| 基于自组织模糊神经网络溶解氧控制方法研究 |
| |
| 引用本文： | 许进超,,杨翠丽,,乔俊飞,,马士杰,.基于自组织模糊神经网络溶解氧控制方法研究[J].智能系统学报,2018,13(6):905-912. |
| |
| 作者姓名： | 许进超 杨翠丽 乔俊飞 马士杰 |
| |
| 作者单位： | 1. 北京工业大学 信息学部, 北京 100124;2. 计算智能与智能系统北京市重点实验室, 北京 100124 |
| |
| 摘 要： | 针对污水处理过程中溶解氧浓度难以控制的问题，提出了一种基于自组织模糊神经网络（self-organizing fuzzy neural network， SOFNN）的溶解氧（dissolved oxygen， DO）控制方法。首先，采用激活强度和神经元重要性两个评判标准，来判断神经元对网络的贡献及活跃程度。然后，对不活跃的神经元进行删减，以此来对神经网络结构进行自适应的调整，从而满足实际控制要求，提高控制精度。其次，采用梯度下降算法对SOFNN神经网络的各个参数进行实时调整，以保证网络的精度。最后，将该自组织方法用在Mackey-Glass时间序列预测中，结果表明所提出的自组织模糊神经网络具有较好的预测效果；同时将所提出的SOFNN方法在BSM1仿真平台上进行实验验证。结果表明，所提出的自组织模糊神经网络控制方法能够对溶解氧浓度进行较好地控制，具有一定的自适应能力。
|
| 关 键 词： | 污水处理 溶解氧 过程控制 神经网络 自组织 |
Dissolved oxygen concentration control method based on self-organizing fuzzy neural network |
| |
| XU Jinchao,,YANG Cuili,,QIAO Junfei,,MA Shijie,.Dissolved oxygen concentration control method based on self-organizing fuzzy neural network[J].CAAL Transactions on Intelligent Systems,2018,13(6):905-912. |
| |
| Authors: | XU Jinchao YANG Cuili QIAO Junfei MA Shijie |
| |
| Affiliation: | 1. Faculty of Information Technology, Beijng University of technology, Beijing 100124, China;2. Beijing Key Laboratory of Computational Intelligence and Intelligence System, Beijing 100124, China |
| |
| Abstract: | It is difficult to control the dissolved oxygen (DO) concentration in wastewater treatment processes. To solve this problem, this paper proposes a dissolved oxygen control method based on a self-organizing fuzzy neural network (SOFNN). First, two judging criteria, firing strength and neuron importance, were used to determine the contribution and activity of neurons to the network. Then the inactive neurons were deleted to adjust the structure of the neural network to adaptively meet the actual control requirements and improve control accuracy. Second, a gradient descent algorithm was used to update the SOFNN parameters to ensure accuracy of the neural network. Finally, the proposed algorithm was used for the Mackey-Glass time series prediction, and the results showed that the proposed SOFNN had better prediction performance. Furthermore, the proposed SOFNN method was used on the benchmark simulation model no.1 (BSM1). The results indicate that the proposed SOFNN controller can achieve a better control effect for the DO control and has a good adaptive ability. |
| |
| Keywords: | wastewater treatment dissolved oxygen process control neural network self-organization |
|
| 点击此处可从《智能系统学报》浏览原始摘要信息 |
|
点击此处可从《智能系统学报》下载全文 |
|
