Measurement of in-plane thermal diffusivity of thin film using a regression methodology based on Bayesian optimization algorithm
Lock-in thermography is widely used in determining the thermal properties of films by extracting the amplitude and phase of thermal waves. However, achieving high phase detection accuracy typically requires sophisticated infrared (IR) cameras. In this work, we present a regression methodology based...
Uloženo v:
| Vydáno v: | Review of scientific instruments Ročník 96; číslo 8 |
|---|---|
| Hlavní autoři: | , , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
United States
01.08.2025
|
| ISSN: | 1089-7623, 1089-7623 |
| On-line přístup: | Zjistit podrobnosti o přístupu |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Shrnutí: | Lock-in thermography is widely used in determining the thermal properties of films by extracting the amplitude and phase of thermal waves. However, achieving high phase detection accuracy typically requires sophisticated infrared (IR) cameras. In this work, we present a regression methodology based on Bayesian optimization to determine the in-plane thermal diffusivity of thin films. Unlike conventional approaches that rely on lock-in algorithms, where amplitude and phase are treated as intermediate quantities, our method directly incorporates time-sequential thermograms into the regression process. The thermal diffusivity and phase offset are automatically extracted by minimizing the mean absolute error between measured and simulated normalized temperatures. This method is validated through measurements on a stainless-steel film, demonstrating that accurate results can be achieved using only a few thermograms per modulation period. These findings highlight the feasibility, robustness, and reduced system requirements of the proposed approach, making it particularly promising for characterizing composite films and advanced thermal interfacial materials. |
|---|---|
| Bibliografie: | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ISSN: | 1089-7623 1089-7623 |
| DOI: | 10.1063/5.0285181 |