Digital embedded refractometer with temperature compensation

Abstract

Temperature is one of the single most important factors influencing accurate refractometer readings and is one of the largest sources of error in measurement. A model based on neural networks, has been implemented to generate solution concentration data, knowing fluid temperature measurements in this solution. The neural network chosen in this study, is the hidden layer feed forward network. Its learning rule is based on the backpropagation of the error. Its training basis consists of concentration measurement, temperature data recorded every time. An in-line process measurement of refractive index (RI) works as a real-time predictive tool for the final concentration in process parameters is proposed. The acquired signal is transmitted, via the Bluetooth, to the processing and diagnostic unit. A predictive modelling using the neural networks enables us to predict its operation under various conditions within a temperature range from 0 up to 40°C.