| 基于改进Elman神经网络的氧化铝浓度控制建模 |
| |
| 引用本文： | 李界家,陈广其.基于改进Elman神经网络的氧化铝浓度控制建模[J].科技广场,2011(9). |
| |
| 作者姓名： | 李界家 陈广其 |
| |
| 作者单位： | 沈阳建筑大学信息与控制工程学院,辽宁沈阳,110168 |
| |
| 摘 要： | 铝电解过程是一个非常复杂的非线性、时变和大滞后的工业过程体系,因而采用常规的控制方法很难达到良好的控制效果。针对此问题本文提出了采用改进的Elman神经网络对其进行建模,介绍了改进Elman神经网络结构及其学习算法;分析了影响氧化铝浓度的主要因素,并根据实际情况确定了输入层和中间隐层的维数,从而确定了模型的结构。通过对现场采集的数据进行了仿真,仿真结果表明：与常规Elman相比,神经网络收敛速度和稳定性上都有明显提高,得到了令人满意的结果。
|
| 关 键 词： | 改进Elman神经网络 建模 仿真 |
Based on Improved Elman Neural Network of Alumina Concentration Control Modeling |
| |
| Abstract: | The aluminum electrolysis process is a very complicated nonlinear,timevarying and Large time delay industrial process system,so it is difficult to use conventional control method reach the good control effect,this paper put forward to solve this problem with the improved Elman neural network model to its,introduces the improvement Elman neural network structure and learning algorithm;Analysis of the influence of the main factors of alumina concentration,and according to the actual situation of the input layers and determine the middle of hidden layer of dimension,so as to determine the structure of the model.Through the collection of the data simulated,and the results show that：the conventional Elman neural network convergence speed and stability have increased significantly,the results obtained. |
| |
| Keywords: | Improve Elman Neural Network Modeling Simulation |
| 本文献已被 万方数据 等数据库收录！ |
|
