【摘要】 对案例推理(CBR)的模型和算法进行了研究和分析,并应用到水泵机组的故障诊断中。利用振动传感器提取特征,运用基于逻辑回归的假设检验方法进行特征约简,并结合SQL Server 2012建立了案例知识库;针对案例推理中相似案例的不唯一性,在多数重用的基础上引入随机打破距离连接的方式,最后结合R语言对方法进行验证,并将诊断结果以表格的形式显示出来。仿真结果表明,该方法可以对故障进行有效、快速的诊断,从而为水泵的故障诊断提供了一种新方法。更多还原

【Abstract】 The model and algorithm of CBR are studied and analyzed,and applied to fault diagnosis of water pump. The use of vibration sensors picks up fault features,and through the hypothesis test based on Logistic regression method to reduce the extracted fault features.The Case knowledge base is established with SQL Server 2012. For non-uniqueness of Similar Cases in Case-based Reasoning,the random break distance connection is introduced on the basis of most reuse,and with the R language to verify the method.Finally,the diagnostic results are displayed in tabular form.The results show that this method can diagnose the fault effectively and quickly,which provides a new method for fault diagnosis of water pump.更多还原