METHOD OF PREDICTING PROPERTIES OF POLYMER PRODUCT

摘要

A method for predicting properties of a polymer product by using the retention time distribution as a process parameter is provided to maintain accuracy of prediction regardless of changes in process parameters, and apply the method to the loop reactor type propylene products. A method for predicting properties of a polymer product comprises the steps of: filtering the process parameters obtained from a reactor by using a moving average filter; inputting the filtered process parameters to a neural network theory properties anticipating model to calculate the instant properties of the polymer product; and calculating the accumulated properties of the polymer products from the calculated instant properties by using the mixing rule that calculates the current accumulated properties anticipating value based on the instant properties anticipating value and previous accumulated properties anticipating value, wherein the properties include melt index, xylene solubility and co-monomer(ethylene or 1-butene) content.