Prediction of the Local Particle Velocity in the Spout Region of Cylindrical Spouted Beds Using an Artificial Neural Network

The local particle velocity is an important hydrodynamic parameter in the design and optimization of gas-solid systems, such as the spouted beds. Predicting the particle velocity in the spout region is often complex, and developing a simple and reliable method for estimating this parameter is both valuable and highly beneficial for engineers and system designers. In the present study, the intelligent predictive approach of a multi-layer perceptron neural network (MLP-NN) is used for the first time to investigate the feasibility of measuring the local particle velocity within the spout region of spouted beds. Accordingly, ۲۲۴ measured data points are collected, covering a broad range of factors such as bed diameter, inlet diameter, static bed height, cone angle, minimum spouting velocity, and inlet gas velocity for Geldart D particles with varying densities and diameters. The data are then used for training, validating, and testing the intelligence-based model. The MLP-NN-based model shows a strong predictive capability for estimating particle velocity in the spouted bed, with an Average Absolute Relative Error (AARE) of ۱۳.۳۲%, an R۲ value of ۰.۹۸۴۲, and a Root Mean Squared Error (RMSE) of ۰.۰۳۳ for the test dataset. In addition, for all data, the model achieves an AARE of ۶.۳۰%, an RMSE of ۰.۰۰۹۸, and an R۲ value of ۰.۹۹۴۲. Furthermore, in the present study, a sensitivity analysis is conducted on the data to determine the degree of influence of the input parameters on the target function.

کلیدواژه ها:

Hydrodynamics ، Spouted beds ، Particle velocity ، Multilayer Perceptron algorithm