Study On Optimal Placement Of Triaxial Sensors Based On The Wolf Algorithm

Optimal sensor placement (OSP) is one of the key issues of structural health monitoring system; the effectiveness of the data which collected have great importance on later steps. So the principle of OSP is to make sure the effectiveness of the data. The number of the sensors can not be a lot, considering the limits of the reality:the complex of structures, the cost of sensor placement, and the feasibility of collecting and storing of the data. In other words, we use as less as possible of the sensors to obtain as much as possible of the information of the structure. Because of the disadvantages of the current criterions, which fail in a particular way, the article focus on the following aspects:(1) Considering that the Modal Assurance Criterion (MAC) in the optimal sensor placement (OSP) can only be used for the optimization of one single direction of the triaxial accelerometers, not for all the three directions, a new Triaxial-MAC (i.e. TMAC) is established based on the thought of MAC. This is done by taking the node’s three translational degrees-of-freedom as one unit and obtaining the fisher information matrix of the structure by those of the sensors.(2) In order to solve the information redundancy problem in the optimal sensor placement (OSP), the information redundancy function g(R) is proposed, and then the triaxial modal assurance criteria considering the redundancy of information is established by combining g(R) with the triaxial modal assurance criterion (TMAC), which can ensure both the observability and discriminability of the modal shape.(3) To improve the solving efficiency, a novel distributed wolf algorithm (DWA) is proposed. First, an improved dual-structure coding method is used to guarantee that the original WA can solve the optimization of discrete variables. Then, the wolves are grouped to enhance the search efficiency by the information exchange of the individual wolves. Finally, the sensitivity analysis of the main parametric and computation with the suitable parametric values are done on the benchmark structure developed by the University of Central Florida. The results show that comparing with SWA, DWA greatly increases the search efficiency, which can better solve the OSP problem.(4) A novel hierarchic wolf algorithm (HWA) is put forward to increase the variety of the wolf pack, because of the siege of process. First, an improved dual-structure coding method is used to guarantee that the original WA can solve the optimization of discrete variables, and the artificial uniform distribution method is raised for the initialization of the wolf population to ensure the uniformity of the initial data. Then, the hierarchic method is adopted to avoid that any individual wolf has the similar grade with the wolf king, which may enhance the diversity of the wolf population and improve the searching efficiency of the algorithm. Finally, the sensitivity analysis of the main parametric and computation with the suitable parametric values are performed on the benchmark structure developed by the University of Central Florida. The results show that comparing with SWA, HWA greatly increases the search efficiency, which can better solve the OSP problem.