Accurate and non-contact measurement of volume percentages of oil-water fluids using microstrip sensors independent of the volume of sample using artificial neural network

Abstract

Due to their moderate sensitivity, extremely cheap cost, ease of fabrication method, and, more crucially, the fact that they are non-invasive, planar microwave sensors have attracted a lot of interest from both industries and academics over the past years. These intriguing properties drive this field's research toward opening up a wide range of applications that go beyond oil and gas to include biological, material sensing, pollution monitoring, and other industrial uses. The main focus of this research is on the simulation and fabrication of a high-sensitivity, very small, and repeatable microwave sensor to measure volume fractions of oil and water in real-time. This sensor is designed by Ansys HFSS software and is made on the RT/Duroid 5880 (with εr = 2.2, thickness = 0.787 mm, loss tangent of 0.0009). In a polylactic acid (PLA) box made using a 3D printer, oil and water with different volume percentages will be placed on the microwave sensor in non-contact conditions. To determine volume percentages independent of the volume of the samples, different samples were analyzed in volumes of 5 ml, 10 ml, and 15 ml. The developed sensor includes two passing bands, and when exposed to crude oil with varying amounts of water, the frequencies of these bands, their insertion loss, and their prominence in these frequencies change. Due to the non-linear variations in the insertion loss, frequency, and prominence value of the two passbands, the MLP neural network is used in this study over other approaches for identifying the objective parameter. The MLP neural network's output was the water volume percentage, and its inputs were variations in the frequency, insertion loss, and prominence of the two passbands in the transmission response. Thanks to microwave sensors and artificial neural networks, volume fractions could be detected with high accuracy, independent of the volume of samples. The suggested microwave sensor could be a highly effective way to measure volume percentages in the oil sector because of its high accuracy, compact size, simplicity of transportation, non-contact feature, etc.