On accuracy of a mutually coupled ladder network model high-frequency parameters identification for a transformer winding using gray wolf optimizer method

It is not a secret that the identification of the high-frequency ladder network model (LNM) parameters for the transformer winding is a crucial task. This paper aims to present the application of one of the latest swarm intelligence algorithms, namely, gray wolf optimizer (GWO) for the identification of the high-frequency LNM parameters for the transformer winding.

The physical realizability of a unique ladder network is ensured and it is based on the frequency response analysis and some terminal measurements of a transformer winding.

The test results on a real transformer winding indicated that the identified model, which is improved and detailed, is superior in terms of representing the physical behavior of the transformer winding in high frequency. The efficiency and the superior capabilities of the proposed GWO method are demonstrated by comparing the later with recent algorithms, such as particle swarm optimization-simulated annealing and crow search. Results show that the proposed GWO is better in terms of optimal solution and fast convergence.

The identified LNM model is mutually coupled and able to reflect the physical behavior of the transformer winding in high frequency; therefore, it is more reliable for the diagnosis and analysis.

Contribution has been offered for the identification and the diagnosis of the transformer winding, using robust algorithms for future research.