Fault tree analysis (FTA) has been proven to be a very important tool and has been successfully applied to safety/reliability studies in nuclear, chemical, military, space industries/systems. Hitherto, several useful and popular FTA software/program packages have been developed, like CAFTA+, FAUNET, RiskSpectrum, SAPHIRE, RiskA etc. Minimum Cut Set (MCS) method is the most commonly used traditional FTA method. However, it suffers from low efficiency when solving remarkably large fault trees (FTs). To overcome the shortcomings of the traditional method, several new techniques are proposed such as Binary Decision Diagram (BDD), Zero-suppressed Binary Decision Diagram, (ZBDD) Petri Net (PN), Bayesian Network (BN) and Dynamic Uncertain Causality Graph (DUCG). DUCG is a newly presented Probabilistic Graphic Model to deal with systems with dynamics, uncertainties and logic cycles. DUCG is a good choice to analyze large FTs, in our previous papers, we have proved that any FT can be mapped into a DUCG graph and additional modeling and analytical power can be achieved. DeRisk is a DUCG embedded risk analysis program package written in C# for FTA and is designed as a powerful tool to assist reliability engineers. In this paper, the design schema and the main algorithms of DeRisk are introduced. DeRisk contains five parts: (1) A Graphical User Interface (GUI) Module which interacts with users; (2) A Preprocessing Module which preprocesses FTs (3) An Input Module which allows user to input necessary data by file or by command line; (4) A Calculation Module which offers qualitative/quantitative analysis; (5) An Output Module which outputs the results required by users. Some illustrative examples are used to verify the correctness and effectiveness of DeRisk.
Intelligent fault diagnosis of hot die forging press based on binary decision diagram and fault tree analysis
