Dejan Susa

A field study of aging in paper-oil insulation systems

The paper used to insulate the windings of power transformers is mostly made from wood pulp, a cellulosic material. Over decades the paper is slowly attacked by water, oxygen, oil acids, and high temperatures and eventually degrades to the point where it is no longer an effective insulator. The transformer is then likely to fail. Power utilities need to know when a transformer is nearing the end of its useful life in order to plan its replacement. However, a problem with monitoring the condition of the paper within a transformer is that it may be difficult to obtain a sample to test. Furthermore, a particular sample may not accurately reflect the overall paper condition. A power transformer operating in Australia failed in 2010. Thus we had the opportunity to study the paper condition at various points within the transformer and evaluate the validity of the current understanding of paper aging. In this article we discuss the mechanisms of cellulose degradation, and the associated equations, and apply them to the paper insulation in the failed transformer.

A Simple Model for Calculating Transformer Hot-Spot Temperature

A simple model for calculating the hot-spot temperature is introduced. The model is based on the hot-spot to ambient gradient. The model considers the changes of the oil viscosity and winding losses with temperature. The results are compared with temperatures calculated by IEEE Annex G method and measured results at varying load for the following transformer units: 250-MVA ONAF, 400-MVA ONAF, and 605-MVA OFAF.

Temperature responses to step changes in the load current of power transformers

Comprehensive load/thermal testing was performed on a 400/400/125 MVA, ONAF-cooled transformer. It was shown that the local hot-spots in the windings, core, and structural parts rise much faster at load increase than what an exponential function based on the time constant of top oil would predict. This paper analyzes these fast rises more in detail for hot-spots in windings, core, yoke clamps, and tie plates. An alternative mathematical model of the winding hot-spot response is presented. Different cooling modes, like ONAF and OFAF, are dealt with. Also different modes of oil circulation through the windings, in horizontal (zig-zag) or axial cooling ducts, are compared. The results are verified by fiberoptic installations and tests on several large power transformers in the range of 250 - 650 MVA. Hot-spot responses predicted by the IEEE Loading Guide, Annex G, are also compared to measured values.

The effect of acid accumulation in power-transformer oil on the aging rate of paper insulation

In this article we explain how acids are formed in the oil and in paper insulation and their involvement in paper degradation. The acidity of the oil reached at the end of a number of paper-aging experiments is reported, and relationships between it and temperature, moisture, oxygen, degree of polymerization (DP), and aging time are presented.

Thermal analysis of two transformers filled with different oils

In this article the authors present the results of their studies comparing the thermal performance of two nominally identical transformers, one filled with standard mineral oil and the other with natural ester Envirotemp FR3 oil.

Aging Rate of Grade 3 Presspaper Insulation used in Power Transformers

Paper materials are used as the insulation of power transformers. Over time these materials slowly degrade, until they reach a point when they no longer function effectively as transformer insulation, and the transformer has then reached its end of life. The aging rate of paper is affected by temperature, water, oxygen and acids. Investigations have been performed previously on Kraft and Kraft thermally upgraded types of paper. However, a different type of paper used in transformer insulation, Grade 3 presspaper (which contains cotton), has not been extensively tested and compared to Kraft paper. In these experiments we studied the aging rate of Grade 3 presspaper and compared it to our previous studies of Kraft paper. Traditionally, the aging rate of paper has been studied in sealed vessels. The problem with this approach is that the chemical environment within the vessel will change during aging, and so the aging rate will be affected. In our experiment setup we controlled the water and oxygen content to more accurately determine the aging rate. Similarly to Kraft paper, the aging rate of Grade 3 presspaper with the same water content increased with oxygen content in the oil. Life curves were developed based on the water content of the paper and the oxygen content of the oil.

Temperature distribution in a disc-type transformer winding

The winding hot-spot temperature is an extremely important factor limiting the loading capability of a power transformer. High temperature also accelerates the aging of paper insulation and thus reduces the lifetime of a transformer. Ohmic losses in the winding are the main source of heat inside a transformer. Precise knowledge of these losses is very important to determine the winding temperature distribution. Furthermore, in an oil-immersed power transformer, oil is used both as insulation and as liquid coolant. In a disc-type winding, a non-uniform coolant flow, in horizontal ducts, can contribute to local overheating. This article describes a new approach to estimate the winding temperature distribution, in which losses are computed by the finite element method whereas the coolant circulation and the temperature distribution are obtained from a thermal-hydraulic network representation of the winding.

On-line assessment of power transformer ageing accelerators

This paper discusses the field application of the improved paper filter system used for non-intrusive monitoring of power transformer insulation contamination level. The filter has been applied to three transformer units. The first results indicated promising trend between the levels of the measured contamination and ageing accelerators in the filter and in the transformers. Also, the monitoring procedures are further discussed to improve measuring consistency.