信息融合在石化故障诊断信息采集方面的研究

Research for Fault Diagnosis Method of Data Mining Techniques

本文的目的在于为石化生产故障诊断提供真实、可靠的数据信息,主要的研究内容为通过对现有多传感器的采集过程的问题分析,以预期差值贝叶斯算法为实现方法,解决特征级,决策级信息触合层次,以石化生产过程中常压炉鼓风机为测试实例,建立贝叶斯分类模型,提高采集数据的精度和真实性。

The purpose of this paper is to provide the real diagnosis,reliable data and information for the fault of petrochemical production,the main content is to analysis the process of the existing multi sensor problem,the implementation method is to use the expected difference Bias algorithm, to solve the feature level,to decide the level information fusion levels,take the test case from the petro chemical production process of atmospheric furnace blower,set up of the Bias classification model, to improve data accuracy and authenticity.