The liquid-insulated power transformer is the most critical and expensive component of the power supply system. In order to improve monitoring of the state of health of these devices (transformers) installed on the power grid in the bank of the large metropolis of Abidjan, dissolved gas analysis (DGA) was used. Conventional interpretation techniques were developed to diagnose transformer oil. The database we used consists of those of the Compagnie Ivoirienne d'Electricité (CIE) in 2018. It includes 25 samples containing analysis of the five main dissolved gases (H2, CH4, C2H2, C2H4, C2H6). The five defect classes we considered, namely partial discharge (PD), D1 and D2 electrical defects, T1&T2 and T3 thermal defects, are taken from the IEC and IEEE. For interpretation, five methods were adopted, namely the IEC, Doenenburg, Rogers criteria using the ppm gas concentration ratio technique, the Duval triangle using percentage gas concentrations and the IEEE criterion using ppm gas concentrations. The highest success rate of 72% was obtained when using the Duval triangle criterion. But the greatest consistency with a rate of over 70% was observed using the IEC ratio criterion.