Transformer modelling and diagnostics.

Funding: Public-private partnership project

The activity is part of the MODITRANS project, funded by MCIN/AEI/10.13039/501100011033 by the European Union "NextGenerationEU"/PRTR "4.


File reference number: CPP2021-008580




Distribution transformers are robust machines, with a long service life. However, the transition towards decarbonisation of electricity generation is leading to a technological evolution that may affect its useful life. For example, connection transformers in photovoltaic plants and wind-turbines are subject to harmonic currents derived from the associated power electronics. As a result, these transformers experience an increase in thermal, mechanical and dielectric stresses. In transformers located in transformation centers, dispersed photovoltaic generation increases fluctuations in the low voltage network. To continue complying with the regulations, it is necessary to introduce smart transformers, which regulate the low-voltage-network through an on-load tap changer (OLTC). OLTC integration can affect the reliability of transformers. To improve decision-making in relation to the health and maintenance of transformers operating in these new scenarios, techniques based on artificial intelligence are being developed. This artificial intelligence needs to be fed by transformer parameters, such as measurements relative to the condition of the liquid dielectric or the number of OLTC operations


Project goals:

In this context, the main objective of the project is the investigation and implementation of a methodology that allows the determination of a transformer health index. This health index will take into account parameters associated with the condition of the oil, pressure and level of the liquid dielectric, hot spot temperature (HST), number of OLTC operations and electrical operating parameters of the transformer.

To develop this health index, the project shall also address the following challenges in the field of transformer monitoring:

  • Develop an oil condition sensor based on the measurement of the dielectric constant that provides metrics associated with both the real part and the imaginary part.
  • Thermally model the transformer in the presence of harmonics and develop a tool that provides HST information.

Additionally, The project will also investigate systems to reduce the connection currents of the transformers and how to improve the calculation of the transformers from the thermal modeling data.


Ceit's role in the project:

Ceit leads the development of the oil condition sensor. The tasks of Ceit include the design of the sensor and the associated electronics that allow obtaining metrics of the real and imaginary part of the dielectric constant. The objective is to develop an intelligent oil filler cap that will allow oil condition monitoring to be incorporated, not only in current designs, but also in already operating transformers.