一种适用于航天器某些故障诊断的知识获取方法

尽可能快地找出航天器在工作过程中出现的故障，采取补救措施，可减少或避免重大损失。故障诊断专家系统可帮助人们作出快速诊断，但它的诊断准确率主要取决于其知识的正确性。以往这些知识是靠领域专家给出的。本文根据航天器某些故障的特点，提出了一种处理连续变元示例的知识自动获取方法，用这种方法获取的诊断故障的规则，在实践中检验是正确的

KNOWLEDGE ACQUIRMENT FOR SPACE CRAFT DIAGNOSIS EXPERT SYSTEM

In the period of launching and flighting of a spacecraft,some faults may occured if the faults can be found as earlier as possible and some measures are adopted,some accidents and disasters can be avoided Fault diagnosis expert system can help us found the faults immediately But the diagnosing degree of accuracy is mainly depending on the knowledge the expert system possessed This knowledge is usually given by the field experts according to their experience But because of the complicated relations between the faults and the signals samped from the spacecraft,it is difficult to induce the diagnose rules According to the characters of the spacecraft,this paper proposed a knowledge acquirement method,which can deal with the continuous attributes examples This method is used to learn from the examples supplied by experiment of the spacecraft model on the ground In this method,the diagnosis rules of every fault can be got These rules are proved to be correct in practice