木材干燥窑内判断传感器有效状态方法的研究  

Research on determining effective working-state of sensor during wood drying process

作　　者：张佳薇[1] 郭润龙[1] 

机构地区：[1]东北林业大学机电工程学院,黑龙江哈尔滨150040

出　　处：《传感器与微系统》2009年第5期55-57,60,共4页Transducer and Microsystem Technologies

基　　金：国家自然科学基金资助项目(30771678);黑龙江省青年专项基金资助项目(QC07C57)

摘　　要：为了判断传感器的有效工作状态,提出了基于D-S证据理论和CMAC神经网络的信息融合方法。利用传感器在空间和时间上的冗余或互补信息,依据D-S证据理论,消除故障检测、诊断中的不确定性。采用模块化的小脑模型识别网络建立基本信任分配函数,具有最佳一致逼近特性,同时解决了D-S证据理论确定基本信任分配函数的困难和因输入参数改变影响信任分配函数构建的问题。仿真结果表明:该方法可以对木材干燥过程中传感器工作状态进行有效诊断。To determine the effective working state of sensors, a data fusion method based on D-S theory and CMAC neural networks is presented. D-S theory using the redundancy and complementary information of sensor can eliminate the uncertainty in detecting and diagnosis faults. The evidence of sensor state is obtained through a high modularity of the sensor state recognizer network with modularized CMAC, so the advantage of neural net is promoted and the bug of its unusable for change of input parameter is resolved. The simulation results show that the method can determine effective working-state of sensor during wood drying process.

关 键 词：故障诊断 数据融合 双输入单输出小脑模型神经网络 证据理论 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置]