A 3D CFD Simulation and Analysis of Flow-Induced Forces on Polymer Piezoelectric Sensors in a Chinese Liquors Identification E-Nose

Chinese liquors can be classified according to their flavor types. Accurate identification of Chinese liquor flavors is not always possible through professional sommeliers’ subjective assessment. A novel polymer piezoelectric sensor electric nose (e-nose) can be applied to distinguish Chinese liquors because of its excellent ability in imitating human senses by using sensor arrays and pattern recognition systems. The sensor, based on the quartz crystal microbalance (QCM) principle is comprised of a quartz piezoelectric crystal plate sandwiched between two specific gas-sensitive polymer coatings. Chinese liquors are identified by obtaining the resonance frequency value changes of each sensor using the e-nose. However, the QCM principle failed to completely account for a particular phenomenon: we found that the resonance frequency values fluctuated in the stable state. For better understanding the phenomenon, a 3D Computational Fluid Dynamics (CFD) simulation using the finite volume method is employed to study the influence of the flow-induced forces to the resonance frequency fluctuation of each sensor in the sensor box. A dedicated procedure was developed for modeling the flow of volatile gas from Chinese liquors in a realistic scenario to give reasonably good results with fair accuracy. The flow-induced forces on the sensors are displayed from the perspective of their spatial-temporal and probability density distributions. To evaluate the influence of the fluctuation of the flow-induced forces on each sensor and ensure the serviceability of the e-nose, the standard deviation of resonance frequency value (SD(F)) and the standard deviation of resultant forces (SD(Fy)) in y-direction (F(y)) are compared. Results show that the fluctuations of F(y) are bound up with the resonance frequency values fluctuations. To ensure that the sensor's resonance frequency values are steady and only fluctuate slightly, in order to improve the identification accuracy of Chinese liquors using the e-nose, the sensors in the sensor box should be in the proper place, i.e., where the fluctuations of the flow-induced forces is relatively small. This plays a significant reference role in determining the optimum design of the e-nose for accurately identifying Chinese liquors.