Neural network model predicts naphtha cut point. I. Wadi. OGJ, 25 November 1996, 94(48), 67--70.
Test results of a Middle East refiner's in-house neural network are reported which have shown it to be more accurate than online analysers by some 82%. Details are given of the model development, back propagation, the BP network concept, neural network applications, and model building. 4 refs.