Documente Academic
Documente Profesional
Documente Cultură
h i g h l i g h t s
a r t i c l e i n f o a b s t r a c t
Article history: Internal heat of electrical equipment rises due to various reasons such as contact problems, unbalance
Received 19 January 2013 loading, cracks in insulation, defective relays or terminal blocks, etc until eventual failure which can
Accepted 25 July 2013 cause unplanned outages and possible injuries. Therefore, early prevention of thermal abnormalities in
Available online 2 August 2013
electrical equipment is necessary to avoid failure. Nowadays a fast, reliable, non-contact and cost
effective infrared thermographic inspection system is being utilized widely for detecting where and how
Keywords:
internal and external defects occur in equipment. The system collects data for investigating thermal
Infrared thermography
condition through the analysis of thermal image of equipment. The image is captured by infrared camera
Electrical equipment
Preventive/predictive maintenance
without interrupting the running operation of system. Both preventive and predictive maintenance
Intelligent system programs are adopted for defect diagnosis. Actually, the frequency and sequence of thermographic in-
Discriminant analysis spection depend on various factors including thermal condition, expense of a failure, type of production
and age of equipment, etc. The objectives of this article are to discuss about preventive and predictive
maintenance, thermal defects in electrical equipment and also an intelligent thermal defect identica-
tion system for accelerating the predictive defect diagnosis technique. The intelligent system proposes
the application of articial neural network and statistical features to detect the existence of defect within
equipment by monitoring its thermal condition. Using discriminant analysis, the optimum features were
chosen as the inputs of the neural network. The performances of the neural network were compared
with the performances of discriminant analysis classier. The comparison results showed that
discriminant analysis classier produced better performance with accuracy 82.40%.
2013 Elsevier Ltd. All rights reserved.
1359-4311/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.applthermaleng.2013.07.028
A.S.N. Huda, S. Taib / Applied Thermal Engineering 61 (2013) 220e227 221
heat. With the time and age, the condition of the electrical component under the same condition or the maximum allowable
component gets deterioration due to various reasons such as loose temperature of the component [15]. Several standards for DT value
connection, corrosion and insulators crack, etc. As component de- are found such as Electrical Testing Association (NETA) [15], Na-
teriorates, its resistance increases and also generates more heat. tional Fire Protection Association (NFPA) e NFPA 70-B [16] and
The increase in heat energy can cause the failure of electrical InterNational American Society for Testing & Materials (ASTM) e E
equipment and also re may break out. By utilizing thermography [17], etc. Sometimes thermographers prefer to classify thermal
for inspecting the equipment under loading condition, the defec- condition based on their own inspection experiences instead of
tive component can be identied and classied by its level of using the available standards.
severity [12]. Articial neural networks (ANNs) are potentially powerful,
Infrared thermography senses the heat which is produced in robust and adaptive mathematical model for pattern recognition
the electrical equipment. The thermal proles of electrical and classication [18,19]. Nowadays ANNs are being utilized to
equipment and connectors are captured by using a thermal solve complex non-linear real world problems. Several articles
imager. Thermal prole which is also called thermogram consists about the application of articial neural networks (ANN) for elec-
of a thermal picture and a temperature scale. The different colours trical fault diagnosis system using infrared thermography have
of temperature scale represent the different temperature spots of been proposed in the past few years. In these articles, different
the equipment. Heating components are generally detected as red features were used as inputs of ANN. Almeida et al. [20] used a
or brighter coloured areas in an infrared image. A circuit breaker neuro-fuzzy approach for fault diagnosis of lighting arrester while
(CB) thermogram obtained in the present study is shown in Fig. 1. the inputs of the ANN were thermographic and identication var-
In Fig. 1a, the brighter area (R phase) indicates the higher resis- iables. The system achieved resulting index of wrong diagnosis
tance and results overheating of that phase. The probable reason lower than 10% for classifying thermal condition into normal, sus-
of this high resistance could be a bad or dirty connection. Fig. 1b picious and faulty classes. Shai and Hamzah [21] proposed the
shows the temperature prole across the line L0 showing the application of multilayered perceptron network. The RGB colour
variation of temperature in the equipment surface. Thus defects in scale data and temperature data were used to classify internal faults
components are identied and then recorded. Finally, the main- into four classes namely low, intermediate, medium and high. The
tenance programs are planned and carried out before breakdown result obtained 99.38% of accuracy. Rahmani et al. [22] developed
of equipment. an intelligent system applying support vector machine (SVM) as a
Maintenance cost of electrical equipment plays an important classier and 22 image features of Zernike moments for diagnosing
role for reducing the total operating cost of electrical power system. of ground substation faults. The diagnosis was carried out on 20
The maintenance cost can be increased due to the lack of infor- thermal images and was able to obtain 68.42% of accuracy. Smed-
mation about the maintenance scheduling and kind of mainte- berg [23] and Wretman [24] proposed an intelligent classication
nance needed for the equipment. Typically, the operating system based on ANN for diagnosing three-phase fuses and
performance of electrical system is not monitored periodically different forms of connections problems. The four input parameters
which can play an important role in early prevention of a fault and of ANN were used which are absolute maximum temperature,
for increasing the life of equipment. Instead, the schedule for relative maximum temperature, mean temperature difference
maintenance is determined by using the failure history of equip- compared to the other regions of the image and histogram distance
ment or according to the experiences of maintenance personnel. to the other regions of the image. The test error rate using all four
Infrared thermography is used to monitor the performance of feature parameters as input of ANN was 9.5% and the error rate
equipment which offers a complete information regarding the using only the histogram distance as input was 31.2%. The set of
operating status of equipment for increasing production, data was 74 infrared images. One of the disadvantages of working
decreasing or removing unnecessary repairs, preventing eventual with a small dataset is that the reliability of the results can some-
failure and decreasing the maintenance cost of equipment. times be questionable.
Usually, applying infrared thermography technology, the elec- This paper will investigate the application of infrared ther-
trical thermal defect and level of defect are identied by moni- mography technology as predictive/preventive maintenance pro-
toring its Delta T (DT) criteria [13]. The technique is widely known gram to identify the presence of thermal defect in electrical
as qualitative based temperature measurement system [14]. The equipment. The technique uses multilayered perceptron network,
Delta T criteria of a component is dened as an increase in tem- statistical features and discriminant analysis classier to charac-
perature value above the temperature of a reference value, which is terize the thermal status of hotspot into defect and no defect
typically the ambient temperature, temperature of a similar categories. Discriminant analysis classier was utilized to select
suitable features and also to build the linear regression model of various reasons such as contact problems, unbalance current dis-
each category using suitable features. tributions, cracks in insulators, defective relays or terminal blocks,
etc. Some of the examples of the defects in electrical system which
2. Preventive/predictive maintenance can be revealed by thermographic technology are mentioned here.
Typically, for maintaining the equipment, two types of ap- 3.1. Transformer
proaches, either run-to-failure or preventive maintenance are used.
The run-to-failure approach is simple and straightforward, where Heat is generated due to the losses in transformer and temper-
there is no outow of money on maintenance purpose before ature increases within the core and windings. The transformer is
eventual failure of equipment. The approach waits for equipment generally cooled by either oil or air and the operating temperature is
failure before any maintenance action is taken. Therefore, the higher than ambient. The cooling system helps to maintain the rise
method is more expensive. Thermography based diagnosis system of temperature within tolerable limits. Usually, the thermographic
allows the preventive/predictive maintenance for early prevention inspection is utilized for thermal defects diagnosis of oil-lled
of equipment failure without interrupting the running operation transformers instead of dry transformers. Since, normally, the coil
and thus saves money. temperatures of the dry transformers are so much higher than
According to the historical data in USA [25], effective use of ambient, it is not easy to distinguish the internal defect using ther-
preventive/predictive maintenance will reduce about 33%e50% of mography before it becomes serious. Therefore, other monitoring
maintenance cost that are wasted by most of the manufacturing tools including built-in temperature and pressure gauge can
and production plants and also introduced the following benets perform more accurate measurement for the dry transformer. In the
from various aspects: oil-lled transformers, thermographic inspection usually detects
the defects occurring at the primary and secondary bushing con-
1. Reduced unscheduled downtime or system ultimate failure nections, cooling ns, cooling fans and internal bushing connec-
2. Increased the proper utilization of manpower tions, etc. The problems are detected visually by the following ways:
3. Increased the production capacity of the plant
4. Reduced the maintenance expenditure of equipment U Primary and secondary bushing connections e The loose or dirty
5. Increased the life time of equipment connections, unbalance and overloading problems can cause
overheating in the connections. Thermographic inspections
show hotspots at the defected phase connection compared to
2.1. Preventive maintenance other phases.
U Cooling ns e In an oil-lled transformer, cooling is provided by
Maintenance of electrical equipment according to the statistical the circulation of oil. The oil gets heated due to the heat generated
or historical information of operating capacity, failure history, in the core and windings. Therefore, cooling ns normally appear
mean-time-to-failure (MTTF) instead of tracking equipment per- warm compared to the inactive cooling n. From the thermal
formance is referred as preventive maintenance [25]. Preventive image, if it is seen that one or more ns have comparatively low
maintenance program schedules repairing and rebuilding activities temperature or cool compared to other warm ns, then it can be
for equipment. Suppose, an electrical equipment normally runs 10 concluded that the circulation of oil is being limited in those ns,
months before needed any repairing. Using preventive technique, probably due to the low oil level.
the equipment will be removed from service and rebuilt after 10 U Cooling fans e In higher rating transformer, where more heat
months of operation. In this example, the schedule of repairing of will be generated, cooling fans are used to increase the cooling
equipment is predened which is determined by maintenance rate. In general, the fan in action is normally warm and thus
personnel. Now, if there is no need to rebuild after 10 months, then, thermogram will show whether it is working or not.
there would be wasted of labour and material. Again, if the equip- U Internal bushing connections e If there is any defect in the con-
ment fails before 10 months, then, it is necessary to x problem nections, it will be higher temperature than surface temperature.
after failure which is usually more expensive than schedule main-
tenance. This is a just random decision which is not supported by
any kind of realistic data. 3.2. Defect in circuit breaker
2.2. Predictive maintenance A circuit breaker (CB) always senses the heat produced by the
current. If it nds the abnormality of heat, instantly opens the cir-
Maintenance of electrical equipment based on the direct mo- cuit automatically and protects the electrical equipment from
nitoring of the actual operating condition, regularly collecting damage. Thermographic inspection has the advantages to nd the
measurements, efciency, heat distribution and other indicators abnormalities by analysis the surface temperature variation of cir-
instead of depending on statistical or historical data is called pre- cuit breaker and can prevent these abnormalities earlier.
dictive maintenance. The program schedules all maintenance ac-
tivities based on factual data and if necessary, repairs before the 3.3. Fuse connection
system failure.
Early detection of imminent equipment failures and preventing Fuses are used to protect the system against overloading con-
them before critical condition signicantly reduce downtime, cost dition. Overloading condition in a system can increase the fuse
for maintenance, maximize uptime. connection heating and moreover, the fuse clips can be overheated
due to the loose and dirty connections.
3. Defect in electrical equipment
3.4. Defect in contact
Electrical installations such as transformer, circuit breaker, fuse
and capacitor, etc can cause faults when its internal temperature Electrical connection switches, breakers and other reclosing
reaches at an abnormal level. Most of the anomalies are due to devices operate with contacts. If these contacts become loose,
A.S.N. Huda, S. Taib / Applied Thermal Engineering 61 (2013) 220e227 223
oxidized, dirty or anyhow increase in resistance, then the defect can Table 1
be occurred in the contacts. DT specications of defect and no defect classes with recommended actions.
Class DT ( C) Recommended actions
3.5. Insulation problem Defect DT 15 Major discrepancy; repair immediately
(Priority 1)
Insulation failure can cause short circuit current between two 5 < DT < 15 Probable deciency; repair as time permits
conductors. Overcurrent generates overheating and allows the circuit (Priority 2)
No defect DT 5 Minor overheating; warrants investigation
breaker or fuse to open. Poor insulation can also cause overheating.
(Priority 3)
distribution of each feature using multivariate test. From the 13 0.24 and defect class is located at the separation point which is
features, 10 optimum features have been chosen using the multi- greater than 0.24.
variate test as suitable features for defect recognition. The ignored Apart from determining the suitable features, the discriminant
features which have p-value distribution more than 0.05 are analysis is also able to predict the linear regression model of each
highlighted in Table 4. The 10 optimum features are X2, Y2, Y4, X11, class taking the account of classication function coefcients and to
X13, X15, X16, X17, Y11 and Y12. classify the thermal class (i.e. determine the number of correctly
Table 5 presents the suitable features with its discriminant co- classied data, the number of misclassied data and the accuracy
efcients which will be used during the formation of discriminant for each class). Table 7 shows the classication function coefcients
function. of each class for all selected features.
The discriminant function Zp value is obtained by adding the Now the linear predictive model of each class is obtained by
sum of multiplication between discriminant coefcient and domi- adding the sum of multiplication between classication function
nant features values. A function has a multiple root (i.e. a root with coefcient and dominant features values. The predictive model for
multiplicity greater than one) in the complex numbers, if and only defect class, Dp can be written as
if its discriminant is zero [30]. From Table 5, the discriminant
function is derived as follows:
Dp 1:806X2 0:153X11 0:02X13 0:0342X15 1:998X16
0:034X17 4:02 106 Y2 7:28 1010 Y4 1:09
Zp 0:021X2 0:017X11 0:006X13 0:0321X15 0:068X16
105 Y11 3:01 107 Y12 46:08
0:001X17 1:551 (2)
(1)
and the predictive model for no defect class, Np can be written as
Based on the Zp value, the class mean called centroid is deter-
mined. Table 6 shows the centroid (mid-point) for each thermal
class. Then, the type and range of each class are found by deter- Np 1:768X2 0:123X11 0:01X13 0:908X15 2:118X16
mining the separation points between the classes. The separation
point must be established, since the value can be used to identify 0:035X17 4:84 106 Y2 6:69 1010 Y4 1
the range for each class of the thermal status. 105 Y11 3:20 107 Y12 43:797
From Table 6, it is found that the centroid of no defect class is
(3)
less than defect class. So, the separation point between the no
defect and defect class is found [(1.122 0.642)/2] 0.24. The Finally, the discriminant analysis is used to classify the thermal
no defect class is located at the separation point which is less than status of hotspots into two classes namely no defect and defect.
Table 3
Results for univariate test.
Table 5
Discriminant function coefcients.
w 1ij
Feature Coefcient Feature Coefcient
z1 w2jk
X2 0.021 Y2 0.000
X11 0.017 Y4 0.000
X13 0.006 Y11 0.000 y1
X15 0.321 Y12 0.000
X16 0.068 Constant 1.551 z2
X17 0.001
Table 6 ym
Centroid for each thermal class.
Table 10
Training phase performance.
q
Specificity 100% (9)
qx Training data folds Training phase
X2 1.806 1.768
X11 0.153 0.123 Table 11
X13 0.02 0.01 Testing phase performance.
X15 0.342 0.908 Testing data fold Testing phase
X16 1.998 2.118
X17 0.034 0.035 Sensitivity Specicity Accuracy
Y2 4.02E-06 4.84E-06
1st fold 87.50 72.22 82.0
Y4 7.28E-10 6.69E-10
2nd fold 84.37 85.55 83.0
Y11 1.09E-05 1.00E-05
3rd fold 73.43 91.66 80.0
Y12 3.01E-07 3.20E-07
4th fold 85.71 70.27 80.0
(Constant) 46.08 43.767
5th fold 88.88 56.75 77.0
Average 83.98 75.29 80.40
Table 8
Classication using discriminant analysis.
y total number of defect condition classied as no defect [3] Z. Ge, X. Du, L. Yang, Y. Yang, Y. Li, Y. Jin, Performance monitoring of direct air-
cooled power generating unit with infrared thermography, Applied Thermal
condition
Engineering 31 (2011) 418e424.
x total number of no defect condition classied as defect [4] B.B. Lahiri, S. Bagavathiappan, T. Jayakumar, J. Philip, Medical applications of
condition infrared thermography: a review, Infrared Physics & Technology 55 (2012)
221e235.
[5] C. Bougriou, R. Bessa, R. Le Gall, J.C. Solecki, Measurement of the temperature
Tables 10 and 11 show the training and testing phase performance distribution on a circular plane n by infrared thermography technique,
of MLP network trained using LM algorithm. In the training phase, Applied Thermal Engineering 24 (2004) 813e825.
MLP network produced 88.20%, 89.789% and 85.418% of accuracy, [6] C.A. Balaras, A.A. Argiriou, Infrared thermography for buildings diagnostics,
Energy and Buildings 34 (2002) 171e183.
sensitivity and specicity respectively. From all the folds, the results [7] R. Royo, M.A. Albertos-Arranz, J.A. Crcel-Cubas, J. Paya, Thermographic
using second fold as training dataset provided the best performance study of the preheating plugs in diesel engines, Applied Thermal Engineering
regarding accuracy, sensitivity and specicity which are 97.75%, 37 (2012) 412e419.
[8] M. Manana, A. Arroyo, A. Ortiz, C.J. Renedo, S. Perez, F. Delgado, Field winding
98.88% and 95.89% respectively. fault diagnosis in DC motors during manufacturing using thermal monitoring,
Similarly in the testing phase, the same network provided ac- Applied Thermal Engineering 31 (2011) 978e983.
curacy 80.40%, sensitivity 83.98% and specicity 75.29%. Therefore, [9] D.J. Titman, Applications of thermography in non-destructive testing of
structures, NDT&E International 34 (2001) 149e154.
it is stated that the MLP network trained using LM algorithm and
[10] A.R. Al-Kassir, J. Fernandez, F.V. Tinaut, F. Castro, Thermographic study of
also the proposed 10 features have high capacity to detect the energetic equipments, Applied Thermal Engineering 25 (2005) 183e190.
defect in equipment. [11] M.S. Jadin, Soib Taib, Recent progress in diagnosing the reliability of electrical
equipment by using infrared thermography, Infrared Physics & Technology 55
Table 12 shows the performance comparison between discrim-
(2012) 236e245.
inant analysis classier and multilayered perceptron network. The [12] Improving Electrical System Reliability with Infrared Thermography, www.
DA produced better specicity and accuracy compared to MLP infraredelectrical.com/index.htm.
network, while MLP network achieved higher sensitivity. The value [13] P. Gill, Electrical Power Equipment Maintenance and Testing, Marcel Dekker
Inc., New York, 1998, pp. 385e391.
of sensitivity, specicity and accuracy difference between two [14] Y.C. Chou, L. Yao, Automatic diagnosis system of electrical equipment using
classier are 2.88%, 9.31% and 2% respectively. infrared thermography, in: International Conference of Soft Computing and
Pattern Recognition, Malacca, December 4e7, 2009, 155e160.
[15] Standard for Infrared Inspection of Electrical Systems & Rotating Equipment,
8. Conclusion Infraspection Institute, 2008, 2011. http://www.armco-inspections.com/les/
ir/Electrical%20Rotating%20Std.pdf (accessed 11.06.11).
[17] ASTM, ASTM E 1934: Standard Guide for Examining Electrical and Mechanical
In this paper, an intelligent predictive diagnosis system for Equipment with Infrared Thermography, ASTM International, West Con-
detecting invisible thermal defects in electrical equipment by shohocken, Pennsylvania, 2005.
infrared thermography technology is proposed. The diagnosis sys- [16] NFPA, NFPA 70B: Recommended Practice for Electrical Equipment Mainte-
nance, National Fire Protection Association, Quincy, Massachusetts, 2006.
tem employed the multilayered perceptron network with Leven- [18] A. Darabi, X. Maldague, Neural network based defect detection and depth
bergeMarquardt training algorithm to classify the thermal estimation in TNDE, NDT&E International 35 (2002) 165e175.
conditions of equipments into two classes namely no defect and [19] G. Manduchi, S. Marinetti, P. Bison, E. Grinzato, Application of neural network
computing to thermal non-destructive evaluation, Neural Computing & Ap-
defect. Total fteen features including rst, second order statistical
plications 9 (1997) 148e157.
features and also mutual features (i.e. the numerical difference [20] C.A.L. Almeida, A.P. Braga, S. Nascimento, V. Paiva, H.J.A. Martins, R. Torres,
between the value of hot and reference features) were calculated W.M. Caminhas, Intelligent thermographic diagnostic applied to surge ar-
resters: a new approach, IEEE Transactions on Power Delivery 24 (2) (2009).
from sample images. Using discriminant analysis feature selection
[21] M.A. Shai, N. Hamzah, Internal fault classication using articial neural
method, 10 optimum features were chosen as the inputs of the network, in: Proceedings of 4th International Power Engineering and Opti-
neural network. The performance of the neural network was mization Conference, Shah Alam, June 23e24, 2010, 352e357.
compared with the discriminant analysis classier. The comparison [22] A. Rahmani, J. Haddadnia, O. Seryasat, Intelligent fault detection of electrical
equipment in ground substations using thermo vision technique, in: Pro-
results showed that the discriminant analysis classier produced ceedings of 2nd International Asia Conference on Mechanical and Electronics
better performance with 82.40% of accuracy. By developing the Engineering, vol. 2, August 1e3, 2010, pp. 150e154. Kyoto.
system, the technical department can schedule the recommended [23] M. Smedberg, Thermographic Decision Support e Detecting and Classifying
Faults in Infrared Images. Master thesis, Royal Institute of Technology,
actions. This will play a signicant role for fast and reliable in- Stockholm, Sweden, 2006.
spection to prevent equipment failure potentially. [24] D. Wretman, Finding Regions of Interest in a Decision Support System for
Analysis of Infrared Images. Master thesis, Royal Institute of Technology,
Stockholm, Sweden, 2006.
Acknowledgements [25] The basics of predictive/preventive maintenance. Web: http://www.cedesa.
com.mx/pdf/uke/ukeTi30_serie_nota.pdf (accessed 02.08.11).
[26] Why thermography is good for your business. Web: http://www.newark.com/
The authors gratefully acknowledge the contribution of Dr. pdfs/techarticles/Why_Thermography_is_Good_for_Business.pdf (accessed
Kamarul and Mr. Shawal, Universiti Malaysia Pahang for helping in 01.08.12).
thermal image acquisition. The work is supported by Fundamental [27] S. Kabir, Imaging-based detection of AAR induced map-crack damage in
concrete structure, NDT&E International 43 (2010) 461e469.
Research Grant Scheme (203/PELECT/6071211), Universiti Sains
[28] A.S.N. Huda, S. Taib, M.S. Jadin, D. Ishak, A semi-automatic approach for
Malaysia. thermographic inspection of electrical installations within buildings, Energy
and Buildings 55 (2012) 585e591.
[29] D.J. Hand, Discrimination and Classication, John Wiley and Sons, New York,
References 1981.
[30] Discriminant. Web: http://en.wikipedia.org/wiki/Discriminant (accessed
[1] M.R. Clark, D.M. McCann, M.C. Forde, Application of infrared thermography to 25.03.12).
the non destructive testing of concrete and masonry bridges, NDT&E Inter- [31] S. Haykin, Neural Networks: a Comprehensive Foundation, second ed., Pren-
national 36 (2003) 265e275. tice Hall, New Jersey, 1988.
[2] X.P.V. Maldague, et al., Theory and Practice of Infrared Technology for Non- [32] C. Schaffer, Selecting a classication method by cross validation, Machine
destructive Testing, Wiley, New York, 2001. Learning 13 (1993) 135e143.