Measurement of particle size distribution in suspension based on artificial neural network

This paper proposes a method for measuring the particle size distribution in suspension, which combined with artificial neural network and ultrasonic attenuation effect. The artificial neural network is used for active learning instead of the traditional inversion algorithm, which can effectively solves the problem that the traditional measurement method is too dependent on the selection of the inversion algorithm and the calculation of the theoretical model is large. Ultrasonic attenuation experiments were carried out on the suspension, using a focused ultrasonic measurement system to obtain an attenuation signal with the information of the particles in suspension. The eigenvalues are extracted from the denoised signal, and 15 eigenvalues such as rise time, peak time and attenuation coefficient are obtained, to form the input feature matrix of the neural network model. The sieving method was used as a control test to obtain a theoretical value of the particle size distribution, and also as a theoretical output matrix of the neural network model. In addition, three suspension samples with different distribution methods were used as test samples to evaluate the proposed method. The experimental results show that the results obtained by the neural network are in good agreement with the results obtained by the sieving method, the measurement efficiency is improved, and the measurement time is greatly reduced.