High-accuracy crude oil purity detection using a non-invasive microwave sensor with proximity-coupled patch antennas

This paper presents a non-invasive microwave-based sensor for high-precision purity detection of crude oil based on the proximity-coupled rectangular microstrip patch antennas. The sensor employs mutual electromagnetic coupling between two closely placed slotted rectangular-shaped antennas in order to extract dielectric metrics from oil-water mixtures. This design contrasts with more traditional single-antenna sensors, which rely entirely on S 11 or return loss, as our design utilizes additional scattering parameters (S 21 , S 22 ) for improved sensitivity and selectivity when characterizing oil purity. This antenna system operates in two primary frequency bands, specifically 6.9–7.6 GHz for detecting dielectric-based bandwidth variation and 1.6 GHz for characterizing signal transmission. The sensor is fabricated on the FR4 substrate and experimentally validates its capabilities with a wide range of water-oil mixtures. The results showed good agreement with simulated and measured data, and validated the high accuracy of the proposed sensor. To improve sensing precision and decrease nonlinearity due to operation at higher frequencies, artificial neural networks (ANN) models are used in the data processing step. Altogether, the ANN model increased prediction accuracy and model reliability for oil clarity classifications. Measured results indicate an S11 < − 10 dB in the 6.9–7.6 GHz with peak gain of 7.6 dBi, and the ANN regression yielded R²=0.995 with low prediction error, confirming enhanced detection accuracy.