论文部分内容阅读
研究了轧制力设定控制的人工智能实现方法。在建立训练及预测数据模式库的基础上开发了轧制力预报人工神经网络软件，同时对实际生产中的历史数据进行了模拟在线预报，达到了较高精度。另外，给出了一种利用人工神经网络进行模拟正交实验的新思路。
The artificial intelligence method of setting control of rolling force was studied. Based on the establishment of training and forecasting data pattern library, an artificial neural network software for rolling force forecasting was developed. At the same time, the on-line forecasting of historical data in actual production was simulated with high accuracy. In addition, a new idea of using artificial neural network to simulate orthogonal experiment is given.