摘要
针对机床热误差补偿系统中温度传感器故障问题,采用神经网络方法建立各温度传感器间关系模型,提出基于模型的温度传感器故障诊断方法,通过所建立的数学模型来恢复故障传感器数据。笔者给出了相关温度传感器数据神经网络模型的建立方法以及故障传感器检测的策略。仿真实验对比表明:在无传感器故障情况下热误差补偿系统可将20μm内机床热误差控制在1.66μm范围内,而当有传感器故障情况下机床热误差可控制在1.69μm内,相差无几。研究工作表明笔者所提出的传感器失效检测和恢复方法是可行的。
In this paper,neural network method is used to establish the relationship model between the temperature sensors and their faults in a thermal error compensation system.We propose a model-based temperature sensor fault diagnosis method and use an established mathematical model to restore fault sensor data.The related neural network model approach and the detection strategy is given for the fault sensor.Finally,through contrast of two kinds of situations in the simulation experiment,we find that one can compensated 20 μm thermal deformation to 1.66 μm with no fault sensor,and the other can compensated 20 μm thermal deformation to 1.66 μm with fault sensors,showing that the method for fault sensor detection and its data restore is feasible.
出处
《机械科学与技术》
CSCD
北大核心
2010年第6期805-808,共4页
Mechanical Science and Technology for Aerospace Engineering
基金
江苏省精密与微细制造技术重点实验室开放课题基金项目(JSPM200704)资助
关键词
热误差补偿
神经网络
传感器故障诊断
恢复
thermal error compensation
neural network
sensor fault diagnosis
recovery
