数据挖掘在船舶推进系统故障诊断中的应用.docVIP

摘 要 数据挖掘是近年来随着人工智能和数据库技术的发展而出现的一门新兴技术，实质是从大量数据集合中有效地自动地发现隐含的、先前未知的、对决策有潜在价值的信息。这些传统的数据挖掘技术的一个重要特点是需要足够大的样本数据，数据必须足够大才能够获得有价值的信息。但在某些情况下，即使有足够大的样本数据也不一定能获得有价值的知识。如样本数据量不大或有残缺、整体数据规律复杂但局部数据却有很强的规律性等此类情况下，传统的数据挖掘技术很难获得有价值的信息。 灰色数据挖掘的思想精髓就是将灰色系统理论的优点作用于数据挖掘领域，使得数据挖掘发现的知识更加有效可信。根据数据挖掘和灰色理论的特点，灰色数据挖掘的基本思想就是利用灰色理论的概念和技术，结合数据挖掘系统成型的、自动化的知识发现技术，去构建“少样本、贫信息”系统的知识自动发现信息系统。将灰色关联度分析方法应用于船舶柴油机推进系统的故障诊断，克服了一般诊断方法需要大量样本、收敛慢、求解全局极值相对困难等缺点。 本文应用成熟的数据挖掘技术并结合灰色理论，构建灰色数据挖掘体系结构。主要是把柴油机作为一个灰色系统，对其进行故障诊断实际上就是利用系统的己知信息来确定未知信息，从而对柴油机的状态进行综合评价和分析判断。首先进行灰色系统建模、灰色模型预测、灰色关联度分析等；其次，从状态信号中提取出恰当的特征参数，构成能够反映柴油机正常状态和各种故障状态特征的基准模式；最后，计算出待检模式（待诊断状态）与各基准模式之间的关联度，根据最大关联度原则，判断出待检状态与哪种基准状态的关联最大，进行状态识别，确定出最可能出现的故障。 本文实验说明，，，Abstract Data mining is a new emerging technology with the development of artificial intelligence and database technology in recent years. In essence, data mining technology efficiently and automatically discover information from the collection of large amounts of data, which is previously unknown implicitly and is potentially valuable for decisions. An important feature of these traditional data mining techniques is that it is in the need of a large enough sample data, which must be large enough to be able to obtain valuable information. However, even if there is a large enough sample data, it does not necessarily have access to valuable knowledge in some cases. If the amount of sample data is not or has defects, the overall data on the law is complex but there is a strong regularity of local data and other such cases. It is difficult to obtain valuable information of traditional data mining techniques. The essence of gray data mining is to make the advantage of the role of gray system theory in the field of data mining, and it makes the discovery of knowledge more effective and credible. According to the theory of data mining and the characteristics of gray, the basic idea of gray data mining is to use the concepts and techniques of gray theory, combined with data mining system form