基于主观贝叶斯推理的多传感器分布式故障检测融合方法

发布时间：2018-10-05 15:11

【摘要】：针对复杂数控加工中心故障预测中各传感器检测信息呈现不确定性的问题,提出基于不确定性推理的多传感器分布式检测融合算法。该算法通过利用主观贝叶斯推理,获取局部检测装置的判决规则,并选取合适的局部判决规则送到融合规则中心,将来自不同传感器的观测数据进行综合分析,最后产生全局判决。以复杂立式加工中心为对象建立测试平台,利用多传感器样本获取方法进行机床不同运行状态及运行环境下的故障样本获取试验。试验表明在含有大量不确定性信息的故障诊断系统中,基于主观贝叶斯推理的分布式检测融合算法具有故障信息识别率高、诊断速度快的优点,其诊断错误率明显低于单个传感器的诊断错误率,且诊断错误率要低于串行分布式检测融算法。
[Abstract]:A multi-sensor distributed detection fusion algorithm based on uncertainty reasoning is proposed to solve the problem of uncertainty in the detection information of each sensor in the fault prediction of complex NC machining center. By using subjective Bayesian reasoning, the decision rules of the local detection device are obtained, and the appropriate local decision rules are selected to be sent to the fusion rule center, and the observed data from different sensors are analyzed synthetically. Finally, a global judgment is produced. Taking the complex vertical machining center as the object, the test platform is established, and the multi-sensor sample acquisition method is used to obtain the fault samples of machine tools under different running states and operating environments. The experiments show that the distributed detection fusion algorithm based on subjective Bayesian reasoning has the advantages of high fault information recognition rate and fast diagnosis speed in fault diagnosis systems with a large amount of uncertain information. The diagnostic error rate is obviously lower than that of single sensor, and the diagnostic error rate is lower than that of serial distributed detection fusion algorithm.
【作者单位】：北京理工大学机械与车辆学院;北京信息科技大学现代测控教育部重点实验室;石家庄铁道大学机械工程学院;
【基金】：国家自然科学基金资助项目(51275052)
【分类号】：TG659;TH165.3

【参考文献】