面向城市污水处理厂自动控制的数据挖掘研究

项目名称： 面向城市污水处理厂自动控制的数据挖掘研究

项目编号： No.50808071

项目类型： 青年科学基金项目

立项/批准年度： 2009

项目学科： 金属学与金属工艺

项目作者： 李晓东

作者单位： 湖南大学

项目金额： 20万元

中文摘要： 鉴于城市污水处理厂故障诊断和控制规则的获取直接影响到出水的稳定和能耗的降低，分别采用数据挖掘的方法对城市污水处理厂进行故障诊断和异常征兆模式识别，并对控制模型和规则的数据挖掘进行研究，此外用数据挖掘的方法解决传感器的最优投入问题，研究建立城市污水处理厂数据挖掘的统一构架和方法。此课题对提高城市污水处理厂的自动控制水平有着十分积极的意义。主要研究内容包括： （1）故障诊断模型和规则以及异常征兆模式的数据挖掘。本研究在实验室建立小型的城市污水处理设施，利用数据挖掘方法得出该设施的故障诊断规则和异常征兆模式。 （2）控制模型和规则的数据挖掘。对城市污水处理设施运行数据进行挖掘，并将挖掘到的控制模型和规则在污水处理设施上进行验证。 （3）传感器投入评估。通过数据挖掘的方法获得传感器投入的最优解，达到传感器投入成本、诊断精度和控制效果三个性能指标综合最优的目标。

中文关键词： 城市污水处理厂；数据挖掘；故障诊断；控制规则；传感器投入

英文摘要： The city sewage treatment plant fault diagnosis and control rules directly impact on water stability and reduced energy consumption. So data mining methods were used in fault diagnosis and pattern recognition of abnormal signs on municipal sewage treatment plant. Control models and the rules were obtained based with data mining research. In addition, data mining methods were used to solve the optimal input of the sensor. The study established a unified framework of urban sewage treatment plants data mining. The research has a very positive sense in improving the level of automatic control of the city sewage treatment plant. The main contents include: (1) Data mining of fault diagnosis model and rules, as well as abnormal signs of the mode. In this study, a sewage treatment facility was built in the laboratory. Then data mining methods was used to get the facility rules of fault diagnosis and abnormal signs mode. (2)Data mining of control model and rules. Data mine with urban sewage treatment facilities operating data, then used mined models and rules to control sewage treatment facilities. (3) Assessment of sensor costs. Data mining methods were used to obtain the optimal solution of sensor costs, diagnostic accuracy and control effect .

英文关键词： Municipal wastewater treatment plant; Data Mining; Fault diagnosis; Control rules; Sensor costs