Abstract:

For the problem of diagnosis identification and location for hydraulic screw down system of strip rolling mill,by establishing an accurate mathmatical model,
unknown input observer method was applied.The faults were analyzed through the model in the form of system state equations,and direction vectors were determined for different faults.By taking the system faults as unknown input and applying unknown input observer and generalized residual set theory,an unknown input observer group was designed.The fault diagnosis was achieved by the generalized residual set which was robust to special fault direction.The simulation results verify the correctness and feasibility of this research.

Key words: hydraulic screw down system, fault diagnosis, unknown input observer, robust

摘要：

为解决板带轧机液压压下系统故障辨识和定位问题,通过建立系统精确数学模型,应用未知输入观测器方法进行故障诊断和隔离。在建立系统状态方程的基础上对系统故障进行分析,确定了不同故障的方向矢量;应用未知输入观测器和广义残差集理论将系统故障作为未知输入，通过设计未知输入观测器组获得对特定故障方向具有鲁棒性的广义残差集进行故障诊断;仿真及试验验证了诊断方法的正确性和可行性。

关键词: 液压压下系统, 故障诊断, 未知输入观测器, 鲁棒