Research On Sensor Selection And Optimization Technology In BIT Design For Mechatronics Equipments

Since information acquirement for BIT equipments in mechatronics systems depends on sensors intensively, the proper sensor selection and configuration are very important, which is aggravated by related factors, such as diversities and complexities of the host mechatronic systems, and operation environment. A bad sensors scheme results in high testing costs and poor testability and diagnosability. To solve such problems, sensor selection and optimization are studied in this paper deeply. The main contents are described in brief as follows.1. Since FMEA is the base of test selection, the FMEA sheets of three types of mechatronics equipments, which are mechanical transmissions, slewing structure and electric-control & drive system, are obtained after introduction of the function, method and program of FMEA. Their fault modes, reasons, potential effects and detecting methods are identified respectively. Furthermore, faults are classified based on their severity.2. Aiming at the problem of sensor selection, the selection guidelines and steps for test-points are summarized firstly. Then the guidelines for sensors selection are also investigated according to test requirements of host systems. The selecting methods of three kinds of sensors for BIT in mechatronics equipments, which are vibration sensors, temperature sensors and speed sensors, are outlined in the end.3. A digraph model describing the relationship between sensor and faults, which can be converted into bipartite graph, is proposed for sensors optimization. Meanwhile, an infinity Boolean algebra system is introduced to describe the characteristics of sensor selection and optimization. Based on the fault information matrix obtained from the bipartite graph model and the Boolean algebra system, a MINLP model orienting sensors optimization is proposed. On the assumption of the single-fault resolution, several testing parameters such as FDR, FIR is considered, and optimal multiple objectives, such as minimum total sensor costs, minimum sensor failure probabilities and so on, are achieved. The case study highlights the validity of the MINLP model.4. Based on the methods proposed in this paper, the scheme of optimal sensor configuration of the BIT system of a mechatronics platform is achieved.