Back to search

ENERGIX-Stort program energi

Transformer Insulation Life Assesment

Awarded: NOK 4.8 mill.

The project was going to investigate how the aging models developed in laboratory research fit with what is experienced in reality, i.e. on live transformers. This was done by: - Examining the aging of the windings in scrapped, obsolete transformers, - Collecting live material from transformers, and examine whether the equilibrium curves used today fits the reality. This should be done in a rig built in a previous project. One would also examine the equilibrium curves for aging indicators such as water-soluble acids, furans and alcohols shown to correlate with aging. - Characterizing live material and age this further in the laboratory to check if established aging kinetic models corresponds with reality. Several transformers are examined, providing knowledge of how aging can vary. This project developed a recommendation on how simple, cheap material sampling can be done on transformers that is wrecked or discarded to learn more about ageing from service. There is now an initiative to establish a database for transformer scrapping data in Norway, which with time will be a reference in terms of reinvestment analysis. In this work, we work within a CIGRE working group. Work on the equilibrium curves for water-soluble acids, which together with water contributes greatly to accelerate pa per aging, faced major problems. In an attempt to develop measurement procedures one failed to extract all acids added to the paper. Work on this was extended, and it was found that the acids entered into chemical bonds (esterification) with cellulose. This important result was published in 2014 in an IEEE journal. To find suitable transformers that were decommissioned from which one could sample "live" material proved difficult. Several attempts stranded because scrap dealers did not keep control of the opening of the scrapped transformer. Ultimo 2013, we succeeded with the help of ABB to find material. This is tested in a partitioning rig for establishing equilibrium curves. Partitioning curves for service aged systems are established for water, furanes, methanol and ethanol and published in a technical report. For the water soluble carboxylic acids, that catalysis hydrolytic ageing the significance of esterification is documented. The esterification of acids in the cellulose reduces the amount of free acids that contribute ageing. This complicates any attempt to establish quantitative models for effects from the acids on ageing. Cellulose sampled from this transformer tested in the laboratory. Results indicate that should consider simplifying the kinetic ageing models that are based on laboratory studies of singular ageing mechanisms. For transformers with both water and oxygen present a simplified scheme is suggested. A report that sums up the Norwegian research and put forward our thoughts on tools for man aging an aging transformer park is written and acknowledged by the CIGRE technical committee. It will be published at the CIGRE session in Paris in August this year.

The project addresses the potential of a prolonged life of transformers without increasing risk by applying good diagnostic methods to reinvest at the correct moment. The basic mechanisms that govern the ageing of cellulose have been revealed in fundament al laboratory research projects, and factors (e.g. water, acids, oxygen and temperature) driving the ageing are identified. Still it is difficult to apply this knowledge because of large uncertainties exist in both measuring the ageing drivers in oil and pressboard, and in relating these to the parameters in the proposed Arrhenius formulas for calculating the ageing rates. The basic idea behind the projects is to make a link between the present state-of-the-art basic knowledge and the real-world condition s in a service aged transformer. In the project we will therefore work on service aged cellulose materials; in these materials the conditions will be determined and then ageing studies performed to be able to calculate parameters in the Arrhenius formulae from e.g. water and acid content in the paper. Two unique test rigs are available: -One laboratory rig for studying partitioning of ageing accelerators and diagnostic markers between cellulose and oil to be better able to assess paper condition from sta ndard oil sampling. -The other rig - for in-service installation - for contamination of paper samples placed in an external container to the condition of the paper inside the transformer by circulating oil out from the transformer past the sample and bac k into the transformer. This can be used to reduce uncertainties in condition assessment in traditional indirect methods. The general idea is to increase the applicability of theoretical knowledge by calibrating with in-service experience. A methodology and tools will be developed so that one can start with life-assessment of fleets - with a considerable uncertainty - , and then for selected units improve the accuracy by applying more accurate investigations

Funding scheme:

ENERGIX-Stort program energi