Partial Discharge Investigation and Failure Analysis On Distribution Network Using Acousic Camera, 2021, Suwansari
Partial Discharge Investigation and Failure Analysis On Distribution Network Using Acousic Camera, 2021, Suwansari
1518 Pracharat 1 Rd., Wongsawang, Bangsue, @Area Building, 5th floor, 5/15 Nonsi 5 Rd., Chongnonsi,
Bangkok, Thailand Yanava, Bangkok, Thailand
*[email protected] †[email protected]
Abstract— This paper presents practical partial camera’s screen. The maintenance strategies
discharge (PD) investigation by using mobile acoustic according to determined PD type are mentioned.
camera with HV and MV equipment in distribution
network. Four types of PD are observed including II. ACOUSTIC CAMERA FOR PD DETECTION
corona discharge at energized electrode with low and PD acoustic sensor is designed to cover various
high discharge intensity, arcing between live parts or
measurement conditions, where on-site/on-line PD
live part with other components as well as surface or
detection can be applied. In this paper, a commercial
internal discharge with their corresponding PD
patterns. The PD severity as low, medium and high PD acoustic detector named NL camera shown in Fig.
together with its percentage probability can be analyzed 1 is applied to detect PD. The mobile acoustic PD
by applying NL cloud system based AI application. detector equips with 124 low noise microphones,
Laboratory test was performed to confirm the video camera and AI application used to identify and
effectiveness of the PD camera. Practical experiences locate PD site [4].
with different MV equipment are presented i.e. HV and
MV conductor, distribution transformer, pin and post
insulator, lightning arrester, fuse, spacer, and etc. The
results illustrate that the PD acoustic camera could
effectively detect PD problems in power equipment. The
maintenance strategies according to each determined
PD type and measured equipment could be suggested
for efficient maintenance.
Fig. 1. PD detection using acoustic camera
Keywords— acoustic camera, partial discharge, PD
severity, NL cloud system, artificial intelligence The working principle of this PD acoustic detector
is based on receiving sound wave emitted from a
I. INTRODUCTION discharge source. The analyzing tool is called NL
PD diagnosis is one of the most successful cloud system based on AI application. The sensor can
techniques for condition monitoring and HV and MV be used in the frequency range of 2 kHz to 35 kHz as
asset assessment. Unconventional PD detection has shown in Fig. 2 to locate a PD source by using 124
been currently developed for onsite PD investigation low-noise microphones with real-time sound
using different sensing techniques i.e. UHF, HFCT, visualization to detect the emitted sound wave. It is
acoustic sensors [1]. PD detection using acoustic completely inaudible and detectable by human sense.
sensor is one of the most convenient tool to perform
on-site PD diagnostic test with MV/HV equipment,
such as overhead power line, cable termination, cable
joint, transformer, and others [2-3]. This paper aims to
practically investigate the capability of PD acoustic
detector named NL acoustic camera with different
MV equipment along distribution network such as
distribution transformer, MV conductor, pin and post
insulator, lightning arrester, fuse, spacer, and etc. The
Fig. 2. Optimal frequency for PD detection (20-30 kHz).
performance of PD acoustic diagnostic tool is
investigated. The acoustic camera applies a principle III. LABORATORY TESTS
of detecting high frequency sound wave and analyzing
by artificial intelligence or AI cloud system to help A. PD Testing Circuit
analyze the type of PD and the severity that occurs. In Fig. 3, the test circuit is indirect measurement
The PD results are expressed in with Phase-Resolved circuit. It obtains PD signal occurring in opposite
PD (PRPD) patterns. Then, PD type is identified while direction of synchronized voltage source (Um) while
PD location can be detected and shown on the Im is discharge current. The coupling device is an
Authorized licensed use limited to: University of Prince Edward Island. Downloaded on July 04,2021 at 05:57:37 UTC from IEEE Xplore. Restrictions apply.
2021 International Electrical Engineering Congress (iEECON2021) (Full Manuscript)
March 10-12, 2021, Pattaya, THAILAND
182
Authorized licensed use limited to: University of Prince Edward Island. Downloaded on July 04,2021 at 05:57:37 UTC from IEEE Xplore. Restrictions apply.
2021 International Electrical Engineering Congress (iEECON2021) (Full Manuscript)
March 10-12, 2021, Pattaya, THAILAND
In Fig. 9, a low intensity corona discharge was loosen contact at clamp. Since the discharge intensity
found with 100% probability at a recloser in a MV was high, a planned schedule maintenance to tighten
feeder due to loosen contact at it terminal. Since the the loosen clamp should be conducted as soon as
discharge intensity was still low, a planned schedule possible.
maintenance to tighten the loosen bolt and nut at
recloser’s terminal is recommended.
183
Authorized licensed use limited to: University of Prince Edward Island. Downloaded on July 04,2021 at 05:57:37 UTC from IEEE Xplore. Restrictions apply.
2021 International Electrical Engineering Congress (iEECON2021) (Full Manuscript)
March 10-12, 2021, Pattaya, THAILAND
Total
recommended. spans (km) negative positive between internal
discharge discharge components discharge
1 233 3.5 0 2 0 3 7
2 300 4.5 0 2 3 3 8
3 100 1.5 0 1 0 4 5
4 83 2.5 2 2 0 2 6
5 200 6.0 0 3 0 0 3
6 100 3.0 0 3 0 8 11
Fig. 18. Surface discharge at partial insulated conductor touching 7 67 2.0 0 1 1 7 9
tension wire. 8 83 2.5 0 1 0 0 1
9 117 3.5 0 2 0 3 5
In Fig. 19, a high intensity corona positive 10 117 3.5 2 2 0 18 22
discharge was found with 90% probability at a spacer 11 83 2.5 2 1 1 0 4
of a 24 kV feeder due to loosen contact at spacer 12 50 1.5 0 0 0 2 2
terminal. Since the discharge intensity was high, a Total 1533 36.5 6 20 5 50 81
planned schedule maintenance to tighten the loosen Percentage of failure = 5.28%
contact should be conducted as soon as possible.
TABLE II. PD TYPES VERSUS MV EQUIPEMNT.
PD Types Failure Components Failure Causes
corona negative lightning arrester, lossen fitting, lossen
discharge conductor, drop out contact at equipment’s
12 cases fuse, conductor to terminal, pollution
insulator, spacer
corona positive conductor, pin aged/degraded equipment,
Fig. 19. High intensity corona positive at 24 kV spacer. discharge insulator, suspension pollution
20 cases insulator, spacer
In Fig. 20, a high intensity arcing was found at discharge between conductor to pin arcing between conductor
touching point of 24 kV conductor, insulator, components insulator, and components, damaged
preformed armour rods and clamp with 100% 5 cases conductor to /degraded insulation, aged
probability. Because discharge intensity was high and suspension insulator equipment, , pollution
hot spot may occur, a planned schedule maintenance surface/internal conductor, spacer, PD between conductor and
discharge lightning arrester other components & tree,
to check at the contact is requested. 50 cases damaged insulation, aged
equipment, pollution
REFERENCES
[1] L. J. Wang, W. J. Ning, C. Wu, L. H. Wang, H. J. Wang, S.
L. Jia, H. Lü. “Experimental investigation of typical partial
discharge signals in medium-voltage switchgear,” 2014
International Symposium on Discharges and Electrical
Fig. 20. Arcing at touching point of 24 kV conductor.
Insulation in Vacuum (ISDEIV), 28 Sept.-3 Oct. 2014,
Table I summarizes network data such as area, Mumbai, India.
number of spans, distance and PD types, and numbers [2] E. Grossmann, K. Feser. “Sensitive online PD-measurements
of onsite oil/paper-insulated devices by means of optimized
of failure. The percentage of failure is 5.28% out of acoustic emission techniques (AET),” IEEE Transactions on
1533 spans. The highest % failure is surface/internal Power Delivery, vol. 20, no. 1, Jan. 2005.
discharge. In Table II, the PD types, failures causes [3] S. Kornhuber, K. Jani, M. Boltze, G. Valtin. “Partial
and failure components were summarized. discharge (PD) detection on MV switchgears using non-
invasive sensing methods,” 2012 IEEE International
V. CONCLUSIONS Conference on Condition Monitoring and Diagnosis. 23-27
Sept. 2012, Bali, Indonesia.
PD acoustic camera has been developed to detect
[4] Oliver Merilaid. “Air Leak and Partial Discharge
the defect of insulation system. The results show the Frequencies: What You Need to Know,” White Paper NL
capability of PD acoustic system applied to detect Acoustics. 11 May, 2020. www.nlacoustics.co.
various PD sources. The PD acoustic camera is able to
184
Authorized licensed use limited to: University of Prince Edward Island. Downloaded on July 04,2021 at 05:57:37 UTC from IEEE Xplore. Restrictions apply.