Estimation of moisture content distribution in porous foam using microwave tomography with neural networks

TL;DR

This paper demonstrates a neural network approach to rapidly estimate moisture content distribution in porous foam using microwave tomography, aiming for real-time control in industrial drying processes.

Contribution

The study introduces a neural network-based method for real-time moisture content estimation in microwave tomography, utilizing synthetic data for training and testing.

Findings

Neural network achieves reconstruction in less than one second.

Method accurately estimates moisture distribution in synthetic scenarios.

Potential for real-time control in industrial microwave drying.

Abstract

The use of microwave tomography (MWT) in an industrial drying process is demonstrated in this feasibility study with synthetic measurement data. The studied imaging modality is applied to estimate the moisture content distribution in a polymer foam during the microwave drying process. Such moisture information is crucial in developing control strategies for controlling the microwave power for selective heating. In practice, a reconstruction time less than one second is desired for the input response to the controller. Thus, to solve the estimation problem related to MWT, a neural network based approach is applied to fulfill the requirement for a real-time reconstruction. In this work, a database containing different moisture content distribution scenarios and corresponding electromagnetic wave responses are build and used to train the machine learning algorithm. The performance of the trained network is tested with two additional datasets.