论文部分内容阅读
针对污水处理过程出水氨氮难以在线测量的问题,文中提出了一种基于递归RBF神经网络的软测量方法来预测氨氮。首先,提取与出水氨氮相关的主元变量,剔除主元变量的异常数据。其次,利用递归RBF神经网络建立主元变量与出水氨氮的蕴含关系,完成出水氨氮软测量模型的设计。最后,将提出的出水氨氮软测量方法应用于污水处理实际运行过程,结果表明,基于递归RBF神经网络的软测量方法能够实现出水氨氮的在线预测;同时,与其他方法的比较结果显示基于递归RBF神经网络的软测量方法具有较好的预测精度。
Aiming at the problem that it is difficult to measure the ammonia nitrogen in effluent of sewage treatment process, a soft sensor based on recursive RBF neural network is proposed to predict ammonia nitrogen. First, the main variables related to effluent ammonia and ammonia were extracted and the anomalous data of the main variables were excluded. Secondly, recursive RBF neural network is used to establish the implication relationship between principal variables and ammonia nitrogen in effluent, and the design of the soft ammonia-nitrogen measurement model is completed. The results show that the soft sensor based on recursive RBF neural network can realize the on-line prediction of ammonia nitrogen in effluent, and the comparison with other methods shows that the proposed method based on recursive RBF The neural network soft-sensing method has a good prediction accuracy.