Effects of multiple factors on the photoacoustic detection of glucose based on artificial neural network

In this paper, to study the effect of multiple factors on the photoacoustic detection of glucose, a Nd: YAG 532nm pumped optical parameters oscillator (OPO) pulsed laser induced photoacoustic detection system were established. The lateral model was used to capture the photoacoustic signals of glucose by using the ultrasonic transducer. The photoacoustic signals were averaged in 512 times. In the experiments, the photoacoustic experiments of different concentrations, temperatures, laser energies and flow velocities for glucose aqueous solutions were performed. Meanwhile, the effects of concentration, temperature, energy, flow velocity on the photoacoustic detection of glucose were investigated. Not only the relationships between the each factor and the photoacoustic detection of glucose were built, but also the coupled relationship between the multiple factors on the photoacoustic detection of glucose was also established by using the artificial neural network. In the artificial neural network, three levels neural network includes four input parameters and one output target was used. Prediction results show that the coupled relationship can better present the practical condition of glucose detection, which can offer the potential research value for the photoacoustic detection of diabetes mellitus.