Classificação de transformadores de distribuição de energia elétrica quanto à DHTV usando Rough Sets

Power quality has become an important research field due to two factors: the growing competitiveness in power system and the increase of non-linear loads in the system, serving as a cause of voltage distortion in distribution systems. In this sense, this work has the aim of improve power quality by indentifying power distribution transformers which have voltage total harmonic distortion (VTHD) above the brazilian limits. It was used a database with monitored data and electrical characteristics obtained during a monitoring campaign in Energy Company at Paraná (COPEL). This information was used to classify the transformers by VTHD, according to a physical division in 5 regions in Paraná State. The VTHD limit used was approved by Power System National Operator (ONS) at Brazil and it has the value of 6% for Power Systems with voltage under 69kV. It was applied the Rough Sets Theory (RST), developed by Zdzislaw Pawlak in the 80’s, which has increasing usage to Power Systems analysis in classification problems and removal of irrelevant information in databases. The application of RST was successful and it allows the acquirement of rules which makes possible the estimation of VTHD in another power distribution transformers. So, with this VTHD estimated it is possible to guide a monitoring campaign avoiding time and money losses. To validate the application of rough sets Theory the rules obtained were applied to the original database to assure the discernibility capability of created sets. The results were compared with other techniques such as quadratic scores and logistic regression, applied to the same problem, and RST had better classification capability. Key-words: Rough Sets. Power Quality. Voltage Total Harmonic Distortion. Classification. LISTA DE ILUSTRAÇÕES Figura 1 Exemplo de regiões positiva, de fronteira e negativa............................. 15 Figura 2 Definição das regiões e dos conjuntos de aproximação ........................ 16 Figura 3 – Sistema elétrico hipotético .................................................................... 17 Figura 4 Representação gráfica dos conjuntos redução e núcleo básico ............ 20 Figura 5 Representação da onda fundamental e da componente harmônica ...... 35 Figura 6 Onda fundamental somada à componente harmônica .......................... 35 Figura 7 Harmônica versus corrente harmônica .................................................. 36 Figura 8 – Período de monitoração e período de análise ...................................... 48 Figura 9 – Exemplo de DHTV – abaixo de 6% ....................................................... 49 Figura 10 – Exemplo de DHTV – acima de 6% ...................................................... 49 Figura 11 – Espectro da tensão no período de maior distorção – abaixo de 6% ... 51 Figura 12 – Espectro da tensão no período de maior distorção – acima de 6% .... 51 Figura 15 Importação do banco de dados ........................................................... 57 Figura 16 Geração dos redutos ........................................................................... 58 Figura 17 Geração de regras ............................................................................... 58 Figura 18 – Regras geradas................................................................................... 60 Figura 19 Resultado da aplicação das regras ...................................................... 61

[1]  Mo-Yuen Chow,et al.  A classification approach for power distribution systems fault cause identification , 2006, IEEE Transactions on Power Systems.

[2]  J.O.P. Pinto,et al.  Data Mining in Distribution Consumer Database using Rough Sets and Self-Organizing Maps , 2006, 2006 IEEE PES Power Systems Conference and Exposition.

[3]  Z. Pawlak Rough Sets: Theoretical Aspects of Reasoning about Data , 1991 .

[4]  Desire L. Massart,et al.  Rough sets theory , 1999 .

[5]  Huisheng Wang,et al.  Fuzzy-neuro approach to fault classification for transmission line protection , 1998 .

[6]  A. A. Fouad,et al.  Application of artificial neural networks in power system security and vulnerability assessment , 1994 .

[7]  N. Kagan,et al.  A methodology for monitoring harmonic distortions in electric power distribution systems , 2002, 10th International Conference on Harmonics and Quality of Power. Proceedings (Cat. No.02EX630).

[8]  Ping-Feng Pai,et al.  Rough set theory with discriminant analysis in analyzing electricity loads , 2009, Expert Syst. Appl..

[9]  F. Choong,et al.  Expert System for Power Quality Disturbance Classifier , 2007, IEEE Transactions on Power Delivery.

[10]  A. K. Ghosh,et al.  The classification of power system disturbance waveforms using a neural network approach , 1994 .

[11]  Maurílio Pereira Coutinho Detecção de ataques em infra-estruturas críticas de sistemas elétricos de potência usando técnicas inteligentes. , 2007 .

[12]  Jerzy W. Grzymala-Busse,et al.  Rough Sets , 1995, Commun. ACM.

[13]  Pradipta Kishore Dash,et al.  Classification of power system disturbances using a fuzzy expert system and a Fourier linear combiner , 2000 .

[14]  Waldemar Rebizant,et al.  Application of Artificial Neural Networks , 2011 .

[15]  Ilona Jagielska,et al.  An investigation into the application of neural networks, fuzzy logic, genetic algorithms, and rough sets to automated knowledge acquisition for classification problems , 1999, Neurocomputing.

[16]  Chen-Fu Chien,et al.  Using Rough Set Theory to Recruit and Retain High-Potential Talents for Semiconductor Manufacturing , 2007, IEEE Transactions on Semiconductor Manufacturing.

[17]  Y.J. Cao,et al.  Performance of the novel rough fuzzy-neural network on short-term load forecasting , 2004, IEEE PES Power Systems Conference and Exposition, 2004..

[18]  G. Lambert-Torres,et al.  Application of rough sets in power system control center data mining , 2002, 2002 IEEE Power Engineering Society Winter Meeting. Conference Proceedings (Cat. No.02CH37309).

[19]  Salvatore Greco,et al.  Rough sets theory for multicriteria decision analysis , 2001, Eur. J. Oper. Res..

[20]  T. A. Short,et al.  Power quality monitoring of a distribution system , 1994 .

[21]  Marzena Kryszkiewicz,et al.  Rough Set Approach to Incomplete Information Systems , 1998, Inf. Sci..

[22]  Ramin Yasdi,et al.  Learning Classification Rules from Database in the Context of Knowledge Acquisition and Representation , 1991, IEEE Trans. Knowl. Data Eng..

[23]  P.A. Crossley,et al.  Building Knowledge for Substation-Based Decision Support Using Rough Sets , 2007, IEEE Transactions on Power Delivery.

[24]  Madhuchhanda Mitra,et al.  A Rough-Set-Based Inference Engine for ECG Classification , 2006, IEEE Transactions on Instrumentation and Measurement.

[25]  Aleksander Øhrn,et al.  Discernibility and Rough Sets in Medicine: Tools and Applications , 2000 .

[26]  Andrzej Skowron,et al.  Rough set methods in feature selection and recognition , 2003, Pattern Recognit. Lett..

[27]  B. Kulicke,et al.  Neural network approach to fault classification for high speed protective relaying , 1995 .

[28]  James F. Peters,et al.  Classifying Faults in High Voltage Power Systems: A Rough-Fuzzy Neural Computational Approach , 1999, RSFDGrC.

[29]  Chen-Fu Chien,et al.  Rough set theory for data mining for fault diagnosis on distribution feeder , 2004 .