A implementation of Principal Component Analysis (PCA) Algorithm for determining the Functional Age of Power Transformer, for the Paper Titled "Reliable Estimation of Dissipation Factor of In-service Power Transformer", by Debmalya Pramanik (ORCiD:0000-0002-3955-8170) and Dr. Arijit Baral (ORCiD:0000-0002-1905-9059).

Insulation failure is a severe threat to high voltage equipment, and its protection using a reliable and efficient diagnostic tool has always been the interest to power utilities. Many traditional and newer techniques are available. Due to the complex aging process of oil-paper insulation, experts generally perform assessments after carefully evaluating measurement data. The paper presents a methodology to analyze recovery voltage measurement data to estimate aging sensitive performance parameters (dissipation factor).

power transformer, dissipation factor, tan delta, return voltage, recovery voltage, central time constant, principal component analysis, regression, oil moisture, initial rate, machine learning, curve fitting

Significant figures related to the paper is added here for reference. Images files are available here, and the overall flowchart of the proposed algorithm and PCA is created using draw.io founded by Gaudenz Alder.

Significant Figures from Conference Paper

Fig.: RVM Spectrum of trf1		Fig.: The Scree Plot to determine Optimal Components
image		image
Fig.:		Fig.:
image		image
Fig.:		Fig.:
image		image
Fig.:		Fig.:
image		image
Fig.:		Fig.:

Figure 1 Two-Electrode Model for Capturing RV Data


Figure 2 RVM Spectrum of trf1


Figure 3 The Scree Plot representing the Percentage of Explained Variance of all the Individual Principal Components calculated from PCA considering all the Transformer Parameters


Figure 4 First Principal Component (PC-1) vs tan 𝛿


Figure 5 PC-1 against Dissipation Factor with Class Label based on User-Defined Boundaries


Figure 6 Proposed Curve to Estimate tan 𝛿 w.r.t. PC-1


Figure 7 Final Proposed Polynomial Equation to Determine tan 𝛿 considering an Error Band of 0.25 𝜎^2

This is licensed to © Debmalya Pramanik, Arijit Baral MIT License. If you find this document useful, please cite the original paper as (or refer to citation file):

D. Pramanik and A. Baral, "Reliable Estimation of Dissipation Factor of In-service Power Transformer," 2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon), Mysuru, India, 2022, pp. 1-6, doi: 10.1109/MysuruCon55714.2022.9972517.

@INPROCEEDINGS{9972517,
  author={Pramanik, Debmalya and Baral, Arijit},
  booktitle={2022 IEEE 2nd Mysore Sub Section International Conference (MysuruCon)}, 
  title={Reliable Estimation of Dissipation Factor of In-service Power Transformer}, 
  year={2022},
  volume={},
  number={},
  pages={1-6},
  keywords={Voltage measurement;Fitting;Estimation;High-voltage techniques;Aging;Oil insulation;Reliability;power transformer;dissipation factor;tan delta;return voltage;recovery voltage;central time constant;principal component analysis;regression;oil moisture;initial rate;machine learning;curve fitting},
  doi={10.1109/MysuruCon55714.2022.9972517}}

Principal Component Analysis (PCA) tries to find the axes with the maximum variance [1]. The decomposition.PCA() function is written using the mathematical formulation and step-by-step guide provided by Sebastian Raschka.

[1] Raschka, S. (2015). Python Machine Learning. Packt Publishing Ltd.

Paper is still under review and modifications, thus the content may change significantly.