Online Estimation of Liquid Viscosity and Density Based on Artificial Neural Network Approach

Abstract

This paper presents an online estimation of density and viscosity from the sensor response based on artificial neural network approach. The working principle of sensor is based on the vibration of piezoelectric actuated cantilever beam at resonance. In addition to that, an Artificial Neural Network approach is utilized for the online estimation of unknown viscosity and density of liquids, based on the dynamic characteristics of the resonant cantilever sensor. The estimation of density and viscosity of unknown liquid is done by the observing the dynamic characteristics of the cantilever vibration which greatly depends on the Q factor and frequency of the cantilever vibration at resonance. The dynamic characteristics of the piezo actuated cantilever sensor is analysed using Euler–Bernoulli beam equations. In this study, the back propagation neural network (BPNN) has been used for the online estimation of the liquid viscosity and density. By the neural network training, it is possible to estimate the accurate density and viscosity of unknown liquid with dynamic response of the beam at any conditions. The output obtained from neural network models are compared with analytical results.