Artificial neural network to predict the growth of Leptospirillum ferrooxidans in 9K defined medium

LAVALLE, A; CURIA, L.; LAVALLE, L.; GIAVENO, A.; DONATI, E.

International Journal of Engineering Research and Applications

IJERA

Año: 2012 vol. 2 p. 1406 - 1416

2248-9622

An artificial neural network (ANN) model was carried out to predict the cell concentration and ferrous iron concentration in the growth of strains of Leptospirillum ferrooxidans in batch culture, at different temperatures and pH values. A custom network with three interacting blocks was developed to carry out the prediction. The temperature, pH and time were fed as inputs to the network. In each block, a Multi Layer Perceptron was used and a tapped delay line was added in the first block. The appropriate architecture of the neural network model was determined through several steps of training and testing. The final ANN model was found to provide an efficient and a robust tool in predicting two variables simultaneously, in complex conditions such as non-linear and time-variant biological processes.