Research On Optimization Technology Of Diagnostic Strategy Based On Multi-signal Flow Model

With the rapid development of high technology, the structure and technology of electronic equipments is becoming more and more complicated, which brings great challenges to the testing, diagnosis and maintenance. As an important technology to improve the availability of equipment and reduce life cycle cost of the system, design for testability has been paid more and more attention. Testability analysis based on multi-signal flow graph model is a kind of relatively effective methods for design for testability at present. As an important aspect in DFT, diagnostic strategy optimization design can reduce the time of fault diagnosis, the manpower of system maintenance, the costs of equipment and support, and has important significance to improve the operational readiness of system.Diagnostic strategy optimization is the test sequencing problem in fault detection and isolation, and its aim is to construct a test sequence which can isolate the fault state of the system at less cost with faster speed. In this regard, the domestic and foreign scholars have conducted a lot of research and have obtained many research results on binary test, but still lack in-depth study on multi-valued test. Aiming at the deficiency of the classical algorithm on binary test, this paper has made improvements, and carried out research on multi-value test.Firstly, this paper introduces the modeling process of multi-signal flow graph. It establishes multi-signal flow graph model according to the schematic diagram of electronic equipment, and then gets fault-test dependency matrix which the diagnostic strategy optimization design requires. Secondly, it introduces the basic theory of the diagnostic strategy optimization technology, and does research on the diagnostic strategy optimization design problem of binary test. It analyses the advantages and disadvantages of the two classic diagnostic algorithms of strategy optimization(greedy algorithm and AO* algorithm) on binary test, presents limited search AO* algorithm. Results of simulation examples show that the improved algorithm can reach a trade-off between the optimality of the diagnosis tree and the efficiency of algorithm. Thirdly, it studies in the diagnostic strategy optimization design problem on multi-value test on the basis of the algorithm on binary test, and AND-OR TREE heuristics search algorithm based on Hoffman code is put forward. The example simulation validates the algorithm is feasible in solving the diagnostic strategy optimization problems on multi-value testing. Finally, according to the requirements of the input and output of diagnostic strategy optimization algorithm, this paper designs a graphical interface which can be used for small-scale system and make the operation more simple and output results more intuitive.