EVALUATION OF THE DEFECTS IN TYPE-I CNG CYLINDER USING

ACOUSTIC EMISSION TECHNIQUE

Suparerk Sirivedin (1), Tonphong Kaewkongka(2), Jirapong Lim(3)

(1)
King Mongkut’s University of Technology North Bangkok
Bangsue, Bangkok 10800, Thailand Email : ssv@kmutnb.ac.th
(2)
Department of Physics, Chulalongkorn University
Pathumwan, Bangkok 10330, Thailand Email : tonphong.k@chula.ac.th
(3)
Department of Production Engineering,
King Mongkut’s University of Technology North Bangkok
Bangsue, Bangkok 10800, Thailand Email : jir@kmutnb.ac.th

the vehicles. Therefore, this test method is impractical due

ABSTRACT
Defects on compressed natural gas (CNG) storage
cylinders can often result in damage of costly cylinders.
Acoustic emission (AE) testing was performed on type I
steel cylinder. These studies suggested that the elastic
stress waves or acoustic emissions generated during the
microscopic dislocation can propagate across the storage
cylinder surface to be detectable by the sensors attached at
the ends of the cylinder. This paper presents a preliminary
study on the signal transmission and propagation of
acoustic emission (AE) signatures across the cylinder with
and without a predefined surface crack. During the test, a
gradual increase in hydrostatic pressure up to 400 bars was
applied to the cylinder. The AE signals were recorded as a
function of time and the increased pressure. It is therefore
obvious that AE can be used to capture the defects due to
crack propagation.
Index Terms— CNG Cylinder, Acoustic Emission,
Surface Crack Propagation,
1. INTRODUCTION
To assure the safety of vehicles using compressed
natural gas (CNG) is an important issue in Thailand. There
have been several accidents from CNG-cylinder explosions
in the past. The periodical inspection requires the first
inspection after the cylinder has been using for 3 years.
However, the inspection is normally carried out by visual
inspection. Consequently, the risk is relied on the
inspection by licensed professional engineers since the
internal defects cannot be examined.
to time-consuming and high inspection cost.
Recently, acoustic emission (AE) technique has been
standardized for gas cylinder inspection as described in
ISO/DIS 16148.2 [1]. Using AE technique, the inspection
can be quickly performed without disassembling the
cylinder from vehicle. However, Craig Webster [2] has
found out that the AE test based on ISO/DIS 16148.2 is
unable to evaluate defects in Type I steel cylinder. Mark P.
Connelly and Han Dinh [3] proposed the AE technique
called “Source Location Acoustic Monitoring (SLAM)” to
inspect US Postal Service vehicles. They concluded that
the inspection cost for acoustic emission SLAM test is
about half of the cost for conventional hydrostatic test.
They also claimed that the SLAM test has the ability to
locate both the external/internal flaws.
Stephen J. Hudak [4] has used AE technique to
evaluate remaining life-time of the cylinders. He found
that at the initial crack of about 25 percent of the wall
thickness the cylinder has the remaining life-time for 95
years!, while the deeper crack of 40 percent of the wall
thickness the cylinder is last for 5 years under the cyclic
load caused by daily refueling pressure ranging from 300
to 3,000 psi.
A major benefit of AE inspection is that it can allow
the whole volume of the cylinder to be tested non-
intrusively in a pressurized operating condition. Generally,
the global AE inspection is used to identify areas with the
presence of defect problems and other NDT methods are
then used to identify more precisely location and the root
cause of the AE sources.

The fully inspection can be accomplished by

ultrasonic test but this method requires removal of the
cylinder from

The aim of this research is to study the feasibility of condition

monitoring of CNG storage cylinder using AE sensors and its
propagation of the stress waves through the type I CNG steel cylinder
with and without surface crack.
 2. EXPERIMENTAL APPARATUS

Acoustic emission is a natural phenomenon of

stress wave generation and propagation spontaneously
when a material is subjected under stress. Plastic
deformation and growth cracks are the primary sources of
acoustic emission in metals. The acoustic signal can be
detected by a piezoelectric transducer, which converts the
mechanical energy carried by the elastic wave into an
electrical signal as shown in Figure 1.
The AE inspection is usually carried out during a
controlled loading or pressurization of the material or
specimen. A conventional AE parameter, AE hit rate is
used to identify the presence of the acoustic emission
activities produced during the microscopic failures.

Figure 2 SIMPAL system from Holroyd

The AE signal enveloper converts the amplified

Figure 1 Acoustic emission system
The resonant type of acoustic emission transducers
(Holroyd Instruments, UK: Model ASS-1) are mounted on
the circumference of the surface of the valve.
signal to ‘rf’ signal which is digitized to personal computer
for further data logging and processing. The SIMPAL
system equipment (Holroyd Instruments, UK) was used as
acoustic emission data logger and processing in this work
(see Figure 2). The AE data can then be captured to data
logger (PC). In the experiment, four AE sensors were
attached to the cylinder without crack (see Figure 3), and
with a surface crack (see Figure 4). Characteristic and
dimensions of surface crack are shown in Figure 5.

It provides the 100 kHz of resonant frequency which

responses well with the material degradation and
microscopic crack initiation. The acquired signal is then
amplified with 60 dB gain pre-amplifier.

Figure 3 Installation of AE sensors on CNG cylinder

 NGV--Node 0 ((11//09//2008 ,, 1::09::01)
Log Hiitt Rate vs Recorrded Date/Tiime
1.00 11.00
10.00
9.00
8.00
7.00
6.00
5.00
4.00

Figure 4 Installation of AE sensors on CNG cylinder 3.00

2.00
1.00
with a longitudinal surface crack 0.00
-1.00

The cylinder used in this experiment has a capacity of Figure 6 Installation of CNG cylinder in a test chamber
60 litres. It has a diameter of 300 mm, the length of 990 .10

mm and the averaged wall thickness of 8 mm. The cylinder

3. EXPERIMENTAL RESULTS
was installed in a chamber of a high-pressure hydrostatic
testing machine (see Figure 6).
3.1 Hydrostatic test on CNG cylinder without crack
Recorded Date/Time

A hydrostatic test with an applied pressure up to

300 bars was performed on a CNG cylinder without a
surface crack. The parametric input indicates the pressure
.01
level (with the maximum voltage 10 Volt corresponding to
300 bar and minimum 0 Volt corresponding to 0 bar). The
first acoustic emission hit rate (AE Hit Rate) was detected
at the applied pressure of 90 bars (see Fig. 7).

Figure 5 Surface crack characteristic

The threshold was set at 40 dB based on observed Pa

Lo g H it R a te

ra

background noise. A gradual increase in hydrostatic m

e tr
ic
pressure up to 400 bars was applied to the cylinder.
1 1 /9 /2 5 5 1 14 :3 9:20

Figure 7 AE hit rate v.s. pressure for CNG cylinder

without surface crack
crack. A crack length, l = 30 mm and a crack depth, a = 2 mm was
introduced in a longitudinal direction of CNG cylinder by hand grinder.
The accumulated AE counts of 7 times were detected at the applied
pressure between 120-140 bars.

3.2 Hydrostatic test on CNG cylinder with a crack

A hydrostatic test with an applied pressure up to 300

bars was performed on a CNG cylinder with a surface
 Figure 8 AE counts v.s. pressure for CNG cylinder
with a surface crack at hydrostatic test
3.3 Burst test on CNG cylinder with a surface crack
A burst test was performed on a CNG cylinder with a
surface crack. A crack length, l = 50 mm and a crack
depth, a = 6 mm was introduced in a longitudinal direction
on the same CNG cylinder, which enlarged a crack size by
hand grinder. At the applied pressure between 100-170
bars, more than 100 AE counts were detected.
AE signal was not detected after 170 bars until the
applied pressure was increased to 310 bars. The
accumulated AE counts more than 2 million times were
detected at the applied pressure between 310-340 bars. The
CNG cylinder leaked at the location of surface crack at the
applied pressure of 340 bars.
Figure 9 AE counts v.s. pressure for CNG cylinder
with a surface crack at burst test
4. CONCLUSIONS
Due to “Kaiser effect” in which the material
recognizes its service stress, the applied pressure to cause
AE hit must be greater than the service pressure. When the
surface crack was introduced onto the CNG cylinder, a
greater number of acoustic events are likely to occur than
CNG cylinder without crack. As the surface crack size is
increasing, a large number of acoustic events were detected
at the same pressure level. Therefore, acoustic emission
technique can be used to evaluate and assess how severe of
the defects in Type-I CNG cylinder.
5. ACKNOWLEDGEMENTS
The authors would like to thank Energy Technique
Development Institute, Department of Energy Business of
Thailand for testing facilities. They would also like to
thank Iron & Steel Institute of Thailand for financial
support throughout this paper.
 6. REFERENCES

[1] ISO/DIS 16148.2, “Gas cylinders-Refillable seamless steel gas

cylinder- Acoustic emission examination for periodic inspection”,
International Organization for Standardization, 2006.
[2] C. Webster, “Development of Non-Destructive Evaluation
(NDE) Techniques for CNG Fuel Tanks”, Report for
Transportation Development Centre of Transport Canada,
Canada, 2007.
[3] M.P. Connolly, H. Dinh, “Fleet Inspection of Compressed
Natural Gas Cylinders for Natural Gas Vehicles Using Source
Location Acoustic Monitoring” SAE Technical Paper Series No.
961174, SAE International, 1996.
[4] S.J. Jr. Hudak, “Assuring the Safety of Natural Gas Vehicles”
Technology Today Magazine, Sept. 1991.