Optimal Feature Oreinted Classification of One and Merged Disturbances of Power Quality Through Supervised Learning
Aslam Shaik1, A. Srinivasula Reddy2
1Aslam Shaik, Research Scholar, JNTUH, Hyderabad (Telangana), India.
2Dr. A. Srinivasula Reddy, Prof & Principal, CMREC, Hyderabad (Telangana), India.
Manuscript received on 07 June 2019 | Revised Manuscript received on 30 June 2019 | Manuscript Published on 04 July 2019 | PP: 956-964 | Volume-8 Issue-1S4 June 2019 | Retrieval Number: A11760681S419/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The emergence of power quality topic in power systems has gained an increased research interest due to its widespread applicability in different applications. Since the quality of power is devalued due to various disturbances, detection of such disturbances is required. This paper provides refined power quality disturbances classification mechanism which is associated with various compositions like the presence of noise, and accumulation of two or more disturbances. Considering the both linear and non-linear dependencies between PQDs, two optimal feature extraction techniques, Joint Mutual Information as well as Correlation based feature selection, are proposed. These optimal set of features are processed for PQDs classification by employing pair of supervised learning algorithms, Support Vector Machine (SVM) along with Decision Tree (DT). Extensive simulations are conducted over different PQ disturbances at different Signal to Noise Ratio (SNR) levels give away the performance constructiveness of proposed approach. The robustness is derived through the provision of a tradeoff between the computational time and classification accuracy.
Keywords: Power Quality Disturbances, Mutual Information, Correlation, SVM, DT, Swell, Sag, Accuracy, Computational Time.
Scope of the Article: Classification