A new method for using the UV-Visible spectrum of oil to evaluate the paper insulation quality of used transformers

Power transformers are one of the most vital and expensive devices in the electricity transmission and distribution network. Therefore, monitoring and evaluating their operational condition is a vital issue for managers in the electricity industry to prevent failure in the power distribution network, maintain their highly reliable performance, manage their assets, and increase the productivity of the network. The results of many research works have indicated that the useful life of a transformer is directly related to the life of its paper insulation system. The measurement of furan compounds released into the insulation oil due to the degradation of cellulose in insulation paper is a common method for assessing the operational condition of the transformers. High-performance liquid chromatography (HPLC) has been introduced as one of the relatively well-established methods for measuring furan compounds according to ASTM-D-5837. Regarding the need for high profession and expertise for operating and using HPLC system as well as the high price of this system, it is inevitable to develop fast, simple, cheap, and user-friendly systems and techniques with high precision without any need for high operator profession. The use of ultraviolet-visible (UV-Vis) spectroscopy has been proposed as a new, effective, and economically preferable alternative method for measuring the concentration of cellulose degradation compounds in insulation oil. However, the excessive turbidity of the oil in older transformers caused by the cellulose degradation products exerts a relatively high ambiguity on the results of this technique and may make its results somehow unreliable. In this paper, a new method is proposed to improve the optical spectrum results and solve the turbidity problem of the used oil samples. This method is composed of the dilution of oil samples with adequate solvents to overcome their high turbidity in combination with the fuzzy logic to evaluate the UV-Vis results. The proposed method was implemented on oil samples taken from 23 transformers. A fuzzy logic algorithm is used to estimate the furan compound contents from the optical spectrum of the oil. The comparison between the results of the method developed in this work and the results of HPLC analysis reveals a relative error as low as 11%, proving that the developed method can be a safe, fast, and cost-effective solution for electricity distribution network operators to assess and monitor the status of network transformers.