Article Dans Une Revue Journal of Engineering Research and Reports Année : 2025

Application of Regression Analysis to Explore the Relationship between Combustible Gas in Insulating Oil of Power Transformers

Résumé

This paper firstly explains the theory of regression analysis - model, least squares method, coefficient of determination, model assumptions, significance test, etc., and how to calculate through actual data to carry out what these essential nouns –such as R2 , SSR, SSE, SST, F, and t and derive relevant parameters to being analysis and interpretation. Secondly, understand what those parameters are, the author took some data from references to calculate. The results are presented in reports through the EXCEL application software SPSS system to clearly, the same time it provides analysis and interpretation. Finally, the combustible gases in the insulating oil of power transformers - hydrogen (H2), methane (CH4), ethane (C2H2), ethylene (C2H4), acetylene (C2H6) and carbon monoxide (CO), these gases were used to diagnose what condition for transformer operation its normal or abnormal, they are important roles in diagnose. Historical case data from Taiwan Power The company was selected, and a regression analysis was performed to understand the relationship between the various gases. The study found that C2H4 had no effect on the increase or decrease of the amount of H2 gas. This paper attempts to arouse the awareness and interest of cross-disciplinary professionals in regression analysis and thus discover new methods for oil-gas diagnosis. Based on the principle of technology sharing, the research results were written into a paper as a reference by scholars and maintenance personnel in the field of power engineering.

Fichier non déposé

Dates et versions

hal-05203289 , version 1 (08-08-2025)

Identifiants

  • HAL Id : hal-05203289 , version 1

Citer

Ming-Jong Lin. Application of Regression Analysis to Explore the Relationship between Combustible Gas in Insulating Oil of Power Transformers. Journal of Engineering Research and Reports, 2025, 27 (8), pp.181-185. ⟨hal-05203289⟩
20 Consultations
0 Téléchargements

Partager

  • More