Sensor technology plays a pivotal role in various aspects of the petroleum industry. The conventional quartz crystal microbalance (QCM) liquid-phase detection method fails to discern the viscosity and density of solutions separately, rendering it incapable of characterizing the properties of unknown liquid solutions. This presents a formidable challenge to the application of QCM in the petroleum industry. In this study, we aim to assess the feasibility of exclusively utilizing a single QCM sensor for liquid viscosity measurements. Validation experiments were conducted, emphasizing the influence of temperature and solution concentration on the viscosity measurement results. The results indicate that the QCM liquid viscosity response model can achieve viscosity measurements in the temperature range of 20 to 60 °C and concentration range of 10%–95% glycerol solution using a single QCM, with a maximum error of 7.32%. Simultaneously, with the objective of enhancing the model’s measurement precision, as an initial investigation, we employed a backpropagation neural network combined with genetic algorithm (to optimize the measurement data. The results demonstrate a substantial improvement in the measurement accuracy of the QCM sensor, with a root mean square error of 3.89 and an absolute error of 3.07% in predicting viscosity values. The purpose of this research was to extend neural networks into the evaluation system of QCM sensors for assessing the viscosity properties of liquid in the oil industry, providing insights into the application of QCM sensors in the petroleum industry for viscosity measurement and improving measurement accuracy.
