确定氧化钇微细粉制备条件的神经网络方法

Neural networks method for parameter determination of yttrium oxide ultrafine making

氧化钇微细粉的制作过程中 ,微细粉的颗粒粒径及品质质量涉及到制作过程中的料液浓度、沉淀温度、草酸浓度、分散剂用量、热分解温度等多种影响因素 ,须经大量的探索和试验才得以确定每一项工艺参数。本文提出建立人工神经网络模型 ,将已有的实验参数作为神经网络的学习样本 ,训练网络模型 ,然后利用经过学习后的模型的容差和智能特性 ,为氧化钇微细粉的制备确定所需的工艺条件。

In processing of yttrium oxide, the diameter of the particle and quality of the products are related with many factors, such as liquid concentration, deposition temperature, pyrogenation temperature. To determine each technical parameter, many researches and experiments are needed to try. In this paper, a method is presented to help determining intelligently the technical parameters for yttrium oxide making. Based on model of artificial neural networks, input the real experiment parameters as the learning samples to train the model of neural networks, then use trained model, which has tolerance and intelligence, to predict technical conditions. The high correction of model is proved by simulation experiment.