研究目的
To develop a robust algorithm for thickness identification of thin nonconductive coatings using terahertz time domain spectroscopy.
研究成果
The terahertz method is promising for assessing non-conductive coatings, offering advantages over other NDT methods. It enables thickness measurements in less than 100 ms per location, but further progress in scanning speed and data processing is necessary to make it a reliable and effective tool. Calibration curves were developed for thickness estimation, though they require linearization for limited ranges.
研究不足
The THz method has limited resolution, making signal components from thin coatings nearly invisible. The thinnest oil films showed issues due to vessel bottom necrosis. The calibration curves are not completely linear over the full thickness range (30 μm to 2 mm), and additional data processing and scanning speed improvements are needed for reliable application.
1：Experimental Design and Method Selection:
The study utilized a terahertz time domain spectroscopy (THz-TDS) system in reflection mode with two configurations (v-type and collinear) to measure coating thickness. An adaptive LMS IIR filter with a delay element was used for signal processing to estimate thickness.
2：Sample Selection and Data Sources:
A metal cylindrical vessel filled with transformer oil was used as the test specimen to simulate nonconductive coatings. The oil layer thickness was varied by adding 0.25 ml increments, resulting in thickness changes of over 60 μm.
3：25 ml increments, resulting in thickness changes of over 60 μm.
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Terahertz time domain spectroscope (model Tray 4000 by Picometrix/Luna Inc.), metal vessel, transformer oil.
4：Experimental Procedures and Operational Workflow:
Measurements were conducted using the THz-TDS system in both configurations. Signals (A-scans) were collected, processed using the adaptive filter to extract coating response, and thickness was estimated by locating peaks in the processed signals.
5：Data Analysis Methods:
The least mean square (LMS) algorithm was applied to filter signals and calculate error. Calibration curves were generated from the processed data to relate signal characteristics to coating thickness.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
