Parameter sensitivity analysis of a spouted bed

A three-layer neural network, based on 192 sets of experimental data obtained by the authors, was built in order to simulate the influence of four main process parameters on the minimum and stationary spouting velocity and on the pressure drop in a spouted bed. The simulations with the neural net are in good agreement with experimental learning data, and the overall average absolute error is 5.43 %. The neural net was used in the framework of a Monte Carlo simulation using CRYSTAL BALL (R). The result of 100,000 trials revealed the percentage contributions of the process parameters to the variance of the selected state variables. Since there is significant uncertainty concerning the design equations of spouted beds, the results of this work should be useful for the design of such devices.