Iron Loss Evaluation via Infrared Thermography Imaging Using Thermal Inverse Methods

Achieving high energy efficiency is the primary objective in all areas of electrical energy applications. In this context, the design of high-efficiency energy conversion devices, such as electrical machines, relies critically on accurate knowledge and quantification of iron losses in their magnetic core, which must be considered from the earliest design stages. Yet, iron losses in magnetic materials are notoriously difficult to assess reliably under real operating conditions. Conventional fluxmetric methods, including the standardized Epstein frame and single sheet tester, are widely used but rely on dedicated sample preparation and are therefore destructive. Moreover, they provide only a global evaluation and fail to capture local iron losses, highlighting the need for alternative approaches. In magnetic materials, iron losses are dissipative phenomena, that are entirely dissipated as heat within the core material. Thus, this fundamental principle underpins the present study. The approach consists of using the heat dissipated by ferromagnetic materials under an AC magnetic excitation field as a direct means of characterizing their iron losses. The objective is to assess how these losses evolve under different operating conditions in terms of: excitation frequency and waveform; manufacturing-induced mechanical stresses; mechanical stresses during operation; the magnetic anisotropy of certain materials, and temperature-induced changes in magnetic properties over time (magnetic aging), while analyzing their spatio-temporal distribution within the material. This is of particular interest for electrical machines, where both the spatial distribution and magnitude of iron losses depend on the operating conditions. This involves using infrared (IR) thermography to carry out non-contact temperature field measurements in a magnetic steel strip subjected to an AC magnetic field. The measured temperature data are then used to reconstruct the dissipated power through thermal modelling of the heat transfer within the material (using thermal quadripoles) and various inverse mathematical methods. Unlike conventional approaches, this method is designed to allow local and real-time estimation of iron losses under realistic magnetic excitation conditions. To ensure the accuracy and robustness of the proposed method, different inverse heat conduction problem solutions have been considered: Tikhonov regularization, Kalman filter, truncated singular value decomposition (TSVD), and the function specification method (FSM). The latter are implemented and compared with the conventional iron loss measurement method to assess their effectiveness. Finally, this approach leads to the development of an innovative and complementary method, that, once properly calibrated, enables the quantitative measurement of iron losses using only a quantum IR camera and an inverse thermal model.

Work In Progress