Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : March 16, 2023 - March 17, 2023
To achieve the in-situ imaging of reaction-diffusion phenomena during aqueous acid-base reactions, regression models constructed based on near-infrared (NIR) spectroscopic data measured using Fourier transform infrared spectroscopy were transferred to the image construction of HCl, NaOH, and NaCl concentrations through multivariate calibration standardization methods. In order to find the optimal method of model transfer, we compared the transfer performance of three methods: direct standardization (DS), spectral space transformation (SST), and calibration model transformation based on canonical correlation analysis (CTCCA). As a result, the transfer performance of DS was the best; the root mean square errors of prediction were 0.131, 0.069, and 0.027 M for HCl, NaOH, and NaCl, respectively. This result indicated that the proposed algorithm was a promising calibration transfer method for the NIR imaging.
