Test point optimization of control moment gyro (CMG) based on multi-signal flow model and improved genetic algorithm

The increasing integration degree of circuits in electronic systems makes faults detection and isolation more and more difficult. Design for testability (DFT) was applied to solve the problem. In this procedure, designers are required to set an optimum arrangement of sensors to achieve the best testability. But the search of the best test points combination is usually conducted by manual work. To simplify this blind and inefficient procedure, a test point optimization algorithm based on multi-signal flow model is proposed in this paper and is successfully applied in a CMG. First, the basic conception of modeling is introduced and the analysis of testability is converted to the calculation based on dependency matrix. Then the test points optimization problem is described and analyzed by an improved genetic algorithm. At last, the algorithm is examined by the CMG´s A/D board subsystem. The result shows the algorithm can find the best solution with excellent efficiency.