基于SDG模型的控制系统故障诊断方法

利用深层定性知识模型、符号有向图模型，对生产流程中常见的单回路和串级控制系统进行了建模与故障诊断研究。提出基于假设-验证的双向推理算法，解决了传统的故障诊断方法从未涉及复杂系统中的多个控制回路对系统故障存在的屏蔽作用的问题。通过对某厂常压蒸馏装置进行基于SDG模型的故障诊断，验证了基于SDG模型的故障诊断方法应用在包含多个控制回路的复杂系统故障诊断时的有效性与可行性。

Fault diagnosis of control system based on signed directed graph model

The problem of modeling and fault diagnosis of the single loop and cascade loop control systems was studied based on the deep quantitative knowledge model, signed directed graph (SDG）.A bidirectional inference method based on supposition and validation was introduced, and the problem of the shielding of the multiple control loops in the complex system with respect to the system fault in the traditional fault diagnosis was resolved.The method of fault diagnosis based on the SDG model was applied to the actual distillation equipment under atmosphere pressure, which proved the validity and feasibility of the SDG fault diagnosis method in the complex systems with multiple control loops.