn2 of drift angle for
A dj = Ajag/(2pfn)2
jth Floor gj each layer by Kb-
(j-1)th Floor hj
gj = (Aj – Aj-1)/hj/(2pFn)2ag value times PGA,
A = Kbjag peak ground
1st Floor acceleration.
g1 Kbj = (Aj – Aj-1)/hj/(2pFn)2
Ground Floor
Seismic Acceleration ag ag: PGA
Aj: Amplification Factor at jth Floor
Fn: Natural Frequency of the Building
hj: Height of jth Layer
Large K-value means that the portion is vulnerable.

Please see the detail of Kb value on reference below as a name of KT.

Nakamura, Y. and et al.: Vulnerabiliy Investigation of Roman Collisseum using Microtremor,
12th WCEE, paper #2660, 2000.
 Kb-value of TC estimated from
simultaneous measurement
200 This Figure shows a destructive
180 index Kb value derived from the
estimated amplification mode of
160
Trajan's Column horizontal vibration.
Step Numbers

140
Kb Distribution According to this figure, the
120
by Simul. H/H influence of the first order vibration
100 mode is very large; and it is possible
3rd 19.165Hz_EW to ignore higher modes.
80
2nd 8.984 Hz_EW
60 1st 1.587 Hz_EW
Distribution of Kb values by the first
3rd. order vibration grows higher toward
40 20.288 Hz_NS
to the top, and reaches the
2nd 8.984 Hz_NS
20 maximum value of 83μ/Gal at the
1st. 1.587 Hz_NS
top. This means that strong motion
0
of PGA 100 Gal will cause 1/120 of
0 20 40 60 80 100
Kb (m/Gal) drift angle at the top of column near
the terrace.
Amplification spectra of AC estimated
from individual measurement
100
EW component 190
162
10
134
106
78
Amplification Factor (Reference Point MC7)

50
1 22
0
-14
0.1
0.1 1 10 100
100
190
NS component 162
134
10
106
78
50
1 22
0
-14
0.1
0.1 1 10 100
100
190
UD component 162
10 134
106
78
50
1 22
0
-14
0.1
0.1 1 10 100
Frequency in Hz
 H/V spectral ratio of AC
100
190
EW/UD 162
10
134
106
78
Amplification Factor

50
1
22
0
-14
0.1
0.1 1 10 100
100
NS/UD 190
162
10
134
106
78
50
1
22
0
-14
0.1
0.1 1 10 100
Frequency in Hz
 Results of Microtremor
Measurement for AC
Table 2. Results of Microtremor Measuements for AC
AC Individual F Hz maxAF maxKb/Kg
EW 1st 1.29 64.0 936
2nd 6.59 15.2 17.9
3rd 15.02 12.4 2.3
NS 1st 1.22 53.3 957
2nd 6.98 19.4 15.9
3rd 15.14 20.9 3.0
Longitudinal 1st 18.80 57.7 1.0
Torsional 1st 9.13 Metal Fasteners?

2nd 25.95
AC H/V F Hz maxAF maxKb/Kg
EW 1st 1.25 18.8 212
NS 1st 1.22 13.9 144
Ground: MC7 1st 1.22 2.54 5.3
CMGW 1st 1.37 2.76 5.6 Damage of AC at West Side
Based on the predominant frequencies of longitudinal and torsional vibration,
propagation velocities are estimated 2300m/s and 1100m/s, respectively. But from
them estimated Poisson's ratio is over 1, abnormal value. This suggest extremely
low rigidity for torsional vibration. Furthermore, resonance frequency of surface
ground is near 1.3Hz, almost equal to the predominant frequencies of AC. They
affect against AC when earthquake occurred, then AC might suffer damage.
 Mode diagram of AC derived from
individual measurement
200
180
On the 1st mode of AC,
amplification factor near the
160
terrace becomes larger same
140 as the case of TC.
Step Numbers

3rd 15.015 Hz_EW

120 2nd 6.592 Hz_EW
1st 1.294 Hz_EW
100 3rd. 15.137 Hz_NS
80 2nd 6.982 Hz_NS
1st. 1.221 Hz_NS
60
Marcus Aurelius's
40
Column Mode Shape by
20
Indiv.H/H
0
-20 -10 0 10 20 30 40 50 60 70
-20
Amplification Factor
 Comparison between the first mode shapes of H/V and
H/H of AC from individual measurements
200

This is a 1st mode shape

150 comparing HH spectral ratio
Marcus and HV spectral ratio.
It seems that large HH
Aurelius's Column
Step Numbers

100 spectral ratio can be

1st Mode modified by HV spectral
Indiv. 1.294Hz_EW ratio similar to the
H/V 1.245 Hz_EW
50
Indiv. 1.221 Hz_NS
discussion for TC.
H/V 1.221 Hz_NS

0
0 10 20 30 40 50 60 70
Amplification Factor
-50
 Comparison of estimated Kb-values for
TC ( ) and AC ( )
200 This is the comparison of Kb
180
values of TC and AC.
160
Kb values of TC grow lager
Kb Distribution of
Step Numbers

140 towards to higher point

120 Trajan's (simul.) and reaching maximum 83 m/Gal.
100 Marcus Aurelius's Kb values of AC show
80
(H/V) Columns variation at each point, and
60 H/V 1.245 Hz_EW the maximum value is 212
40
Simul. 1.587 Hz_EW m/Gal near the terrace.
H/V 1.221 Hz_NS
20 Simul 1.587 Hz_NS Drift angle over 1/120 will
0 occur with over 100Gal for
0 50 100 150 200 250 TC, and in case of AC the
Kb (m/Gal)
same drift angle will occur
even under 50Gal.
 Summary
Columns of Trajan and Marcus Aurelius were studied.

The fundamental frequencies of horizontal vibration are 1.6 Hz for

TC and 1.3 Hz for AC, and they are caused by bending vibration.

For TC, Predominant frequencies of torsional and longitudinal

vibration are estimated 15.5 Hz and 25.6 Hz, respectively.
From these values the propagation velocities of shear wave and
longitudinal vibration are estimated as 1900 m/s and 3100 m/s,
respectively, about 65% of typical value, Poisson's ratio is estimated
about 0.33 and it is reasonabe. Metal Fasteners?

For AC, the torsional and longitudinal vibrations are 9.1 Hz and 18.8
Hz, respectively, and the velocity of shear wave and longitudinal
vibration are estimated as 1100 m/s and 2300 m/s, respectively.
Longitudinal velocity is more than 50% lower than typical value,
and the shear wave velocity is too low to estimate Poisson's ratio.
It seems to be caused by low natural frequency of torsional
Damage of AC at West Side
vibration, and damage for vertical direction on the west side of
column body can be observed.
 Summary-2
The destructive index Kb-value of the columns derived from the fundamental
mode of bending vibration is 83 μ/Gal in maximum for upper part of TC and is
212 μ/Gal for top of AC with large variation. Thus, drift angle over 1/120 will
be caused by about 100 Gal of PGA for TC and lower than 50 Gal of PGA for
AC.

Considering the columns consist of the stacking marble brock, their stabilities
might be lost by large deformation. It is necessary not only to calculate
numerically based on proper numeric model in detail but also to observe
seismic motion at various level of columns.

I would like to recommend the investigation using microtremor repeatedly in a

certain period for monitoring the changing status of columns.
Microtremor measurement in this research could be
realized under kind and strong corporation of
Professors Valente and Tallini.
Thank you very much for their efforts.

Thank you for

your kind attention