Vol. 55, Issue 1, pp. 49-62 (2025)

Abstract

To measure the state of ethanol differentiation process, a fiber option sensing based process state monitoring system is designed. It includes a laser, demo module, PC processing module, fiber option sensing network, and state feedback control unit. A data mining algorithm for multi-parameter demo is proposed to accurately achieve temperature and strain field classification and combine accuracy to the different weights of temperature and strain on different positions, improving the correlation between wavelengths offset and state parameters. The experiment compared four common abnormal situations, The experiment adopts the method of collecting the temperature field inside the tank and the stress field outside the tank, with strain testing of 100–5000 με and temperature testing range of 0–120 °C. The results showed that the average temperature sensitivity after calibration was 0.0102 nm / °C, and the linearity was 0.9959. The average strain sensitivity is 0.499 pm/με, and the linearity is 0.9982. Feedback control has the ability to adjust state fluctuations online, and the feedback time varies for different types of anomalies. The temperature and strain wavelength deviations after correction for all four cases are less than ±1 °C and ±50 με.