International Journal of Latest Research in Science and Technology ISSN (Online):2278-5299

Vol.1,Issue 1 :25-29 ,May -June(2012)

http://www.mnkjournals.com/ijlrst.htm

Fuzzy-Logic Applications in Transformer Diagnosis Using

Individual and Total Dissolved Key Gas Concentrations
Hasmat Malik1, R.K. Jarial2, H.M. Rai3
1,2Electrical Engineering Department, NIT Hamirpur (HP)-177005,
3
Ex- Prof. NIT Kurukshetra, India
1
[email protected], [email protected], [email protected]

Abstract -The gases generated in oil filled transformer can be used for determination of incipient faults. Dissolved gas analysis (DGA) of
transformer oil has been one of the most power full methods to detect the faults. The various methods such as liquid chromatography,
acoustic analysis, and transformer function techniques are require some experience to interpret observations. The researchers have used
artificial intelligence (AI) approach to encode these diagnostic techniques. This paper presents an expert system using AI techniques which
can diagnose multiple faults in a transformer theoretical and practical fuzzy-logic information model. We also concluded by identifying
limitations, recent advances and promising future research directions over seventy and more power transformers.

Keywords - Transformer, Dissolved gas analysis, Fuzzy-logic, diagnostic, Expert system

I. INTRODUCTION condition of the cellulose material within an in-service

transformer. The various Artificial Intelligence (AI)
Reliable and quality power is need of the hour for the Techniques [8-17] that have been used by the researchers in
economic development of a country. For providing reliable the past have been considered and some conclusions have
electrical energy, it is very necessary to have highly reliable been drawn based upon the observations.
associated electrical equipments in electrical power stations. Fuzzy logic with fuzzy-logic IF-THEN rules firing
The transformer, being key element in the transmission and scheme have been used to analyse the health condition of
distribution of electrical energy need special care [1]. For transformers. The extracted data from DGA has been used
improving its satisfactory operation under normal power for input and output membership function and the best
system conditions, the system reliability is of utmost parameters for this network have been presented graphically.
importance. Large oil filled electrical power equipments, Finally result given by fuzzy logic expert system has been
such as transformers and reactors, is a critical element of an compared with IEEE/ANSI standards.
electrical power system. The reliability and continuous II. FUZZY-LOGIC
performance of these equipments is then vital key to the Fuzzy-Logic (FL) is a relatively new artificial
profitable generation and transmissions. To minimize the intelligence technique. FL means approximate reasoning,
capital expenditure of the electrical power system, it is very information granulation, computing with logical words and
common to operate these equipments at or close to the so on. Fuzzy systems are rule-based systems that are
limits of their design parameters [2]. Power transformers constructed from a collection of linguistic rules. FL is a
are the vital link of the power system. Monitoring and convenient way to map an input space to an output space. It
diagnosis techniques are essential to decrease maintenance provides mathematical strength to the emulation of certain
and improve reliability of equipment [3]. Due to highly perceptual and linguistic attributes associated with human
competitive electrical energy market it is required to cognition. The theory of fuzzy logic provides an inference
enhance the system’s reliability and availability with cost mechanism under cognitive uncertainty. This view of
effectiveness. To maintain higher system reliability, its key network as a parameterized function will be the basis for
components such as power transformer is required to applying standard function optimization methods to solve
undergo periodic diagnostics and reduce its losses for the problem of neural network training.
effective utilization of capital resources.
In this paper an attempt has been made to present the A. Expert System for Transformer Diagnosis:
various conventional and non-conventional DGA methods
The schematic diagram of fuzzy logic transformer
[1, 4, 7] used by various agencies and utilities to find the
diagnosis (FLTD) expert system is shown in Fig.1. FLTD is

25
 Malik et. al., International Journal of Latest Research in Science and Technology.

a novel fuzzy-based approach that deals with heterogeneous transformer so as to compensate the conflict between
data of both linguistic and numeric types, imprecise, vague excessive cost due to over sampling and neglected danger
in-formation, and concepts encountered in the mechanical- owing to long sampling period. In general, TDCG uses the
fit process and facilitate the expression of the reasoning sum of the 6 key gas values (as formula 1) and the TDCG
process of an experienced observer with minimal rules. gas generation rate to determine the operating procedure and
The Fuzzy Logic Transformer diagnosis process represents predict suitable oil sampling interval as shown in Table 1
a fuzzy-logic-based complete transformer diagnosis process [19].
comprising the following three phases: Table 1
phase I: tentative selection of total dissolved combustible Action Based on Dissolved Combustible Gas as per [19]
gas (TDCG), and TDCG_rate; Status/ TDCG_Rates Sampling Intervals and Operating
TDCG (ppm/day) Procedures for Gas Generation Rates
phase II: mechanical-fit process; Sampling Operating Procedure
phase III: estimation and optimization of Sampling Interval Interval (SI) (OP)
(SI), and operating Procedure (OP). Condition 1 <10 6 Monthly Continue normal
A general schematic of FLTD representing the second <720 (SIA) operation (OPA)
10-30 Quarterly
phase is shown in Fig. 1. It involves the following phases: (SIQ)
1) Fuzzification; >30 Monthly Exercise caution.
2) Knowledge representation; (SIM) Analise individual
3) Inference scheme; gases to find cause.
4) Defuzzification. Determine load
dependant (OPB)
Condition 2 <10 Quarterly Exercise caution.
721-1920 (SIQ) Analise individual
10-30 Monthly gases to find cause.
>30 (SIM) Determine load
dependant (OPB)
Condition 3 <10 Monthly Exercise caution.
1921-4630 (SIM) Analise individual
10-30 Weekly gases to find cause.
>30 (SIW) Plan outage. (OPC)
Condition 4 <10 Weekly Exercise caution.
>4630 (SIW) Analise individual
10-30 Daily gases to find cause.
(SID) Plan outage. (OPC)
>30 Consider removal from
service

Figure 1: General scheme of FLTMWE Although the TDCG method is widely used in solving
fault diagnosis problem, but in the certain cases, it is very
TDCG = C2H2 + C2H4 + H2 + CH4 + C2H6 + CO (1) and hard to determine the correct group of the TDCG value
especially when the TDCG value falls near the boundary
TDCG_Rate = (St-So)/T (2) line as shown in the TDCG rules set in Table 1. The fuzzy
Where St = Current TDCG; and So = Previous TDCG logic technique is advantages in solving this problem,
T = Time in days. which is explained in section III.
for example, data taken of sample no. 5 from Table 3 as For the TDCG diagnostic method, the sum of the six
H2=16, CO=935, C2H4=28, C2H6=6, CH4=12, C2H2=29, fault gases and the gas generation rate are required to
then TDCG=1026 (using formula 1) determine the health condition of a power transformer.
Assume that the previous TDCG is 630 for sample no. 5, Based on these results, a fuzzy model is developed using its
and the duration from last sampling date is 30 days. The input variable, TDCG, and TDCG_Rate. Membership
TDCG rate can be calculated using formula 2 as: functions (MF) for input variables are established based on
TDCG_Rate= (1026-630)/30 the variation of TDCG, and TDCG_Rate as shown in Fig.
= 13.2 2(a-b). The membership functions for the output variables
Hence, in similar way TDCG_rate can be calculated for all (expected diagnosis such as SI and OP) are shown in Fig
samples of Table 3. 2(c-d).
In Section "B", membership function of input and output
variables for fuzzy FLTD expert system are discussed.
Section "B" also presents the fuzzy logic based results.
Section III describes the simulation FL based results.
B. Input and Outputs of Fuzzy-Logic System:
Total Dissolved Combustible Gas (TDCG) in transformer
fault detection concept is useful in finding out the suitable
oil-sampling interval based on the health condition of the

ISSN: 2278-5299 26
 Malik et. al., International Journal of Latest Research in Science and Technology.

Rule 8 If (TDCG is COND.3) and (TDCG_Rate is 10-30) then

(SI is SIW)(OP is OPC)
Rule 9 If (TDCG is COND.3) and (TDCG_Rate is >30) then (SI
is SIW)(OP is OPC)
Rule If (TDCG is COND.4) and (TDCG_Rate is <10) then (SI
10 is SIW)(OP is OPC)
Rule If (TDCG is COND.4) and (TDCG_Rate is 10-30) then
11 (SI is SID)(OP is OPC)
Rule If (TDCG is COND.4) and (TDCG_Rate is >30) then (SI
12 is SID)(OP is OPD)
(a) However, this method can not specify the type of fault
that occurs in the transformer. This method is only able to
detect whether the transformer is in good or bad condition.

(b)

(c) Fig. 3: Test results of fuzzy transformer diagnosis with trapezoidal

membership function

(d)
Fig. 2: Trapezoidal membership functions: (a) input variable of TDCG;
(b) input variable of TDCG_Rate; (c) output variable of SI; (d) output
variable of OP.

Based on the experimental results, a set of fuzzy rules

relates the input linguistic variables to the output are
developed as shown in Table 2. The expected diagnosis for
transformer using input variables is shown in Fig. 3. 3D
Fig. 4: Surface viewer of OP w.r.t. TDCG and TDCG_Rate
surface viewer outputs for SI and OP with respect to (w.r.t.)
TDCG and TDCG_Rate are shown in Fig. 4 and Fig. 5
respectively.
Table 2
Fuzzy inference rules for Transformer Health
Rule 1 If (TDCG is COND.1) and (TDCG_Rate is <10) then (SI
is SIA)(OP is OPA)
Rule 2 If (TDCG is COND.1) and (TDCG_Rate is 10-30) then
(SI is SIQ)(OP is OPA)
Rule 3 If (TDCG is COND.1) and (TDCG_Rate is >30) then (SI
is SIM)(OP is OPB)
Rule 4 If (TDCG is COND.2) and (TDCG_Rate is <10) then (SI
is SIQ)(OP is OPB)
Rule 5 If (TDCG is COND.2) and (TDCG_Rate is 10-30) then
(SI is SIM)(OP is OPB)
Rule 6 If (TDCG is COND.2) and (TDCG_Rate is >30) then (SI Fig. 5: Surface viewer of SI w.r.t. TDCG and TDCG_Rate
is SIM)(OP is OPB)
Rule 7 If (TDCG is COND.3) and (TDCG_Rate is <10) then (SI
is SIM)(op is OPC)
ISSN: 2278-5299 27
 Malik et. al., International Journal of Latest Research in Science and Technology.

III. RESULTS AND DISCUSSION It can be seen from the Table no. 4 that the new method
Using fuzzy logic method, seventy numbers of 6.3-52 is generally in agreement with ANSI/IEEE method for
MVA, 50 Hz, having voltage ratio 220/11kv, 132/33kv, power transformers diagnosis. Compared with ANSI/IEEE
132/11kv, 66/11kv power transformers of Himachal C57.104 method, the fuzzy logic method also has some
Pradesh State Electricity Board (HPSEB), India were advantages. For example, due to no matching condition,
diagnosed and some typical results are given in Table no. 3. five transformers could not be diagnosed by the ANSI/IEEE
These results are taken by using Kalman Transport-X DGA method but are diagnosed by the fuzzy logic method. In
analyzer. For this method, 4 operating procedures (OP) and some cases, the deterioration of insulating paper may be
5 sampling intervals (SI) are determined by choosing the only at the early stage or intermittent which did not produce
highest degree of membership value obtained from the sufficient gases to give a stronger indication. However,
fuzzy inference rules. The operating procedure and such information obtained should be useful for future trend
sampling interval can be classified into the linguistic analysis.
variable based on the degree of membership function as
Table 5(a)
shown as Table 5(a) and 5(b). Transformer gas concentration for fault diagnosis
Table 3 Membership Degree (out of 10) of Membership Degree (out of
Transformer gas concentration for fault diagnosis Fuzzy output Result for SI 10) of Fuzzy output Result
S. H2 CO2 CO C2H4 C2H6 CH4 C2H2 TDCG SIA >8 for OP
N. SIQ 6.0-7.84 OPA >7.45
1 6 2552 295
27 137 79 <0.5 544 SIM 4.97-5.85 OPB 6.25
2 <5 2261 51
63 160 4 2 280 SIW 2.96-3.84 OPC 3.75
3 10 8185 358
21 84 18 16 506 SID <1.9 OPD <1.29
4 7 9320 726
327 106 171 <0.5 1337
5 16 5004 935
28 6 12 29 1026
Table 5(b)
6 1866 229 10
111 2 64 1265 3318 Transformer gas concentration for fault diagnosis
7 1879 198 6
36 5 29 521 2477
Degree of Membership Condition
8 505 2626 340
817 82 256 881 2881 (out of 1)
9 3619 2395 41
80 4 61 900 4705 1 Most Encourageable
10 3524 103 10
115 4 69 1163 4885 0.75 – 0.99 Encourageable
Table 4 0.5 – 0.74 Preferable
Action Based on Fuzzy-Logic and comparison with IEEE Std.
Action for Gas Action for Gas
Generation Rate as Generation Rate as Let us an example for Interpretation:
TRF N0.

TDCG

TDCG
_Rate

per IEEE Std. C57.104 per Fuzzy Logic Most encourageable operating procedure :-
Sampling Operating Sampling Operating Exercise caution. Analyse for individual gases. Plan outage
Interval Procedure Interval Procedure
(SI) (OP) (SI) (OP)
Preferable sampling interval: - Week.
1 544 <10 SIA OPA 8.98-8.99 8.75
10-30 SIQ 6-7.84 7.9-8.75 IV. CONCLUSION
>30 SIM OPB 4.98 6.25
2 280 <10 SIA OPA 8.98-8.99 8.75 This paper has presented a framework for performing
10-30 SIQ 6-7.84 7.49-8.75 diagnostics using fuzzy-logic expert system. FL techniques
>30 SIM OPB 4.97-4.98 6.25 has provided efficient solution to decide the sampling
3 506 <10 SIA OPA 8.98-8.99 8.75 interval (SI) of transformer oil and operating procedure (OP)
10-30 SIQ 6-7.84 7.79-8.75 for maintenance in transformer. AI has emerged over the
>30 SIM 4.97-4.98 6.25
4 1337 <10 SIQ 6.97-6.99 6.25
years and has made remarkable contribution to the various
10-30 SIM 4.97-5.85 6.25 fields of engineering. In fact, they have already been
>30 OPB 4.97-4.98 6.25 successfully applied in many industries. In case of Electrical
5 1026 <10 SIQ 6.97-6.99 6.25 systems the models presented by AI are tested and
10-30 SIM 4.97-5.85 6.25 practically used. They have solved many different categories
>30 4.97-4.98 6.25 of faults analysis. A lot a research is still required in order to
6 3318 <10 4.97-4.99 3.75 understand electrical system behaviour.
10-30 SIW 2.96-3.84 3.75
>30 2.96-2.98 3.75
7 2477 <10 SIM 4.97-4.99 3.75 ACKNOWLEDGMENT
10-30 SIW 2.96-3.84 3.75
OPC The authors would like to thank Himachal Pradesh
>30 2.96-2.98 3.75
8 2881 <10 SIM 4.97-4.99 3.75 Electricity Board (HPSEB), Shimla (India) for providing
10-30 SIW 2.96-3.84 3.75 the transformer's oil samples for experimental testing
>30 2.96-2.98 3.75 results in this study.
9 4705 <10 2.99 3.75 The authors are also thankful to Technology Information
10-30 SID 1.03-1.9 2.51-3.75 Forecasting and Assessment Council and Centers of
>30 OPD 1.03 1.26
Relevance & Excellence (TIFAC-CORE) on Power
10 4885 <10 SIW OPC 2.96-298 3.75
10-30 SID 1.01-1.9 2.51-3.75
Transformer Diagnostics at NIT Hamirpur for providing
>30 OPD 1.01-1.02 1.25

ISSN: 2278-5299 28
 Malik et. al., International Journal of Latest Research in Science and Technology.

necessary infrastructural facilities for carrying out the [10] O. Vanegas, Y. Mizuno, K. Naito , and T Kamiya, “ Diagnosis of Oil
research work. Insulated Power Apparatus using Neural Network Simulation,” IEEE
Transactions on Dielectrics and Electrical Insulations, Vol 4, No3,
June 1997, pp 290-299.
[11] Hong- Tzer Yang and Yann-Chang Huang, “Intelligent Decision
REFERENCES Support for Diagnosis of Incipient Transformer Faults Using Self-
[1] Hasmat Malik, Tarkeshwar and R.K. Jarial, "An Expert System for Organizing Polynomial Networks,” IEEE Transactions on Power
Incipient Fault Diagnosis and Condition Assessment in Transformers Systems Vol 13, No: 3, August 1998, pp 946-952.
Energy Management", IEEE Int. Conf. on Computational [12] J. L. Guardado, J. L. Naredo, P. Moreno and C. R. Fuerte, “A
Intelligence and Communication Systems, pp. 138- 142, Oct. 2011. comparative study of Neural Network Efficiency in Power
[2] Sin P. Lai, A.Abu-Siada, Syed M.Islam and Gary Lenco, Transformer Diagnosis Using Dissolved Gas Analysis,” IEEE
“Correlation between UV-Vis Spectral Response and furans Transactions on Power Delivery, Vol. 16, No. 4, October 2001, pp
measurement of transformer oil”, in Condition Monitoring and 643-647.
Diagnosis, 2008. CMD 2008. International Conference on, 2008, [13] A. S. Farag, M. Mohandes and Ali Al Shaikh, “Diagnosing Failed
pp.649-662 Distribution Transformers using Neural Networks,” IEEE
[3] A.Abu-Siada, “Remnant life estimation of power transformer using Transactions on Power Delivery, Vol. 16, No. 4, October 2001, pp
oil UV- VIS spectral response”, IEEE.2009. 631-636.
[4] Dornerburg, E., Strittmatter, W., ‘Monitoring oil cooling [14] K. F. Thang, R. K. Aggarwal, A. J. MacGrail and D. G. Esp,
transformers by gas analysis’, Brown Boveri Review, May “Application of Self-Organizing Map Algorithms for Analysis and
1974;61:238–47 Interpretation of Dissolved Gases in Power Transformers,” Power
[5] Kelly, J.J., ‘Transformer fault diagnosis by gas-gas analysis’, IEEE Engineering Society Summer Meeting, 2001. IEEE, Vancouver, BC,
Transactions on Industry Applications, December 1980;16(4):777–82 Canada,07 /15/2001 Vol. 3,pp 1881-1886.
[6] Rogers, R., ‘IEEE and IEC codes to interpret incipient faults in [15] M. H. Wang, “Extension Neural Network for Power Transformer
transformer, using gas in oil analysis’, IEEE Transactions on Incipient Fault Diagnosis,” IEE Proceeding Generation Transmission
Electrical Insulation, October 1978;13(5):349–54 Distribution, Vol 150, No. 6, Nov 2003, pp 679-656.
[7] Mineral Oil Impregnated Electrical Equipment in service. Guide to [16] D.V.S.S Siva Sarma and G.N.S Kalyani, “ANN approach for
the interpretation of Dissolved Free Gas Analysis”. CEI, IEC 60599, Condition Monitoring of Power Transformers Using DGA,” (0-1803-
second edition 1999-03. 8560/04©2004 IEEE) pp 444-447.
[8] N.A Muhamad, B. T Phung and T R Blackburn, “Comparative study [17] C.-H. Yann, "Evolving neural nets for fault diagnosis of power
and analysis of DGA methods for Mineral Oil Using Fuzzy Logic,” transformers", IEEE Trans. Power Deliv. 18 (3) (2003) 843–848.
Proceedings of 8th International Power Engineering Conference [18] R.E James and Q Su “Condition Assessment of High-Voltage
(IPEC2007), pp 1301-1306 Insulation in Power System Equipment”, IET Power and energy
[9] Diego Roberto Moaris and Jacqueline Gisete Rolim, “An artificial series 53, Published by The Institution of Engineering and
Neural Network Approach to Transformer Fault Diagnosis,”IEEE Technology, London, United Kingdom 2008.
Transactions on Power Delivery , Vol. 11, No. 4, April 1996, pp [19] IEEE Standard Guide C57.104 Guide for the Interpretation of Gases
1836- 1841. Generated in Oil-Immersed Transformer, 1992

ISSN: 2278-5299 29