Cognitive artificial-intelligence for doernenburg dissolved gas analysis interpretation

TELKOMNIKA, Vol.17, No.1, February 2019, pp.268~274
ISSN: 1693-6930, accredited First Grade by Kemenristekdikti, Decree No: 21/E/KPT/2018
DOI: 10.12928/TELKOMNIKA.v17i1.11612  268
Received July 28, 2018; Revised October 10, 2018; Accepted November 6, 2018
Cognitive artificial-intelligence for doernenburg
dissolved gas analysis interpretation
Karel Octavianus Bachri*
1
, Umar Khayam
2
, Bambang Anggoro Soedjarno
3
,
Arwin Datumaya Wahyudi Sumari
4
, Adang Suwandi Ahmad
5
1,2,3,4,5
Bandung Institute of Technology, Ganesha St. 10, Bandung, Indonesia, 40132, +62-22-2502260
*Corresponding author, e-mail: karel232@students.itb.ac.id
1
, umar@hv.ee.itb.ac.id
2
,
b.anggoro55@yahoo.com
3
, arwin.sumari@yahoo.com
4
, adangsahmad@yahoo.com
5
Abstract
This paper proposes Cognitive Artificial Intelligence (CAI) method for Dissolved Gas Analysis
(DGA) interpretation adopting Doernenburg Ratio method. CAI works based on Knowledge Growing
System (KGS) principle and is capable of growing its own knowledge. Data are collected from sensors, but
they are not the information itself, and thus, data needs to be processed to extract information. Multiple
information are then fused in order to obtain new information with Degree of Certainty (DoC). The new
information is used to identify faults occurred at a single observation. The proposed method is tested using
the previously published dataset and compared with Fuzzy Inference System (FIS) and Artificial Neural
Network (ANN). Experiment shows CAI implementation on Doernenburg Ratio performs 115 out of 117
accurate identification, followed by Fuzzy Inference System 94.02% and ANN 78.6%. CAI works well even
with small amount of data and does not require trainings.
Keywords: cognitive artificial-intelligence, DGA interpretaion, information fusion, knowledge growing
system
Copyright © 2019 Universitas Ahmad Dahlan. All rights reserved.
1. Introduction
Dissolved Gas Analysis (DGA) interpretation is the most reliable method in transformer
fault diagnosis, as it is capable of detecting incipient faults before they becomes
catastrophic [1–3]. Transformer oil degrades over time during operation [4–6]. At the presence
of stresses, radices will be released from the oil as shown in Table 1 [7]. The radices will then
form combustible gases as shown in Table 2 [5], [8].
Table 1. Released Radices [7]
Energy Level (KJ/mol) Fault
Radix
Released
< 338
Partial
Discharge
H*
, CH3
*
338 Low-Energy CH2
*
607 Arcing CH*
720 Thermal-Low C*
> 960 Thermal-High C*
Table 2. Gases Formed by Radices [5], [8]
Radices H*
CH3
*
CH2
*
CH*
H*
H2 CH4 CH4 CH4
CH3
*
CH4 C2H6 - -
CH2
*
CH4 - C2H4 -
CH*
CH4 - - C2H2
The more faults occurs, the more combustible gases are produced in the oil. These
gases accelerate the degradation process. There are two kinds of transformer paper
degradation processes, Hydrolysis and Pyrolysis [8], [9]. Pyrolysis is related to heat, while
Hydrolysis is related to water. Pyrolysis produces Oxygen, which will then oxidize oil and paper
insulator. Hydrolysis causes depolymerization of the oil insulator and later on produces Carbon
Monoxide and Carbon Dioxide, which are acidic oxides that accelerate Hydrolysis even more.
There are several conventional methods commonly used in DGA interpretation, one of
which is Doernenburg Ratio Method (DRM). Conventional methods have limitation, they have
low accuracy [10–12], moreover, they have high failure rate [1]. There have been several
researches working on DGA interpretation. They started from labeled dataset to test their
proposed methods [10]. All of them used Artificial Intelligence (AI) to imitate human experts,

TELKOMNIKA ISSN: 1693-6930 
Cognitive artificial-intelligence for doernenburg dissolved gas... (Karel Octavianus Bachri)
269
some used FIS [13], [14], some used ANN [1], [12], while some others used modification of
standard-AI method [11]. Ratio DGA interpreting methods have been developed to identify
faults [1], [10–14] as seen in Table 3.
Table 3. Related Works
Author, year Problem Solution Result
M Duval,
2001
conventional methods have Low
accuracy
Dataset creation In-depth description of the
five main type of fault [10].
F Zakaria,
2012
DGA is reliable, but conventional
methods have low accuracy
Muti-layer feed forward
back preopagation ANN
Increase accuracy, use three
layer as the optimal number
of layer [12]
S. A. Khan,
2014
conventional methods have Low
accuracy
ANFIS on IEC 599 Fault identification and
location [11].
M Beykverdi,
2014
IEC Ratio method has high failure
rate
Combination of ANN
Nero-ICA
Failure rate has been
successfully decreased[1]
F D Samirmi,
2015
Basic ontology model requires
improvement
Fuzzy Ontology
Reasoning
Multi-agent oriented System
improves accuracy of basic
ontology model[13]
Trianto,
2016
AI and conventional methods
have different characteristics
Combining AI and
conventional method
Accuracy 93.7% [14]
Normal AI methods have some disadvantages. ANN requires a large number of
samples in training [15], while FIS requires complex computing resources and the lacks of
design techniques [16]. Adaptive Neural Fuzzy Inferece System (ANFIS) improves the
accuracy, but increase complexity of the system.
The most recent development of AI is Cognitive Artificial Intelligence (CAI) [17]. CAI
uses a method knowns as Knowledge Growing System (KGS) that is able to solve multi-input
and multi-output problems occured in the environment. This method is developed from the
combination of Bayesian Inference Method (BIM), Maximum a Posteriori (MAP) theorem, and
Linear Opinion Pool (LOP) to obtain decision options [17–19]. In the area of DGA interpretation,
DGA problem requires multi-input processing in order to make decision in a multi-output
situation, and hence, this method is suitable for this situation.
2. Proposed Method
CAI method solves problems based on Knowledge Growing System (KGS) principle
and can be described using Figure 1.
Knowledge
Base
Information
Multi-Source
Information fusion
and Inferencing
Information
with DoC
DoC
satisfied?
New
Knowledge
N
Stored to knowledge base
New
Knowledge
satisfied?
Ultimate
Knowledge
Y
Stored to knowledge base
KNOWLEDGE
INFORMATION
N
Knowledge
fusion
and
Inferencing
Y
Figure 1. Diagram of KGS [17]
The diagram consists of two parts, the Information Part and the Knowledge Part. In the
Information Part, information from multi sources are fused to extract new information with
Degree of Certainty (DoC). If the DoC satisfies certain value, the information will then be sent to
the knowledge part as the Current Knowledge for further processes.

 ISSN: 1693-6930
TELKOMNIKA Vol. 17, No. 1, February 2019: 268-274
270
In the Knowledge Part, the Current Knowledge will be fused with the Existing
Knowledge in order to obtain the New Knowledge with DoC. If the DoC satisfies certain value,
the knowledge will then become the Ultimate Knowledge.
The main component of CAI is Information Fusion [17]. Information Fusion processes
information by imitating the way human brain processes information. Multi-source information
were processed in order to obtain new information.
ASSA2010 is the new name given by the method’s inventors, to the original method’s
name namely A3S (Arwin-Adang-Aciek-Sembiring) [17]. It is the algorithm used to fuse
information that that is derived from BIM, MAP, and LOP (BIM + MAP + LOP) [17–19] as shown
in (1). The probability of Hypothesis B given A is the probability of Indication A given B times the
probability of B per all the possible events.
𝑃(𝐵𝑗|𝐴𝑖) = ∑ (
𝑃(𝐴 𝑖|𝐵 𝑗)𝑃(𝐵 𝑗)
∑ 𝑃(𝐴 𝑖|𝐵 𝑘)𝑃(𝐵 𝑘)𝑚
𝑘=1
)𝑛
𝑖=1 (1)
where P(Bj|Ai) is the probability of Hypothesis Bj given Ai, P(Ai|Bj) is the probability of Indication
Ai given Bj. P(Bj) is the probability of Hypothesis Bj itself, and Σ(P(Ai|Bk) P(Bk) is the combination
of all possible events. They are put into arrangement in the Indication-Hypothesis matrix as
fused information (P(υi
j
)). BIM + MAP is unable to do decision making in multi-input multi-ouput
situation as faced in DGA situation. ASSA2010 improves BIM + MAP by combining the concept
of LOP so it has capability in making decision in multi-input multi-ouput situation based on the
New Knowledge Probability Distribution (NKPD) it produces after performing computation on
Indication-Hypothesis data.
NKPD comprises of DoC for each hypothesis, which indicates how a Hypothesis can be
believed regarding to the presence of the combination of Indications. NKPD can be calculated
using (2):
𝑃(𝜓𝑖
𝑗
) =
∑ 𝑃(𝑣𝑖
𝑗
)𝛿
𝑖=1
𝛿
(2)
where δ is the number of sensors and P(ψi
j
) is the inferenced fused-information. The system
keeps receiving information from sensors in the form of P(ψi
j
) according to (3).
𝑃(𝜙 𝛾
𝑗
) = {
1, 𝑃(𝜓 𝛾
𝑗
) >
𝑃(𝜓 𝛾
𝑗
)
𝜆
0, 𝑃(𝜓 𝛾
𝑗
) ≤
𝑃(𝜓 𝛾
𝑗
)
𝜆
(3)
where λ is the number of fused information. The New Information is fused with the previous
information to produce NKPD over Time (NKPDT), which is as shown in (4).
𝑃(𝜃𝑗) =
∑ 𝑃(𝜙 𝛾
𝑗
)Γ
𝛾=1
Γ
(4)
where Γ is the number of observation and P(ϕγ
j
) is the NKPD of each observation. Decision will
be made based on the highest DoC in the NKPDT matrix as shown in (5):
𝐷𝑜𝐶 = {
𝑚𝑎𝑥[𝑃(𝜓 𝛾
𝑗
)], 𝑠𝑖𝑛𝑔𝑙𝑒 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛
𝑚𝑎𝑥[𝑃(𝜃𝑗)], 𝑡𝑖𝑚𝑒 𝑠𝑒𝑟𝑖𝑒𝑠 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛
(5)
where j = 1, 2, 3, …, λ. The proposed method can be described in pseudocode as follows:
1. Read data [gas].
2. Calculate Ratio using (6) to (9).
3. Localize Ratio using Table 4.
4. Create observation matrix as Table 5.
5. Obtain NKPD using (2).
6. determine fault using (5)

271
7. if more data available then goto step 1
The Doernenberg Ratio Method uses gas concentration Ratios that is calculated
using (6)-(9) and works based on Figure 2.
𝑅1 = [𝐶𝐻4]/[𝐻2] (6)
𝑅2 = [𝐶2 𝐻2]/[𝐶2 𝐻4] (7)
𝑅3 = [𝐶2 𝐻2]/[𝐶𝐻4] (8)
𝑅4 = [𝐶2 𝐻6]/[𝐶2 𝐻2] (9)
Figure 2. Doernenberg decision system flowchart [7]
Localization of Ratios is shown in Table 4. The input of Table 4 is ratio Ri, and the
output is score Ai, if R1 exceeds 1, A1 will be given score ‘2’, if R1 is between 0.1 and 1, A1 will
be given score ‘1’, and if R1 is below 0.1, A1 will be given score ‘0’. If R2 is below 0.75, A2 will be
given score ‘0’, and if R2 exceeds 0.75, R2 will be given score ‘1’. If R3 exceeds o.3, A3 will be
given ‘1’, otherwise A3 will be given ‘0’. If R4 exceeds 0.4, A3 will be given ‘1’, otherwise A4 will
be given ‘0’.Observation matrix is shown in Table 5, while NKPD is calculated using (2) and is
shown in Table 6. Since the test data is not time-series, NKPDT does not need to be calculated,
but for time-series data, NKPDT can be calculated using (4) to show trends in a certain time
interval. Fault is identified using (5) for single observation. It is the max of  j
iP for each
sample.
Table 4. Ratios Localization [7]
TD PD A
R1 > 1 <0.1 0.1 < <1
R2 <0.75 ns >0.75
R3 <0.3 <0.3 >0.3
R4 >0.4 >0.4 <0.4
Table 5. Observation Matrix [17]
R Score
Hypotheses
B1 B2 B3
R1 A1  11 ABP  12 ABP  13 ABP
R2 A2  21 ABP  22 ABP  23 ABP
R3 A3  31 ABP  32 ABP  33 ABP
R4 A4  41 ABP  42 ABP  43 ABP

 ISSN: 1693-6930
272
Table 6. NKPD Matrix [17]
Sample
Hypotheses
H_TD H_PD H_A
1  1
1P  1
2P  1
3P
2  2
1P  2
2P  2
3P
… … … …
j  j
P 1  j
P 2  j
P 3
… … … …
n  1
nP  1
nP  1
nP
3. Results and Analysis
The proposed method is verified using previously-published data consisting
117 samples with 9 Partial Discharge (PD), 26 Low-Energy Discharge (LE), 48 High-Energy
Discharge (A), 16 Thermal-Low (TL), and 18 Thermal-High (TH) that is put into groups based on
the fault types. For example, data #1 of PD, R1=0.07. R2=0.00, R3=0, and R4=inf. Using Table 3,
Ratios are then localized and given score, A1=0, A2=0, A3=0, A4=1. These scores are arranged
in observation matrix in Table 5. Table 3 is also used to give scores to each Hypothesis, for
example R1=‘0’, column H_TD, H_PD, and H_A for row R1 is given ‘0’, ‘1’, and ‘0’ respectively
to be arranged in Table 7. Other types of faults can be analyzed using the same method. NKPD
is calculated using (2) and is shown in Table 8.
Table 7. Observation Matrix PD
Data A score
Hypotheses
H_TD H_PD H_A
1
A1 0 0 1 0
A2 0 1 1 0
A3 0 1 1 0
R4 1 1 1 0
… … … … … …
9
A1 1 0 0 1
A2 0 1 1 0
A3 0 1 1 0
A4 1 1 1 0
Table 8. NKPD for PD.
No. H_TD H_PD H_A
1 0.3750 0.6250 0.0000
2 0.2500 0.6250 0.1250
3 0.3750 0.3750 0.2500
4 0.3750 0.3750 0.2500
5 0.2500 0.3750 0.3750
6 0.2500 0.3750 0.3750
7 0.3750 0.6250 0.0000
8 0.2500 0.6250 0.1250
9 0.3750 0.3750 0.2500
Sample number 1, 2, 7, and 8 in dataset PD indicates the major faults were PD, some
with the minor fault TD, while others with A. meanwhile, in sample number 3, 4, 5, 6, and 9, no
major faults occurred, however PD are still identified as the DoCs are among the highest. All
samples in the dataset LE indicate the major faults are Arcing, while minor faults are mostly PD,
this is due to the limitation in Doernenburg method, which identifies only verylow Energy
Discharge (PD) and High-Energy Discharge (A). All samples in the dataset HE indicate the
major faults are Arcing, while some others are identified as TD and also some others as PD.
Dataset TL and TH give TD, with two samples in the dataset TL are identified as PD.
The dataset are also tested using Fuzzy Inference System (FIS) and Artificial Neural
Network (ANN). FIS and ANN are designed to be implemented on the same conventional DRM
using the same dataset. FIS has four input membership functions (R1, R2, R3, and R4) and three
output membership functions (TD, PD, and A). The result is shown in Table 9. For dataset PD,
FIS has 88.9% accuracy, while ANN has 55.6%. For dataset LE, FIS has 92.3% accuracy, while
ANN has 53.9%. For dataset HE, FIS has 100% accuracy, while ANN has 95.8%. For dataset
TL, FIS has 75% accuracy, while ANN has 81.3%. For dataset TH, FIS has 100% accuracy,
while ANN has 77.8%. The accuracy of FIS in the testing is confirmed by various authors, which
lies between 93% to 96% with various datasets and conditions [13], [14], while the accuracy of
ANN varies due to the small numbers of samples [1], [11, 12].
The proposed method performs 87.5% accuracy for dataset TL and 100% accuracy for
other dataset. The overall performance of CAI is 98.3% accuracy, with 115 correct
identifications out of 117 samples. The accuracy of ANN is proportional to the number of
samples.

273
Table 9. Comparison between CAI and other Methods
Fault Type #Data
#Accurate %Accuracy
FIS ANN CAI FIS ANN CAI
PD 9 8 5 9 88.9 55.6 100
LE 26 24 14 26 92.3 53.9 100
HE 48 48 46 48 100 95.8 100
TL 16 12 13 14 75 81.3 87.5
TH 18 18 14 18 100 77.8 100
Overall 117 110 92 115 94.02 78.6 98.3
4. Conclusion and Future Works
In this research, a novel method of DGA interpretation is proposed. The proposed
method is based on CAI with KGS as the core. In this paper CAI adopts Doernenburg Ratio
method. There are some important things that can be noted from experiments.
4.1. Conclusion
The proposed method has successfully identify fault based on DGA with the overall
accuracy of 98.3% from five types of fault with 87.5% accuracy on Thermal-Low, and 100%
accuracy on Partial Discharge, Low-Energy Discharge, High-Energy Discharge, and Thermal
High, which results in higher accuracy than FIS (94.02%) and ANN (78.6%). This CAI-based
proposed method does not require a large amount of data and training as ANN does and does
not require a complex computation as FIS does.
The more the sample number, ANN tends to have more accuracy. The accuracy of FIS
is the proportional to the membership functions and rule base. The more detail the membership
functions, the more accuracy FIS will have, resulting more complexity.
The accuracy of the proposed method is independent to the number of samples. It is
shown in Table 8, where various number of samples resulting 100% accuracy. The complexity
of the proposed method is only proportional on the number of hypotheses and indications.
Another advantage of the proposed method is CAI works well even with small amount of data
and does not require trainings. Meaning that the proposed method is able to capture the ability
of the human brain in performing fast learning to obtain decision options for multi-input multi-
output problems.
4.2. Future Works
The proposed method is a novel method in computing, and AI area that is developed to
solve multi-input and multi-output problems, such as in Biomedic [20], DPA
countermeasure [21], Intrusion Detection System [22], emotion modeling [23], and many other
area requiring a Multi-input-Multi-Output processing such as in big data [24]. The proposed
method can be implemented in software [22] and hardware [21], [25]. For the time being, a
cognitive processor is being developed [26] and is going to be used in an embedded system to
perform such tasks with more power efficiency and higher speed.
Acknowledgment
The first author would like to show his gratitude to Mr. Harry Gumilang for his support
through discussions.
References
[1] M. Beykverdi, F. Faghihi, and others. A new approach for transformer incipient fault diagnosis based
on dissolved gas analysis (DGA). Nova Journal of Engineering and Applied Sciences. 3(2). 2016.
[2] D. Bhalla, R. K. Bansal, and H. O. Gupta. Transformer incipient fault diagnosis based on DGA using
fuzzy logic. in 2010 India International Conference on Power Electronics (IICPE). 2011: 1–5.
[3] H. Ma, Z. Li, P. Ju, J. Han, and L. Zhang. Diagnosis of power transformer faults on fuzzy three-ratio
method. in The 7th International Power Engineering Conference IPEC 2005. 2005: 1–456.
[4] V. Arakellian and I. Fofana. Physicochemical aspects of gassing of insulating liquids under electrical
stress. IEEE Electrical Insulation Magazine. 2009; 25(3): 43–51.
[5] R. Sanghi. Chemistry behind the life of a transformer. Resonance. 2003; 8(6): 17–23.

 ISSN: 1693-6930
274
[6] J. B. Taylor. Discussion on ‘Corona Losses between Wires at High Voltages’ (Harding), ‘The Law of
Corona and the Dielectric Strength of Air-II’ (Peek), and ‘The Electric Strength of Air-III’ (Whitehead).
Boston, Mass., June 25, 1912. Transactions of the American Institute of Electrical Engineers. 1912;
XXXI(1): 1119–1130.
[7] “IEEE Guide for the Interpretation of Gases Generated in Oil-Immersed Transformers - Redline. IEEE
Std C57.104-2008 (Revision of IEEE Std C57.104-1991) - Redline. 2009; Feb: 1–45.
[8] H. Gumilang. Hydrolysis process in PLN P3BJB transformers as an effect of oil insulation oxidation. in
2012 International Conference on Condition Monitoring and Diagnosis (CMD). 2012: 1147–1150.
[9] B. CIGRE Working Group 12.18. 227 guidelines for life management techniques for power
transformers. CIGRE WG. 2003: 12.
[10] M. Duval and others. Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC
10 databases. IEEE Electrical Insulation Magazine. 2001. 17(2): 31–41.
[11] S. A. Khan, M. D. Equbal, and T. Islam. ANFIS based identification and location of paper insulation
faults of an oil immersed transformer. in 2014 6
th
IEEE Power India International Conference
(PIICON). 2014: 1–6.
[12] F. Zakaria, D. Johari, and I. Musirin. Artificial neural network (ANN) application in dissolved gas
analysis (DGA) methods for the detection of incipient faults in oil-filled power transformer. in 2012
IEEE international conference on Control system, computing and engineering (ICCSCE). 2012:
328–332.
[13] F. D. Samirmi, W. Tang, and Q. Wu. Fuzzy Ontology Reasoning for Power Transformer Fault
Diagnosis. Adv. Electr. Comput. Eng. 2015; 15: 107–114.
[14] Trianto, Suwarno, Y. Li, and G.-J. Zhang. Combining conventional and artificial intelligence DGA
interpretation methods using optimized weighting factor. in 2016 International Seminar on Intelligent
Technology and Its Applications (ISITIA). 2016: 37–42.
[15] M.-Y. Chow and P. Goode. The advantages and challenges of machine fault detection using artificial
neural network and fuzzy logic technologies. in Proceedings of the 36
th
Midwest Symposium on
Circuits and Systems, 1993. 1993: 708–711.
[16] I. Bousserhane, A. Hazzab, M. Rahli, M. Kamli, and B. Mazari. Adaptive PI controller using fuzzy
system optimized by genetic algorithm for induction motor control. in International Power Electronics
Congress, 10
th
IEEE. 2006: 1–8.
[17] A. D. W. Sumari, A. S. Ahmad, A. I. Wuryandari, and J. Sembiring. Constructing brain-inspired
knowledge-growing system: A review and a design concept. in, 2010 International Conference on
Distributed Framework and Applications (DFmA). 2010: 1–7.
[18] K. O. Bachri, B. Anggoro, and A. S. Ahmad. Transformer performance calculation using information
fusion based on DGA interpretation. in International Symposium on Electronics and Smart Devices
(ISESD). 2016: 11–15.
[19] A. D. W. Sumari and A. S. Ahmad. Knowledge-growing system: The origin of the cognitive artificial
intelligence. in Electrical Engineering and Informatics (ICEEI), 2017 6
th
International Conference on
2017: 1–7.
[20] L. Goeirmanto, R. Mengko, and T. L. Rajab. Direction of ventricle contraction based on precordial lead
ECG signal. in Cyber and IT Service Management, International Conference on Electronics and Smart
Devices (ISESD). 2016: 1–3.
[21] S. D. Putra, A. S. Ahmad, and S. Sutikno. DPA-countermeasure with knowledge growing system. in
International Symposium on Electronics and Smart Devices (ISESD). 2016: 16–20.
[22] H. R. Talompo, A. S. Ahmad, Y. S. Gondokaryono, and S. Sutikno. NAIDS design using
ChiMIC-KGS. in 2017 International Symposium on Electronics and Smart Devices (ISESD). 2017:
346–351.
[23] T. Sutikno, M. Facta, and G. A. Markadeh. Progress in Artificial Intelligence Techniques: from Brain to
Emotion. TELKOMNIKA Telecommunication Computing Electronics and Control. 2013; 9(2):
201–202.
[24] T. Sutikno, D. Stiawan, and I. M. I. Subroto. Fortifying big data infrastructures to face security and
privacy issues. TELKOMNIKA Telecommunication Computing Electronics and Control. 2014; 12(4):
751–752.
[25] T. Sutikno, N. R. N. Idris, and A. Z. Jidin. Overview on Strategies and Approaches for FPGA
Programming. TELKOMNIKA Telecommunication Computing Electronics and Control. 2014; 12(2):
273–282.
[26] C. O. Sereati, A. D. W. Sumari, T. Adiono, and A. S. Ahmad. Implementation Knowledge Growing
System Algorithm using VHDL. in International Symposium on Electronics and Smart Devices
(ISESD). 2016: 7–10.

Cognitive artificial-intelligence for doernenburg dissolved gas analysis interpretation

More Related Content

What's hot (18)

Similar to Cognitive artificial-intelligence for doernenburg dissolved gas analysis interpretation (20)

More from TELKOMNIKA JOURNAL (20)

Recently uploaded (20)

Cognitive artificial-intelligence for doernenburg dissolved gas analysis interpretation