Advanced distributed fiber-optic liquid flow rate sensing based on optical carrier microwave interferometry

• Establish a mathematical model between dip frequency shift and liquid flow rate. • Propose an ANN-assisted demodulation method to achieve liquid flow rate sensing. • Realize spatially distributed liquid flow rate sensing based on the OCMI technique. • Utilize a single ANN model for the demodulation of multiple OCMI-based FPIs. Liquid flow is widespread in industrial practices, and accurate measurement and control of its flow rate contribute to improved production and process safety. In this paper, advanced distributed fiber-optic liquid flow rate sensing is realized utilizing optical carrier microwave interferometry (OCMI). The liquid flow-induced strain theory for double-coated optical fibers is deduced, and the response of the dip frequency shift in microwave interferograms to the liquid flow rate is explored. Numerical simulations are performed to investigate the factors affecting the axial displacement of the fiber core. Fluid experiments are conducted by combining an OCMI sensing system and a liquid flow loop. The experimental results indicate that the commonly used dip frequency tracking demodulation method cannot effectively measure the liquid flow rate due to the minor changes caused by the liquid flow. Therefore, the artificial neural network (ANN) is introduced and an ANN model is developed to assist in the implementation of OCMI-based distributed liquid flow rate sensing. To the best of our knowledge, it is the first instance of utilizing a single machine-learning model to demodulate multiple OCMI-based sensors simultaneously. The prediction results for multiple Fabry-Perot interferometers are in good agreement with the actual values, with the mean absolute error and mean absolute percentage error being <0.07 m 3 /h and 6.4 %, respectively. The study demonstrates the feasibility and effectiveness of ANN for distributed sensing, which brings a new distributed demodulation scheme for the OCMI technique.