【摘要】 系统的故障诊断问题是一种多知识源问题,反映系统状态的信息有时十分广泛。为对技术系统的故障诊断问题做出更精确、更完全的估计和判决,利用数据融合技术具有分析和处理多源信息的能力,将数据融合技术与故障诊断原理融为一体,形成基于数据融合技术的故障原理,并提出数据融合故障诊断系统模型。数据融合模型分成三级结构:原始信息融合层、特征融合层、决策融合层。每一级内部又有相应子结构。前两层主要对输入信息进行预处理和特征提取,第三层主要对特征信息进行全局决策。通过将Bayes决策理论纳入DFFDSM和时空融合方法的两个例子,说明DFFDSM是一种适应性较好的模型,具有结构清晰、简单的特点。更多还原

【Abstract】 The problem of system fault diagnosis is the matter of multi-knowledge sources, which results in the wide refection of the states of system. In order to make a more accurate and more complete estimation and decision for the problem, the data fusion technology, with the capability to analyze and dispose the multi-knowledge sources, is adopted. Combining the data fusion technology with fault diagnosis theory, the theory based on the technology of data combination is developed and a DFFDSM was presented. There’re three layers of the model: original data fusion layer, feature fusion layer and decision fusion layer. Each layer has its sub-layers. The first two layers primarily handle in advance and abstract the input information. The last layer makes the whole decision for the information with features. Two examples, the application of Bayes theory in DFFDSM and the space-time fusion method, illustrate that the DFFDSM is an adoptive one with clear and simple structures.更多还原