Size estimation of biopolymeric beads produced by electrospray method using artificial neural network

In the present study, an artificial neural network (ANN) was used to determine the size of alginate/chitosan polymeric particles prepared with different operating conditions via ionotropic gelation process by the electrospray method. For evaluation of particle size, effective operational parameters such as polymer concentration, nozzle diameter, flow rate and voltage were investigated and used for training, testing, and validation parts of the neural network. In this design model, a multilayer perceptron (MLP) is used and the optimal number of hidden layers for this network was 2. The number of optimal neurons in the input layer, the first hidden layer and the second hidden layer also, were 4, 16 and 9, respectively. The results show the high accuracy of the proposed model compared to the experimental data in the biocomposite formation process. The sensitivity analysis confirm that all the considered input parameters, directly affect the biopolymeric beads size. The values of mean squared error (MSE), mean relative error (MRE) and mean absolute error (MAE) for the suggested model, were 0.00206, 1.53986, and 0.01772, respectively.

Automated summary

In this study, an artificial neural network (ANN) was used to determine the size of alginate/chitosan polymeric particles prepared with different operating conditions. The proposed model showed high accuracy and sensitivity to the operational parameters in the biocomposite formation process.