MEV Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 Journal of Mechatronics, Electrical Power, and Vehicular Technology e-ISSN: 2088-6985 p-ISSN: 2087-3379 www.mevjournal.com https://dx.doi.org/10.14203/j.mev.2018.v9.17-24 2088-6985 / 2087-3379 ©2018 Research Centre for Electrical Power and Mechatronics - Indonesian Institute of Sciences (RCEPM LIPI). This is an open access article under the CC BY-NC-SA license (https://creativecommons.org/licenses/by-nc-sa/4.0/). Accreditation Number: (LIPI) 633/AU/P2MI-LIPI/03/2015 and (RISTEKDIKTI) 1/E/KPT/2015. Condition assessment of power transformers status based on moisture level using fuzzy logic techniques Vezir Rexhepi *, Petar Nakov Technical University of Sofia, Faculty of Electrical Engineering, Boulevard “Sveti Kliment Ohridski”, 8, 1000, Sofia, Bulgaria Received 19 February 2018; received in revised form 27 April 2018; accepted 2 May 2018 Published online 31 July 2018 Abstract Power transformers are one of the most expensive components; therefore the focus on their status and its continuous operation is the primary task. In the power systems, condition assessment of performance and reliability is based on the state of components, measurements, testing, and maintenance as well as their diagnosis. Hence, condition assessment of power transformer parameters is important regarding their status and finding incipient failures. Among many factors, the most factors that affect the safe operation and life expectancy of the transformer is the moisture in oil. It is known that the low moisture oil in power transformers causes many problems including electrical breakdown, increase the amount of partial discharge, decreases the dielectric withstand strength and other phenomena. Thus, knowledge about the moisture concentration in a power transformer is significantly important for safe operation and lifespan. In this study, moisture level in oil is estimated, and its status classification is proposed by using fuzzy logic techniques for the power transformer monitoring and condition assessment. Moreover, the goal of the study is to find methods and techniques for the condition assessment of power transformers status based on the state of moisture in oil using the fuzzy logic technique. These applied techniques increase the power system reliability, help to reduce incipient failures and give the better maintenance plan using an algorithm based on logic rules. Also, by using the fuzzy logic techniques, it is easier to prevent failures which may have consequences not only for transformers but also for the power system as a whole. ©2018 Research Centre for Electrical Power and Mechatronics - Indonesian Institute of Sciences. This is an open access article under the CC BY-NC-SA license (https://creativecommons.org/licenses/by-nc-sa/4.0/). Keywords: Moisture in oil; power transformers; fuzzy logic techniques; failures; condition assessment; measurements. I. Introduction Power transformers are key components for transmission and distribution infrastructure that transfer electricity from the substations to consumers. Forced component outages caused by failures may collapse the electrical power systems, interrupt the electricity consumption and give an environmental impact. The transformer could fail due to electrical failures, mechanical or thermal stresses. Such defects are sometimes catastrophic and can be lead to an unscheduled outage of the component, therefore causing the need of repairment and replacement. Assessment of power transformer condition is the basis for reliable operation and optimal repairs schedule. Generally, there are four main aspects of transformer condition monitoring and assessment, includes thermal dynamics, dissolved gas, partial discharge, and winding deformation [1]. Continuous monitoring of oil insulation characteristics has become an important task to avoid deterioration of its characteristics under working conditions. Several efforts have been made over the past years to study the electrical, physical, and chemical of insulating oils [2]. Methods for evaluating the status and monitoring of transformer parameters create a possibility of diagnosis and analysis to define accurate measures for necessary steps that should be taken into consideration. Various methods have been explained [3], such as using Fuzzy Logic techniques or the Neural Network methods. The combination of these techniques will improve diagnostic decision-making under the uncertainty inherent in diagnostic information, and it can also capture gradual system degradation [4]. The techniques for condition assessment, measurement, * Corresponding Author, Tel: +383 44 558 305 E-mail: vezir.rexhepi@gmail.com https://dx.doi.org/10.14203/j.mev.2018.v9.17-24 http://u.lipi.go.id/1436264155 http://u.lipi.go.id/1434164106 http://mevjournal.com/index.php/mev/index https://dx.doi.org/10.14203/j.mev.2018.v9.17-24 https://creativecommons.org/licenses/by-nc-sa/4.0/ https://crossmark.crossref.org/dialog/?doi=10.14203/j.mev.2018.v9.17-24&domain=pdf https://creativecommons.org/licenses/by-nc-sa/4.0/ V. Rexhepi and P. Nakov. / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 18 and interpretation of gases have been developed [5]. However the assessment technique should be extended to other parameters such as moisture concentration in oil. Furthermore, through such techniques, a more detailed look at the predictions of failures based on the statistics for transformers is not needed. Faults in different parts of transformer winding or other parts during impulse tests generate distinct signatures in the time-frequency spectrum. Thus, such signs are difficult and impossible for the human eyes to identify; this complex task can be successfully handled and developed by modern classification algorithm such as fuzzy logic [6]. The main objective of this paper is to design a model or technique for condition assessment of the moisture in oil of the power transformers. The assessment of parameter conditions through fuzzy logic techniques also makes a significant contribution to the recognition of transformer status as well as its ranking according to the conditions and quality of parameters in the case of moisture in oil. The model includes the design of the algorithm, which the given inputs are the conditions of moisture in oil (five conditions), and oil level (defining oil by standards). Those inputs are processed by logic rules IF, AND, THEN, to finally results in the condition assessment of the moisture in oil, monitoring measures, diagnosis, and preventive measures. The condition assessment through this logic is also compared with the statistical database. II. Failure analysis in a power system and assessment methods In a power system, unplanned outages, faults, and hidden failures are difficult to track down and to model because those are obviously hidden [7]. The reliability of the power system is classified into adequacy and security. The adequacy is related to the existence of sufficient generation of the electric power system to comply the consumer demand. Meanwhile, the security is related to the ability of power system to respond to the transients and disturbances that occur in the system [8]. Failure is an inability of a part or equipment to carry out its specified function. Meanwhile, a fault has broader meaning than a failure in specific equipment and its related features that needed for proper operation of the equipment [9]. A power transformer condition monitoring focuses mainly on the detection of incipient faults inside the transformer that are generated from the gradual deterioration. Some of these incipient faults may be detected during routine maintenance; however other faults may cause numerous problems before the regular maintenance cycle. Markov method is a stochastic process in which future states are conditional only on the present state and independent of previous states [10]. The Markov process in which there is the number of finite states 𝑆1, 𝑆2, 𝑆3, . . . , 𝑆𝑛 that may exist at any given time. The probability of the process moving from 𝑆𝑖 to 𝑆𝑗 is denoted by the transition probability - 𝑃𝑖𝑗 and the probability of the process remaining in the same state is denoted by the probability -𝑃𝑖𝑖 [11]. Markov method is widely used for reliability and simulation analysis [12]. A system is made of a number of components 𝑛, which at any given time may be operating successfully or not. The successful operation of the entire system depends on the operation or failure of its components. Therefore, the system may exist in one of two states as follow: 1. An operating state, where the system is operating even if some of its components have failed. A fully operational system is one in which no components have failed. 2. A failed state, where the system is not operating because of the failure of one or more of its components [12][13]. Previous researches have discussed the topic of condition assessment of transformers in several contexts. One method is fault identification at the defective condition, commissioning test, and trend analysis [14]. The power transformer monitoring process evaluation starts from the beginnings of incipient failures, their diagnosis, and the final actions of the measures to prevent the defects or failures are presented in Figure 1. Another method is a linear regression that consists of finding a linear function bxwxf  )()(  that gives the best interpolation of a set of training points labeled from Y R . Geometrically this corresponds to a hyperplane fitting of the given points. This technique is known as a method of least squares [15]. Therefore, the training process of a neural network involves tune of the value of the weights and the biases of the networks to optimize network performance, as defined by the network performance function [16].      N i N i iit N ie N mseF 1 1 2 )( 12 )( 1  (1) In the Equation (1),  is network outputs, t target, and mse is the mean square error. By comparing the data in their input, and processing by combining the data, it is possible to obtain results that enable the recognition of the conditions of the transformer operation according to the parameters that had been taken into account [17]. In some condition, assessment methods can be confirmed or automated with the use of electronic performance support systems such as statistics based systems, expert systems, and algorithms [18]. III. Problem formulation using fuzzy logic Probability and fuzziness are related to different concepts. Fuzziness is a type of deterministic Figure 1. Process evaluation of power transformer monitoring Symptoms Diagnosis Cure V. Rexhepi and P. Nakov / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 19 uncertainty. It describes the event class ambiguity, so fuzziness measures the degree in which an event occurs. A linguistic variable can be regarded as a variable of value, either a fuzzy number, or a variable. Fuzzy logic of fuzzy set theory provides a basis for mathematical modeling and language to express quite sophisticated algorithms in a precise manner [19]. The Fuzzy techniques consist of several phases and processes, which are analyzed according to the defined conditions in the assessment range. Through the building concept of the model, inputs, and outputs; the database is processed according to the numerical values of the membership that are allocated to the respective sets. The output is obtained through the set of inputs rules. The implementation of the fuzzy logic technique to a real application requires the following three steps: a) Fuzzification – convert classical data into fuzzy data or membership functions, b) Fuzzy Interference Process – combine membership functions with the control rules to derive the fuzzy output, and c) Defuzzification – use different methods to calculate each associated output to put them into a table or graphical structure [20]. The most popular fuzzy logic systems are Mamdani and Sugeno which use crisp data as inputs. As an extension of the Mamdani model in order to work with interval inputs, the fuzzy sets are represented by triangular fuzzy numbers [21]. The processes and steps of operating conditions assessment and their status according to fuzzy logic are also included. Determining the membership functions of inputs means that the data populations as well as the assessment of transformers parameters are inputted at certain time periods or on a continuous basis. The membership functions of the outputs are determined by members and their impact on the status of power transformers. For different cases involving aspects of issues evaluation that have an impact on specific parameters, such as temperature, state of gases, moisture in oil, paper, tap changer, hot spot partial discharge, and load factor. Therefore, this structure of fuzzy logic operation is built by populating the data at certain time periods in the online or offline mode.                                       hx gxf fxe ex hgfejand xd dxc bxa ax dcbai 0 ),,,,0(, 0 ),,,,0( (2) In Equation (2), the logic technique for defining the inputs and outputs parameters is depicted. This includes the membership determination by the levels and numbers belonging to the sets, which based on the respective measurements, and their comparison is made to reach the most accurate conclusions. Structuring the rules of fuzzy logic includes a set of rules (range of data) that are created according to the classification of transformer occurrences. The entirety of the failures causes is processed according to the logic operators, IF, AND, THEN. Through the processing of the causes, their ranking according to the assessment results is derived from the outputs. It constitutes a logical method of data processing and obtaining results that can be helpful in monitoring and diagnosis of transformers during continuous operation. In this study, a problem of condition estimation of the power transformers that are installed in the transmission system of Kosovo is addressed. The transformers conditions are estimated based on moisture content in the oil using fuzzy logic. Simulation results are compared with real measurements results, and a ranking of the power transformers conditions is presented. IV. Condition assessment model of power transformer parameters Condition assessment of electrical, thermal, and mechanical parameters is one of the fundamental elements for determining the operational continuity of transformers. Fuzzy logic may be applied for condition assessment and status classification of the transformers which emphasize the main parameters such as oil temperature, winding temperature, dissolved gases, moisture in oil, and partial discharge. Oil insulation is an important form of the power transformer insulation system. The remaining life of the transformer largely depends on the oil insulation status. Aging degree and moisture content are two important factors of insulation state assessment. Thus, accurate assessment technique to evaluate the condition of oil in the power transformers insulation has become a popular research topic. A lot of valuable research about oil insulation condition assessment and the influence of moisture have been done, such as; Spectroscopy, Polarization and depolarization Current, and Recovery Voltage Measurement [22]. This paper proposes a fuzzy model which involves two fuzzy input variables, i.e., moisture in oil and moisture level, and transformer condition status as the fuzzy output. The moisture in oil is divided into five functions: Condition 1 (Con1), Condition 2 (Con2), Condition 3 (Con3), Condition 4 (Con4), and Condition 5 (Con5). The moisture level is also divided into five functions: Very Low (VL), Low (L), Medium (M), High (H), and Very High (VH). The transformer condition status is divided into six functions: Very Good (VG), Good (G), Medium (M), Alarm 1 (A1), Alarm 2 (A2), and Alarm 3 (A3). Figure 2 illustrates the proposed fuzzy model. The model includes an algorithm where the two fuzzy inputs are given and processed by logic rules (IF, AND, THEN) to produce a condition assessment. This fuzzy model links the two fuzzy input variables with a transformer condition status. This model provides a clear view of a transformer conditions status and necessary measurement according to the fuzzy output. Figure 3 and Figure 4 show the functions of the fuzzy input and output variables, respectively. V. Rexhepi and P. Nakov. / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 20 Moisture in oil Moisture level Rule 1 Rule 2 Rule 3 Rule 4 Rule 5 Rule 11 Condition assessment Condition 1 Condition 2 Condition 3 Condition 4 Condition 5 Very low Low Medium High Very high Very good Good Medium Alarm 1 Alarm 2 Alarm 3 If moisture in oil is condition 1 and oil level is very low operation then condition assessment is very good Rule 6 If moisture in oil is condition 1 and oil level is low operation then condition assessment is very good If moisture in oil is condition 2 and oil level is low operation then condition assessment is good If moisture in oil is condition 2 and oil level is medium operation then condition assessment is good If moisture in oil is condition 3 and oil level is medium operation then condition assessment is good If moisture in oil is condition 3 and oil level is medium operation then condition assessment is medium If moisture in oil is condition 5 and oil level is very high operation then condition assessment is alarm 3 Figure 2. Fuzzy logic model for condition assessment (a) (b) Figure 3. The functions of fuzzy input variables; (a) moisture in oil; (b) moisture level V. Rexhepi and P. Nakov / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 21 According to oil moisture ranges in the standard of power transformer operation condition, the universe of discourse (UOD) of fuzzy input variables is expressed in Equations (3) and (4), whereas UOD of fuzzy output variable is expressed in Equation (5) [23].                         305 29204 20153 1562 601 econdtion dcondition ccondition bcondition acondition oilinmoisture (3)                         30 2920 2015 158 80 ehighvery dhigh cmedium blow alowvery levelmoisture (4)                          303 29252 25201 2015 158 80 ealarm ealarm dalarm cmedium bgood agoodvery assessmentcondition (5) This fuzzy rule refers to the standard regarding power transformer condition status [24]. Eleven fuzzy rules are proposed in this study and listed in Table 1. Figure 5 and Figure 6 illustrate 11 fuzzy rules that link each relevant fuzzy input variable with the corresponding fuzzy output variable using logic operators (IF, AND, THEN) in Matlab®/Simulink® environment. Figure 7 shows three dimensional form of the parameter levels extension according to the level of moisture in oil, output to the Z axis through measured samples, and data processing based on the rules of fuzzy logic. The standards or codes recommend the counter measures or actions to be done for each power transformer conditions status [25]. Table 2 summarizes the appropriate measure according to the standards/codes. V. Results and discussion Monitoring and diagnosis of the status of transformers in the operator transmission system of Kosovo were conducted from July to August 2017. The fuzzy logic model of transformer condition assessment includes 65 transformers from 20 up to 400 MVA power capacity. The Kelman TRANSPORT X – mobile labs was used to measure moisture in oil and its level. Measurement of moisture in oil and moisture level was conducted manually. Manual diagnostics were conducted by mapping values of the manual measurements to the condition status according to the power transformer assessment code. Simulated diagnostics were carried out by inputting the same values of measurement to the fuzzy logic model to obtain the corresponding transformer condition status. Figure 8 presents the assessment results of the power transformers. Transformer condition status is shown as Figure 4. Output and membership functions Table 1. Fuzzy rule base (11 Fuzzy rules) Condition Status Moisture in oil Con 1 Con 2 Con 3 Con 4 Con 5 M o is tu re L e v e l VL VG L VG G M G G+M M H M+A1 A2 VH A3 Table 2. Recomended measures / actions according to standard Status Recommended Action VG No need any action to make a diagnosis G No need any action to make a diagnosis M Should be under monitoring from time to time A1 Should be under continuous monitoring A2 Need any action to make diagnosis and analysis A3 Should be out of operation for deep analysis V. Rexhepi and P. Nakov. / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 22 the result of manual diagnostics and simulated diagnostic using fuzzy logic. The same results are listed in Table 3. The error is calculated by subtraction the manual diagnostic value from the corresponding fuzzy logic diagnostic value. From the error values, it can be calculated that the mean error is zero and the root mean squared error (RMSE) is 1.62. Explanation of the first raw in Table 4 is as follows: 1st condition assessment; moisture in oil = 3 ppm; moisture level = 7 ppm; and condition assessment = 4.07 ppm. From the data, the result only emerges the 1st rule, where according to this rule, it can be concluded that IF the moisture in oil is condition 1 AND the moisture level is in very low operation THEN the condition assessment is very good, that means the transformers are in very good condition and do not need any diagnosis or analysis. Other raws in Table 4 have a similar explanation. Discussion of the cases related to the status of the power transformers parameters such as moisture in oil is helpful in achieving results for defining the transformer status during their work. It is also Table 3. Condition assessment results Status Number of Transformer Error Fuzzy Logic Diagnostic Manual Diagnostic VG 2 4 -2 G 55 53 2 M 7 5 2 A1 0 2 -2 A2 0 0 0 A3 1 1 0 Total 65 65 0 Figure 5. Assessment model by the fuzzy logic method Figure 6. Definition of the structure rules V. Rexhepi and P. Nakov / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 23 important to perform other analysis. Through the fuzzy logic model, condition assessment or diagnosis of the power transformers can be carried out automatically. Measurements values of moisture in oil and moisture level are fed to the fuzzy logic model, and the transformer condition status is automatically obtained. VI. Conclusion A fuzzy logic model for transformers condition assessment was designed based on moisture content. The fuzzy logic model was applied to 65 transformers condition assessment in order to classify them into 5 statuses, i.e. Very Good, Good, Medium, Alarm 1, Alarm 2, and Alarm 3. The simulation result using the fuzzy logic model was compared with the manual assessment result based on the standard/code. The error between the simulation result and the manual assessment result was calculated. The mean value of error was zero, and the root means square error (RMSE) was 1.62. These results demonstrated that the proposed fuzzy logic model had provided good assessment performance. It can be assumed that fuzzy logic technique constitutes an efficient method for understanding transformer status over the key parameters. It also reflects the connectivity between on-line measurements and time to time measurements in a function to achieve the most accurate and reliable results. Table 4. Examples of some diagnostics using fuzzy logic CA Moisture in oil (ppm) Moisture level (ppm) CA (ppm) Effective Rule Status 1st 3 7 4.07 1 VG 2nd 16 18 14.0 5 G 3rd 20 19 17.2 5 M 4th 24 24 20.3 9 A2 5th 35 35 47.0 11 A3 Figure 7. Three-dimensional graphic of the moisture level in the oil Figure 8. Condition assessment monitoring of the moisture in oil to the power transformers 4 53 5 2 0 1 2 55 7 0 0 1 0 10 20 30 40 50 60 very good good medium alarm - 1 alarm - 2 alarm - 3 C o n d it io n a ss e ss m e n t o f p o w e r tr a n sf o r m e r s Power transformer numbers Simulation by Fuzzy Logic Measurements V. Rexhepi and P. Nakov. / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 17–24 24 Acknowledgement This work was supported by the Kosovo Transmission System - KOSTT. References [1] Jose Luis Martinez, “Condition Assessment of Power Transformers: A Practical Methodology Approach”, 23rd International Conference on Electricity Distribution, June 2015, CIRED, Lyon, France. [2] Nikolina Petkova, Petar Nakov, Valeri Mladenov, “Real Time Monitoring of Incipient Faults in Power Transformer”, Electricity Distribution, Springer, pp. 221 – 240, Berlin Heidelberg, March 2016. ISBN: 978-3-662-49434-9. [3] Ioana Făgarăşan, Sorina Costinas, Sergiu St. Iliescu, “Monitoring and Diagnosis Methods for high Voltage Power Transformers”, U.P.B. Sci. Bull. Series C, Vol. 70, No.3, 2008, Romania. [4] Lefeng Cheng, Tao Yu, “Dissolved Gas Analysis Principle- Based Intelligent Approaches to Fault Diagnosis and Decision making for Large Oil-Immersed Power Transformers: A survey”, Energies, MDPI, April, 2018. [5] S. Tenbohlen, J. Aragon-Patil, M. Fischer, Z.D. Wang, M. Schäfer, I Höhlein Atanasova, “Investigation on Sampling, Measurements and Interpretation of Gas-In-Oil Analysis for Power Transformers”, 42nd International Conference on Large High Voltage Electric Systems,2008, CIGRE, 2008. Paris, France. [6] Shweta Taneja, Bhawna Suri, Himanshu Narwal, Anchit Jain, Akshay Kathura, Sachin Gupta, “A New approach for data classification using Fuzzy Logic”, Cloud System and Big Data Engineering (Confluence), 2016th International Conference, IEEE, January 2016. [7] Lili Zhao, Xueming Li, Ming Ni, Tianyu Li, Yameng Cheng, “Review and prospect of hidden failure: protection system and security and stability control system”, Journal of Modern Power Systems and Clean Energy, pp. 1 – 9, April, 2015. [8] J. P. Deane, Francesco Gracceva, Alessandro Chiodi, Maurizio Gargiulo, Brian P.O. Gallachoir, “Assessing power system security. A framework and a multi model approach”, International Journal of Electrical Power & Energy Systems, Volume 73, December 2015, pp. 283-297. [9] Vezir Rexhepi, “An Analysis of Power Transformer Outages and Reliability Monitoring”, 4th International Conference on Power and Energy Systems Engineering, CPESE, Energy Procedia, Volume 141, September, 2017, pp. 418-422. [10] Wimonmas Bamrungsetthapong and Adisak Pongpullponsak, “Parameter Interval Estimation of System reliability for Repairable Multistate Series-Parallel System with Fuzzy Data”, The Scientific World Journal, May, 2014. [11] Charles M. Grinstead, J. Laurie Snell, Introduction to probability.. American Mathematical Society, USA, July, 1997, pp. 405 – 452. [12] M. Srinivasa Rao, V. N. A. Naikan, “Reliability analysis of repairable systems using system dynamics modeling and simulation”, Journal of Industrial Engineering International, September, 2014, Springer Berlin Heidelberg. [13] G. Levetin, “Computational intelligence in reliability engineering evolutionary techniques in reliability analysis and optimization”, Springer-Verlag Berlin Heidelberg, pp. 11-18, 2007. [14] Jean Sanchez, Mladen Banovic, “A General Overview of Power Transformer Diagnosis”, April, 2014, Research Gate. [15] Guowei Yang, Jia Xu, “A New Fitting Scattered Data Method Based on the Criterion of Geometric Distance”, 2nd AASRI Conference on Computational Intelligence and Bioinformatics, Vol 6, 2014, pp. 41-48. [16] Ivan Nunes da silva, Danilo Hernane Spatti, Rogerio Andrade Flauzino, Luisa Helena Bartocci Liboni, Silas Franco dos Reis Alves, “Artificial Neural Network Architectures and Training processes” in ,Artificial Neural Networks, Springer International Publishing Switzerland, 2017, pp. 21-28. [17] Mark Hudson beale, Martin T. Hagan, Haward B. Demuth, Neural Network Toolbox User’s, Matlab, The Math Works, Inc,2015. [18] Arturas Kaklauskas, “Intelligent Decision Support Systems”, Intelligent Systems, 2015, Springer, doi: 10.1007/978-3-319- 13659-2_2. [19] Rafik Aziz Aliev, “Fundamentals of the Fuzzy Logic-Based Generalized Theory of Decisions”, Fuzzy Sets and Fuzzy Logic, Springer, Berlin, Heidelberg, 2013, pp. 1-64. [20] Mohd Hilmi Hasan, Izzatdin Abdul Aziz, Jafreezal Jaafar, Lukman AB Rahim, Joseph Mabor Agany Manyiel, “A Comparative Study of Mamdani and Sugeno Fuzzy Models for Quality of Web Services Monitoring”, (IJACSA) International Journal of Advanced Computer Science and Applications, Vol.8, No. 9, 2017. [21] LHossein Hassani & Jafar Zarei, “Interval Type-2 fuzzy logic controller design for the speed control of DC motors”, Systems Science & Control Engineering, Published by Taylor & Francis, 2015. [22] Guoqiang Xia, Guangning Wu, Bo Gao, Haojie Yin and Feibao Yang, “A New Method for Evaluating Moisture Content and Aging Degree of Transformer Oil-Paper Insulation Based on Frequency Domain Spectroscopy”, Energies, 2017, 10(08), 1195; MDPI, Basel, Switzerland, August, 20017. [23] CIGRE, Guide on Transformers Intelligent Condition Monitoring (TICM) Systems, Working Group A2.44. September, 2015. [24] CIGRE, Guidelines for Life Management Techniques for Power Transformers, Draft Final Report Rev. 2, 22 June 2002. [25] Mark Tostrud, Brian D. Sparling, Ty A. Foren, “Monitoring, Diagnoostics, or Changing Condition?”, PTPiREE Conference May 12 – 14th, 2015, Gdannsk, Poland. http://cired.net/publications/cired2015/papers/CIRED2015_0024_final.pdf http://cired.net/publications/cired2015/papers/CIRED2015_0024_final.pdf http://cired.net/publications/cired2015/papers/CIRED2015_0024_final.pdf http://cired.net/publications/cired2015/papers/CIRED2015_0024_final.pdf https://doi.org/10.1007/978-3-662-49434-9_9 https://doi.org/10.1007/978-3-662-49434-9_9 https://doi.org/10.1007/978-3-662-49434-9_9 https://doi.org/10.1007/978-3-662-49434-9_9 https://www.scientificbulletin.upb.ro/Arhiva_2008/Seria_C/Nr3C-2008.pdf https://www.scientificbulletin.upb.ro/Arhiva_2008/Seria_C/Nr3C-2008.pdf https://www.scientificbulletin.upb.ro/Arhiva_2008/Seria_C/Nr3C-2008.pdf https://www.scientificbulletin.upb.ro/Arhiva_2008/Seria_C/Nr3C-2008.pdf https://doi.org/10.3390/en11040913 https://doi.org/10.3390/en11040913 https://doi.org/10.3390/en11040913 https://doi.org/10.3390/en11040913 https://e-cigre.org/publication/D1-204_2008-investigation-on-sampling-measurement-and-interpretation-of-gas-in-oil-analysis-for-power-transformers https://e-cigre.org/publication/D1-204_2008-investigation-on-sampling-measurement-and-interpretation-of-gas-in-oil-analysis-for-power-transformers https://e-cigre.org/publication/D1-204_2008-investigation-on-sampling-measurement-and-interpretation-of-gas-in-oil-analysis-for-power-transformers https://e-cigre.org/publication/D1-204_2008-investigation-on-sampling-measurement-and-interpretation-of-gas-in-oil-analysis-for-power-transformers https://e-cigre.org/publication/D1-204_2008-investigation-on-sampling-measurement-and-interpretation-of-gas-in-oil-analysis-for-power-transformers https://e-cigre.org/publication/D1-204_2008-investigation-on-sampling-measurement-and-interpretation-of-gas-in-oil-analysis-for-power-transformers https://doi.org/10.1109/CONFLUENCE.2016.7508041 https://doi.org/10.1109/CONFLUENCE.2016.7508041 https://doi.org/10.1109/CONFLUENCE.2016.7508041 https://doi.org/10.1109/CONFLUENCE.2016.7508041 https://doi.org/10.1109/CONFLUENCE.2016.7508041 https://doi.org/10.1007/s40565-015-0128-9 https://doi.org/10.1007/s40565-015-0128-9 https://doi.org/10.1007/s40565-015-0128-9 https://doi.org/10.1007/s40565-015-0128-9 https://doi.org/10.1016/j.ijepes.2015.04.020 https://doi.org/10.1016/j.ijepes.2015.04.020 https://doi.org/10.1016/j.ijepes.2015.04.020 https://doi.org/10.1016/j.ijepes.2015.04.020 https://doi.org/10.1016/j.ijepes.2015.04.020 https://doi.org/10.1016/j.egypro.2017.11.053 https://doi.org/10.1016/j.egypro.2017.11.053 https://doi.org/10.1016/j.egypro.2017.11.053 https://doi.org/10.1016/j.egypro.2017.11.053 http://dx.doi.org/10.1155/2014/275374 http://dx.doi.org/10.1155/2014/275374 http://dx.doi.org/10.1155/2014/275374 http://dx.doi.org/10.1155/2014/275374 https://bookstore.ams.org/iprob/ https://bookstore.ams.org/iprob/ https://bookstore.ams.org/iprob/ https://doi.org/10.1007/s40092-014-0069-3 https://doi.org/10.1007/s40092-014-0069-3 https://doi.org/10.1007/s40092-014-0069-3 https://doi.org/10.1007/s40092-014-0069-3 https://doi.org/10.1007%2F978-3-540-37372-8 https://doi.org/10.1007%2F978-3-540-37372-8 https://doi.org/10.1007%2F978-3-540-37372-8 https://doi.org/10.1007%2F978-3-540-37372-8 https://www.researchgate.net/publication/303075801_A_General_Overview_of_Power_Transformer_Diagnosis https://www.researchgate.net/publication/303075801_A_General_Overview_of_Power_Transformer_Diagnosis https://doi.org/10.1016/j.aasri.2014.05.007 https://doi.org/10.1016/j.aasri.2014.05.007 https://doi.org/10.1016/j.aasri.2014.05.007 https://doi.org/10.1016/j.aasri.2014.05.007 https://doi.org/10.1007/978-3-319-43162-8_2 https://doi.org/10.1007/978-3-319-43162-8_2 https://doi.org/10.1007/978-3-319-43162-8_2 https://doi.org/10.1007/978-3-319-43162-8_2 https://doi.org/10.1007/978-3-319-43162-8_2 https://www.mathworks.com/help/pdf_doc/nnet/nnet_ug.pdf https://www.mathworks.com/help/pdf_doc/nnet/nnet_ug.pdf https://www.mathworks.com/help/pdf_doc/nnet/nnet_ug.pdf https://doi.org/10.1007/978-3-319-13659-2_2 https://doi.org/10.1007/978-3-319-13659-2_2 https://doi.org/10.1007/978-3-319-13659-2_2 https://doi.org/10.1007/978-3-642-34895-2_1 https://doi.org/10.1007/978-3-642-34895-2_1 https://doi.org/10.1007/978-3-642-34895-2_1 http://dx.doi.org/10.14569/IJACSA.2017.080948 http://dx.doi.org/10.14569/IJACSA.2017.080948 http://dx.doi.org/10.14569/IJACSA.2017.080948 http://dx.doi.org/10.14569/IJACSA.2017.080948 http://dx.doi.org/10.14569/IJACSA.2017.080948 http://dx.doi.org/10.14569/IJACSA.2017.080948 https://doi.org/10.1080/21642583.2015.1013644 https://doi.org/10.1080/21642583.2015.1013644 https://doi.org/10.1080/21642583.2015.1013644 https://doi.org/10.1080/21642583.2015.1013644 https://doi.org/10.3390/en10081195 https://doi.org/10.3390/en10081195 https://doi.org/10.3390/en10081195 https://doi.org/10.3390/en10081195 https://doi.org/10.3390/en10081195 https://doi.org/10.3390/en10081195 https://e-cigre.org/publication/630-guide-on-transformer-intelligent-condition-monitoring-ticm-systems https://e-cigre.org/publication/630-guide-on-transformer-intelligent-condition-monitoring-ticm-systems https://e-cigre.org/publication/630-guide-on-transformer-intelligent-condition-monitoring-ticm-systems https://e-cigre.org/publication/ELT_208_5-guide-for-life-management-techniques-for-power-transformers https://e-cigre.org/publication/ELT_208_5-guide-for-life-management-techniques-for-power-transformers https://docplayer.net/20049006-Monitoring-diagnostics-or-changing-condition.html https://docplayer.net/20049006-Monitoring-diagnostics-or-changing-condition.html https://docplayer.net/20049006-Monitoring-diagnostics-or-changing-condition.html