Signed Directed Graph and Qualitative Trend Analysis-Based Method for Process Monitoring and Fault Diagnosis

A combined signed directed graph (SDG) and qualitative trend analysis (QTA) method for process monitoring and fault diagnosis is proposed. In this method, every node in the model of SDG is monitored When fault occurs, SDG is used to find possible faults and their consistent paths. For every consistent path, the trends of all the nodes on it are extracted and analyzed using QTA. And then backward reasoning according to trends of nodes (matching the trends of nodes with trends stored in the database from fault node to cause node) for every consistent path is done to get the most possible path and cause. Thus, this method combines the completeness property of SDG with the high diagnostic resolution property of QTA. The proposed method is tested on a tank-level 3 system. The result shows that this method provides fast, reliable and accurate fault diagnosis.