Ultrasound classification for recycled glulam sleepers

Abstract

Nowadays, due to environmental concerns and high market value, the railroads dramatically decreased the use of native timber sleepers. Aiming to reducing environmental impact, a recycling process was developed where wooden sleepers that are discarded or sold at low prices by railway companies are transformed into glulam railway sleepers. Vale Company donated 46 railway sleepers for this study, which 35 of these were transformed into 6 new railway sleepers. In this article, the non-destructive tests on discarded railway sleepers and on the thin boards procedures are described. These tests were performed in order to determine the greater longitudinal moduli of elasticity for the upper and lower boards’ distribution of the recycled glulam sleepers (RGS) and also for determining the direction of lamination of the sleepers, to provide a greater reuse of materials. The results indicated that the railway sleepers ultrasound test are not indicated and should be replaced by testing only one board of each sleeper, which is sufficient to classify the pieces.

Full text

Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013 www.ijaser.com
© 2013 by the authors – Licensee IJASER- Under Creative Commons License 3.0 editorial@ijaser.com
Research article ISSN 2277 – 9442
—————————————
524
1 Corresponding author (e-mail:mantilla@dees.ufmg.br)
2 pi@piegenharia.com
3 cynarafiedlerbremer@ufmg.br
4 recosta_88@hotmail.com
5 judymantilla@hotmail.com
Received on September 19, 2013; Accepted on September 29, 2013; Published on October, 29 2013
Ultrasound classification for recycled glulam sleepers
(1) Carrasco, E. V. M., (2) Passos, L.B., (3) Bremer, C. F., (4) Alves, R. C., (5) Mantilla, J.N.R
Federal University of Minas Gerais
DOI: 10.6088/ijaser.020500006
Abstract: Nowadays, due to environmental concerns and high market value, the railroads dramatically
decreased the use of native timber sleepers. Aiming to reducing environmental impact, a recycling process
was developed where wooden sleepers that are discarded or sold at low prices by railway companies are
transformed into glulam railway sleepers. Vale Company donated 46 railway sleepers for this study, which
35 of these were transformed into 6 new railway sleepers. In this article, the non-destructive tests on
discarded railway sleepers and on the thin boards procedures are described. These tests were performed in
order to determine the greater longitudinal moduli of elasticity for the upper and lower boards’ distribution
of the recycled glulam sleepers (RGS) and also for determining the direction of lamination of the sleepers,
to provide a greater reuse of materials. The results indicated that the railway sleepers ultrasound test are not
indicated and should be replaced by testing only one board of each sleeper, which is sufficient to classify
the pieces.
Keywords: non-destructive tests; ultrassound; recycled sleepers; wood.
1. Introduction
This paper presents the nondestructive testing using the ultrasonic method, performed in railway sleepers
for determining the modulus of elasticity and subsequent distribution of the pieces obtained for RGS
fabrication. There are several non-destructive tests, among them the ultrasound technic, which is based on
the principle of mechanical wave propagation in materials, which can be considered a reliable test, cheap
(when compared to the automatic classification machines) and training of the workers is relatively simple.
According to JAMES INSTRUMENTS INC (4), the method of ultrasound was first developed for use in
concrete in Canada by Leslie and Cheesman between 1945 and 1949 and also independently in Britain by
Jones and Gatfield. The device developed by them was a cation ray oscilloscope for measuring
transmission time. BUCUR states that when a longitudinal wave propagates parallel to the fibers, the
wavelengths and the lengths of the cells are of the same order of magnitude, where the propagation occurs
in stochastic scattering regime. Perpendicularly to the fiber (radial and tangential direction), the
wavelengths are greater than the average dimensions of the cells and spreads occur within a Rayleigh
scattering regime. These two propagation conditions are related to the frequency used. According to
TRINCA and GONÇALVES the general equations presented in studies on the propagation of waves in
solid media, assume that the ideal propagation medium is infinite, that is, the wavelength (λ) is much
smaller than the dimensions of the cross section of the body whereby this wave crosses. When the
dimension of the piece to be tested, according to the propagation direction of the wave, approaches the
wavelength (λ), the medium become known as finite and, in this case, the phenomena of refraction and
reflection of the waves begin to affect propagation and, therefore, the basic theory is no longer applied

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
525
directly, being passing the dispersive medium. The transversal section dimension has an effect on
ultrasound wave propagation and longitudinal velocity is strongly affected by transducer frequency below
500 kHz, confirming the importance of adopting during the tests specimen length/wave length above 3.0.
Obstacles occur if there is need to assess wood structures in situ, because in most cases both ends of a
beam or joist, for example, are covered and the measurement cannot be made. Measurements can only be
carried out when placing the transducers parallelly on one side or across the detail facing each other. This
kind of measurement method also provides the local properties of wood Bucur apud Teder et al.
2. Materials and methods
2.1 Materials
For this study 46 sleepers were used. These railway sleepers were classified into three types: Fence post
(Figure 1.a), refuse (Figure 1.b) and firewood (Figure 1.c).
(a) (b) (c)
Figure 1: Railway Sleeper’s classification: (a) fence post; (b) refuse; (c) firewood.
2.2 Tests methods
2.2.1 Railway sleeper’s tabulation
The sleepers were numbered in sequence from 1 to 46 and three dividing lines were marked along the
length of the sleepers, dividing them into 4 equal parts numbered Q1 to Q4 (quadrant 1-4). For each of
these parts, there were four measurements of the height (H) and width (b) (Figure 2.a) and two

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
526
measurements the length (L) for each part (Figure 2.b).
(a) (b)
Figure 2: Measurements of the sleepers dimensions: (a) height and width; (b) length.
To obtain the density of each sleeper, it was still necessary to determine their weights: a scale of maximum
capacity of 200 kg with an accuracy of 10g was used. Any imperfections, fractures, cracks and defects
caused by insects or fungi were measured and cataloged.
2.2.2 Sleepers ultrasound tests
To determine the wave propagation velocities in the main directions of the sleepers ultrasound devices
manufactured by James Instruments and Sylvatest were used, Figure 3.
(a) (b)
Figure 3: Ultrasound devices: (a) James Instruments 150 kHz transducer; (b) Sylvatest 30 kHz transducer.
In the measurement of pulse propagation time along the sleeper Sylvatest device was used with 30kHz
transducers positioned directly (Figure 4) at four points on each sleeper. This device was chosen because
according to DUARTE (3), the minimum distance between the transducers is 30cm, if lower, the readings
of the ultrasonic pulses propagation time do not stabilize.
(a) (b)
Figure 4: Positioning of Sylvatest transducers along the sleeper length: (a) Proposed scheme (b) In loco test.

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
527
To determine the pulse propagation time in the length (L1) of each quadrant ultrasound device of James
Instruments was used, with 150 kHz transducers positioned in indirect way (Figure 5).
(a) (b)
Figure 5: Positioning of the transducers in indirect way MK II along the length of each quadrant of the
sleeper:(a) Proposed scheme (b) In loco test.
Direct measurements were also carried out along the height and width of the sleepers in order to determine
the best direction of the boards cutting.The device MK II was chosen because according to Duarte (3), the
distance between transducers should be 10 to 60cm, otherwise, ultrasonic pulses would not be
stabilize.Using James Instrument device it is indispensable to use a smoothing surface between the
transducers and the wood, with the intention of eliminating the presence of air. In these tests the silicone
type "Dow Corning 111 Lubricant and Valu Sealast" was used.
2.2.3 Sleepers boards ultrasound tests
After cutting the sleepers, size and weight measurements and ultrasound test in at least two pieces of each
sleeper were performed. Each piece was weighed on a scale with a capacity of 50 kg (Figure 6.a). Then
two measurements in length (Figure 6.b), three in width (Figure 6.c) and in thickness measurements were
made. In this case the measurement in the volumes of existing defects was also performed .
(a) (b) (c)
Figure 6: Measurements: (a) weight; (b) length; (c) width.

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
528
The specimens were tested measuring only the velocities along their lengths (Figure 7) using Sylvatest
device and 30 kHz transducers.
Figure 7: Measuring the propagation time of the pulse along the length of the piece
3. Discussion and results
3.1 Sleepers ultrasound tests
In the ultrasound tests performed with Sylvatest and James Instruments devices, various readings were not
obtained due to the fact that the pulses were not transmitted through the air large voids. When there was a
change in reading on a particular point, the reading was performed twice for verification. If this variation
persisted, it was ignored. Table 1 shows the number of measurements in each direction and quantities of
readings obtained. It can be seen that Sylvatest as the James devices obtained approximately 92% of the
readings. It was observed that 60.90% of the readings not obtained by James device occurred along the
width of the sleepers. Due to this fact, it was decided that, to obtain greater use of recycled material, that
the sleepers had to be cut along the width.
Table 1: Number of readings obtained and not obtained in each direction
Direction
Measurements
number of readings
obtained
number of readings not
obtained
James
Sylvatest
James
Sylvatest
James
Sylvatest
Length
552
184
536
171
16
13
Width
552
-
471
-
81
-
Height
552
-
516
-
36
-
Total
1656
184
1523
171
133
13
Table 2 shows the average velocity obtained from Equation (1), for both devices.
Table 2: Average propagation velocities of the pulses in the railway sleepers
Sleeper
Average velocities (m/s)
Sylvatest (30 kHz)
James (150 KHz)
L
L1
b
a
1
4353
3200
2158
1776
2
4711
3892
1677
1633
3
4765
3934
1749
1856

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
529
4
5039
4507
1482
1289
5
4681
4304
1476
1672
6
5324
3476
1691
1564
7
3521
4264
2107
1989
8
4528
4152
1838
1789
9
4602
4259
-
1725
10
4609
4803
1736
1925
11
3969
4085
-
1890
12
4175
4483
-
2280
13
3800
4190
1887
1663
14
3980
4439
1745
1788
15
4563
3260
1671
1540
16
2923
3030
1833
2061
17
3691
3796
1870
2039
18
4143
4593
1558
1885
19
2853
4295
-
-
20
4391
3575
1624
1835
21
4499
4763
1988
1033
22
3809
3097
2124
1402
23
4447
4353
1867
1917
24
4426
2829
949
1692
25
5125
1416
1400
1752
26
4443
2631
4475
1645
27
4519
3468
1195
1546
28
4464
4714
1338
1823
29
3717
3368
1721
1594
30
4640
5011
1652
1344
31
4304
2450
1892
1905
32
4676
4312
1664
1859
33
4184
3432
1867
1986
34
4602
2467
1476
1334
35
4078
3316
-
802
36
2247
3788
1631
1875
37
4132
3168
-
1585
38
3324
1572
1313
1699
39
4806
3705
1292
1293
40
4269
4452
-
1529
41
4704
4806
1272
1419
42
1525
2500
1661
873
43
4346
4821
1820
1687
44
4947
2781
-
1167
45
4294
3173
1461
1166
46
4316
3881
3010
1176

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
530
t
d
V
(1)
Where:
V: velocity of wave propagation, in meters per second;
d: distance traveled by the wave in the material, in meters;
t: time traveled by the wave in the material given by the device in microseconds, which should be
transformed in seconds (x 10-6).
The values of the dynamic longitudinal moduli of elasticity for both devices were obtained by
Equation (2), given by HALABE et al. apud MINÁ (5), Table 3.

d
E
V
(2)
Where:
V: velocity of wave propagation, in meters per second;
Ed: dynamic moduli of elasticity, in mega pascal (x 10-6);
ρ: wood density, in kilograms per cubic meter.
Table 3: Comparison of the longitudinal elasticity moduli obtained by the 2 ultrasonic devices.
Sleeper
E (MPa)
Sleeper
E (MPa)
James
Sylvatest
Difference
James
Sylvatest
Difference
1
12148
22471
10323
24
6273
15357
9083
2
12670
18564
5893
25
1692
22172
20480
3
16829
24697
7867
26
5029
14335
9305
4
12066
15087
3021
27
12620
21417
8797
5
14363
16994
2630
28
18554
16643
-1910
6
11875
27864
15988
29
10356
12611
2255
7
19622
13379
-6242
30
20878
17900
-2.978
8
18020
21427
3406
31
6090
18797
12706
9
15938
18609
2670
32
17479
20546
3067
10
24273
22354
-1918
33
13294
19756
6462
11
17526
16550
-975
34
5193
18070
12876
12
21378
18539
-2839
35
9853
14900
5046
13
16460
13538
-2922
36
11367
4000
-7367
14
16417
13201
-3215
37
10801
18379
7578
15
9664
18929
9265
38
2088
9327
7239
16
10771
10022
-748
39
11086
18656
7570
17
16195
15311
-884
40
20225
18596
-1629
18
17040
13866
-3173
41
20979
20094
-884
19
9476
4180
-5296
42
7295
2713
-4582
20
13243
19973
6730
43
25448
20686
-4762
21
24953
22262
-2691
44
8500
26899
18399
22
9607
14526
4919
45
8295
15186
6891
23
18553
19365
812
46
16453
20345
3891

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
531
The statistical hypothesis of null difference with uncertainty of 95% was used. The range considered was
between 1079 MPa and 5008 MPa, indicating that the elasticity moduli obtained from two devices may not
be considered statistically equal. Then one of these devices did not provide reliable results. As Sylvatest
device showed a small variation during the readings of pulse propagation time and the design of its
transducers allowed a faster work and confidence in the results, it was decided that the longitudinal moduli
of elasticity to be considered in study would be obtained through this device. The values obtained by James
device was discarded because it provided more difficulties during readings, with difficult contact between
transducers and the irregular surface of the sleepers, thus, presenting large variations in readings. Three
sleepers showed longitudinal elastic moduli values below 5000 MPa, probably due to an incorrect
measurement of the waves propagation time, caused by internal defects.
4. Sleepers boards ultrasound tests
Thirty five sleepers were cut, and all the pieces obtained were measured and weighed. Table 4 compares
the values of density and elastic moduli obtained from the sleepers and the sleepers boards. It was also
used the statistical analysis of null difference with uncertainty of 95%. For the density, the interval was
between -14.37 kg/m3 and 18.68 kg/m3 and for the elasticity moduli between -1401 MPa to 74.10 MPa,
indicating that the values obtained in the sleepers before and after being cut can be regarded as statistically
equal.
Table 4: Comparison of sleepers and sleepers boards ultrasonic tests
#
Density (Kg/m3)
E (MPa)
Sleeper
Sleeper board
Difference
Sleeper
Sleeper board
Difference
1
1185
1164
21.10
22471
18058
4412
2
836
823
12.69
18564
18941
-377
3
1087
1086
0.84
24697
23752
944
4
593
627
-33.99
15087
15201
-113
5
775
773
2.16
16994
17844
-850
7
1078
983
95.27
13379
15424
-2044
8
1044
1043
1.51
21427
22276
-849
9
878
894
-15.73
18609
19819
-1210
10
1052
1037
14.20
22354
23491
-1136
11
1050
1137
-87.44
16550
20057
-3507
12
1063
1162
-98.92
18539
22789
-4249
15
908
925
-16.80
18929
20179
-1249
16
1172
1171
0.95
10022
17166
-7143
17
1123
1162
-38.72
15311
17281
-1969
18
807
790
17.58
13866
15152
-1285
20
1035
997
38.38
19973
20193
-219
21
1099
1039
59.71
22262
21062
1200
22
1001
1030
-29.51
14526
17195
-2669
23
978
1014
-35.39
19365
21762
-2397
24
783
787
-3.73
15357
15705
-347

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
532
25
844
794
49.12
22172
17480
4692
27
1048
1066
-17.92
21417
18988
2429
29
912
880
32.67
12611
16804
-4192
30
831
743
87.49
17900
17086
814
31
1014
1014
-0.38
18797
19751
-954
32
939
883
55.80
20546
20475
71
33
1128
1035
92.81
19756
16780
2976
34
853
879
-26.11
18070
21859
-3789
37
1076
1018
57.36
18379
18956
-576
39
807
738
69.49
18656
18676
-20
42
1166
1075
91.37
2713
21493
-18780
43
1094
1082
12.04
20686
20659
27
44
1098
1110
-11.24
26899
29410
-2510
45
823
746
77.31
15186
16718
-1532
46
1091
1105
-13.45
20345
21752
-1407
The sleepers 16 and 42 showed inconsistent and widely dispersed values, therefore were not incorporated
into the analysis. As can be seen in Figure 8, the modulus of elasticity found in the sleeper 42 ultrasound
test was equal to 2713 MPa, and in the sleeper board 21493 MPa, indicating that the first measurement was
impaired by a defect in the board.
Figure 8: Elasticity moduli
5. Conclusionions
In the analysis it is concluded that measuring the elastic modulus can be performed only in one board of
sleeper to simplify and streamline the process of manufacturing glulam railway sleepers. The elasticity
moduli values presented indicated that sleepers ultrasonic measurements is not suitable because that there
are many defects that affect the propagation of the pulses.

Ultrasound classification for recycled glulam sleepers
Carrasco, E. V. M
Int. Journal of Applied Sciences and Engineering Research, Vol. 2, Issue 5, 2013
533
Through this study, the boards with higher elasticities moduli were placed in the upper and lower blades of
the sleepers, due to the fact these boards have best behavior when under tensile stresses. Intermediate
boards should be assembled just concerning with a better use of the boards, because each had, in this case,
different values of width and length.
6. References
1. ASSOCIAÇÃO BRASILEIRA DE NORMAS TÉCNICAS. 1997. NBR-7190: Projeto de
estruturas de madeira. Rio de Janeiro.
2. BUCUR, V.1995. Acoustics of Wood. Florida, USA: Boca Raton: CRC Press.
3. DUARTE, R. de S. 2004. Avaliação do comportamento de ligações com parafusos
auto-atarraxantes em vigas de MLC. 2004. Dissertação (Mestrado em Engenharia de Estruturas) –
Escola de Engenharia – Departamento de Engenharia de Estruturas - Universidade Federal de
Minas Gerais.
4. JAMES INSTRUMENTS INC. Non Destructive Testing Systems. V - Meter instruction manual –
Mark II. Chicago, EUA.
5. MINÁ, A. J. S.; OLIVEIRA, F. G. R.; JÚNIOR, C. C.; DIAS, A. A.; SALES, A. 2004 Avaliação
não-destrutiva de postes de madeira por meio de ultra-som. IPEF - Scientia Forestalis, 65,
188-196.
6. TEDER, M. et alii. 2012. Investigation of the physical-mechanical properties of timber using
ultrasound examination, Journal of Civil Engineering and Management, 795-801.
7. TRINCA and GONÇALVES.2009. Efeito das dimensões da seção transversale da frequência do
transdutor na velocidade de propagação de ondas de ultrassom na madeira. Revista Árvore, 33,
177-184.