Evaluation of power transformer health analysis by internal fault criticalities to prevent premature failure using statistical data analytics approach

•Internal faults are diagnosed using statistical data analytics approach to avoid premature failure of power transformer.•Health indices are formulated using piece wise linear functions of chemical, mechanical and electrical parameters.•Analytical hierarchy process, piecewise liner equation and residual analysis are combined for multi criteria decision making.•Residual analysis is incorporated using linear, cubic and quadratic curves to evaluate normalization. Transformers are key players at transmission and distribution level in electrical grid to deliver reliable, efficient, smooth and quality power at consumer end. Power transformers plays a crucial role in the interconnected power systems and is therefore considered as one of the most important and critical assets. Any deterioration occurring due to electrical, thermal, chemical as well as environmental stresses can be identified from health indices and immediate actions can be proposed to avoid any premature failure. This research paper includes the multi criterion based mathematical approach to identify the health indices of power transformers. The proposed approach gives more precise result compared to the conventional methods used for transformer faults diagnosis like gas ratio and traditional experiment based methods to measure health index. The multi criterion considered here are dielectric strength, acidity, breakdown voltage, dissolved gas analysis, furan compounds, dielectric dissipation factor, moisture presence, interfacial tension, winding DC resistance, tan delta etc. It becomes more reliable and proficient to measure the health indices using multi criteria decision making along with piece wise linear equations and Residual Analysis approach for accurate measurement for newly manufactured transformer as well as in service aged transformers. Further, 100 test data set have been added to analyze the performance and implications of health index of 20 completely healthy transformers, 60 partial deformed transformers and 20 complete deformed transformers or on verge of failure. This work gives detailed insights of health status and insulation condition of new transformer, in service transformers or any failure transformer.