摘要 |
The present invention relates to a method for predicting a parameter of a bitumen having a penetration nP at 25°C, said bitumen being obtained through a process (P) applied to a hydrocarbon and/or a blend of hydrocarbons, comprising the following operative steps: a) obtaining, through experimental analyses and/or databases and/or blending models, the values of N physico-chemical parameters of the hydrocarbon or blend of hydrocarbons; b) feeding the values of the N parameters obtained in step a) together with the nP value of the bitumen, to a neural network trained for correlating the same N physico-chemical parameters of a hydrocarbon or blend of hydrocarbons and the nP values of the bitumens obtained therefrom by applying the process (P) to the parameter to be predicted; c) predicting the parameter of the bitumen on the basis of the result of the correlation obtained by the neural network during step b). The present invention also relates to a system for effecting said method. |