TY - JOUR
T1 - Piezoelectric Smart Patch Operated with Machine Learning Algorithms for Effective Detection and Elimination of Condensation
AU - Zhang, Qian
AU - Wang, Yong
AU - Wang, Tao
AU - Li, Dongsheng
AU - Xie, Jin
AU - Torun, Hamdi
AU - Fu, Yongqing
N1 - This work was supported by the National Natural Science Foundation of China under Grant 51875521, the Zhejiang Provincial Natural Science Foundation of China under Grant LZ19E050002, the Engineering Physics and Science Research Council of UK (EPSRC EP/P018998/1) and UK Fluidic Network (EP/N032861/1) Special Interest Group of Acoustofluidics, International Exchange Grant (IEC/NSFC/201078) through Royal Society and the National NSFC.
PY - 2021/8/27
Y1 - 2021/8/27
N2 - Timely detection and elimination of surface condensation is crucial for diverse applications in agriculture, automotive, oil and gas industries, and respiratory monitoring. In this paper, a smart patch based on a ZnO/aluminum (~5 μm/50 μm thick) flexible Lamb wave device has been proposed to detect, prevent and eliminate condensation, which can be realized using both of its surfaces. The patch is operated using a machine learning algorithm which consists of data preprocessing (feature selection and optimization) and model training by a random forest algorithm. It has been tested in six cases, and the results show good detection performance with average Precision = 94.40% and average F1 score = 93.23%. Principle of accelerating evaporation is investigated in order to understand the elimination and prevention functions for surface condensation. Results show that both dielectric heating and acoustothermal effect have their contributions, whereas the former is found more dominant. Furthermore, the functional relationship between the evaporation rate and the input power is calibrated, showing a high linearity (R2 = 97.64%) with a slope of ~3.6×10-5 1/(s·mW). With an input power of ~0.6 W, the flexible device has been proven effective in the prevention of condensation.
AB - Timely detection and elimination of surface condensation is crucial for diverse applications in agriculture, automotive, oil and gas industries, and respiratory monitoring. In this paper, a smart patch based on a ZnO/aluminum (~5 μm/50 μm thick) flexible Lamb wave device has been proposed to detect, prevent and eliminate condensation, which can be realized using both of its surfaces. The patch is operated using a machine learning algorithm which consists of data preprocessing (feature selection and optimization) and model training by a random forest algorithm. It has been tested in six cases, and the results show good detection performance with average Precision = 94.40% and average F1 score = 93.23%. Principle of accelerating evaporation is investigated in order to understand the elimination and prevention functions for surface condensation. Results show that both dielectric heating and acoustothermal effect have their contributions, whereas the former is found more dominant. Furthermore, the functional relationship between the evaporation rate and the input power is calibrated, showing a high linearity (R2 = 97.64%) with a slope of ~3.6×10-5 1/(s·mW). With an input power of ~0.6 W, the flexible device has been proven effective in the prevention of condensation.
KW - Flexible devices
KW - Condensation detection and elimination
KW - Lamb waves
KW - Random forest algorithm
KW - Respiration detection
UR - http://www.scopus.com/inward/record.url?scp=85114310239&partnerID=8YFLogxK
U2 - 10.1021/acssensors.1c01187
DO - 10.1021/acssensors.1c01187
M3 - Article
VL - 6
SP - 3072
EP - 3081
JO - ACS Sensors
JF - ACS Sensors
IS - 8
ER -
