THE PURPOSE: This paper is devoted to the interpretation methods of dissolved gases chromatography results of power transformer equipment. Today, due to the emergence of a large number of various data analytics systems for power equipment technical state assessment, the issues associated with interpretation of dissolved gases chromatography results are quite acute, since the risk of incorrect diagnosis is quite high with automated recognition of technical state patterns. METHODS: This paper addressed six methods for analyzing the gases: Rogers method, Dornenburg method, IEC standard, nomogram method, ETRA method and Duval triangle, as well as expert judgment approach. These methods differ in the composition of gases used for the corresponding analysis, detected faults, the need to use the boundary concentration criterion, the visual representation of the results, including the ability to visualize dynamics of the indicators changing. As a case study, the problem of defect identification in 110 kV operated power transformers is analyzed on the basis of the presented methods for giving interpretation of the results of dissolved gases chromatography and ways to solve the stated problems. RESULTS: For the available measurements of the gases’ concentration, calculations were made to identify outliers provided by each of the methods described above. When using statistical approach, outliers were detected in all cases where defects were detected by at least one of the methods. An outlier was also detected in one case, which does not correspond to any defect. When applying the block diagram, outliers are also detected in the cases which correspond to a certain defect. However, the method also revealed outliers in "normal" samples. Thus, the block diagram, despite its simplicity, is too sensitive to outliers detection, which can lead to wrong elimination of non-erroneous data. The statistical method, on the other hand, produces a result that is most consistent with the expectations of experts. For this reason, at this stage of research, this method of rejecting measurements is considered. CONCLUSION: The results of the dissolved gases chromatography obtained in this paper indicate that the situations that are difficult to identify even for an expert or a group of experts occur quite often. This fact, in turn, indicates that when applying data analytics systems, it is not only necessary to take into account additional data, but also to create training samples for diagnoses based on precedent information.