基于核函数的加权极限学习机污水处理在线故障诊断

发布时间：2018-01-27 04:50

本文关键词： 加权极限学习机核函数在线建模污水处理故障诊断仿真实验　出处：《化工学报》2016年09期 　论文类型：期刊论文

【摘要】：污水生化处理中的运行故障会引起出水水质不达标、运行费用增高和环境二次污染等严重问题,需要及时准确地对运行故障进行诊断。考虑到污水处理过程运行状态数据的不平衡性造成故障诊断准确率下降,提出了一种基于核函数的加权极限学习机污水处理过程实时在线故障诊断方法。该方法以极限学习机为基础,采用加权的方式处理数据的不平衡特性,通过核函数的非线性映射来提高数据线性可分的程度。仿真实验证明,本文建立的污水处理在线故障诊断模型在线测试精度高,泛化性能好,模型在线更新速度快,能够比较好地满足准确性和实时性,实现对污水处理过程的在线故障诊断。
[Abstract]:The operation failure in wastewater biochemical treatment will cause serious problems such as substandard effluent quality, higher operating cost and secondary environmental pollution. It is necessary to diagnose the operation fault in time and accurately. Considering the imbalance of the running state data of the sewage treatment process, the accuracy of fault diagnosis is reduced. This paper presents a real-time on-line fault diagnosis method for wastewater treatment process of weighted extreme learning machine based on kernel function. The method is based on the ultimate learning machine and uses weighted method to deal with the unbalanced characteristics of data. The nonlinear mapping of kernel function is used to improve the separability of data line. The simulation results show that the on-line fault diagnosis model of sewage treatment established in this paper has high on-line testing accuracy and good generalization performance. The model is updated quickly on line, which can satisfy the accuracy and real time, and realize the on-line fault diagnosis of sewage treatment process.
【作者单位】：华南理工大学自动化科学与工程学院;
【分类号】：TP18
【正文快照】： 引言污水处理是一个复杂的、影响因素非常多的生化过程,污水处理厂难以保持长期稳定的运行,发生故障容易引起出水水质不达标、运行费用增高和环境二次污染等严重问题[1],所以需要对污水处理厂运行状态进行在线监控,诊断出运行故障并及时处理。污水处理过程的故障诊断数据具有

【相似文献】