Research On Optimization Of Sensor Placement For Heterogeneous Wireless Sensor Networks

With the development of structural health monitoring at home and abroad, Comparing with the homogeneous wireless sensor networks, sensor nodes in heterogeneous wireless sensor networks(HWSN) have different allocation of resources, and can better meet the demands of various applications for the technologies of sensor networks. Heterogeneous wireless sensor networks is formed by a low cost, with perception, data processing, storage and wireless communication capabilities of multi-type micro-sensor nodes through self-organizing. Meanwhile, a heterogeneous wireless sensor networks is used to carry out structural health monitoring, with a lot of limitations such as energy consumption, transmission distance. Multi-type wireless senors in heterogeneous wireless sensor networks should be deployed with reason when applied to practical engineering. In this paper, wireless sensor nodes identify structural modes through structural acceleration and strain monitoring, and then this paper proposes a optimal heterogeneous wireless sensor deployment method by establishing optimal goals about improving structural modal identification and reducing energy consumption. clustering issue is studied further in a certain heterogeneous wireless sensor network to reduce the network energy consumption. At last, optimal deployment for heterogeneous wireless sensor network simulates structural monitoring. The main work and results in this paper are as following.Structure modes obtained by heterogeneous wireless sensor networks does not have modal orthogonality, so this paper proposes a new structural modal evaluation criteria-modal identification index. This paper will measure modal identification index by the ratio of modal clarity index(MCI) and modal relative error(MRE) to evaluate structural modal identification. By means of laboratory study truss model, this paper studies uniform schemes and non-uniform schemes. Further study shows that the more two types sensors deployed, the wider heterogeneous wireless sensor network coverage, the higher the structural modal identification. Finally it is concluded that the number and relative position have an effect on the modal identification.After analyzing energy consumption in the process of data transmission in heterogeneous wireless sensor network, this paper considers data acquisition, data processing, data sending, transmission and data reception as the main part of the network energy consumption. This paper proposes a sphere energy consumption model to analyze the network energy consumption. Finally it is concluded that the number and relative position have an effect on the network energy consumption.A two-objective optimization function is constructed based on the modal identification and energy consumption. The two-objective optimization function is the number of wireless sensor nodes and coverage radius. And the optimal value of the variables can be determined by solving the function with discrete particle swarm optimization(DPSO) or other methods. According to the optimal number and radius of the two type sensors, the heterogeneous wireless sensor networks based SHM can identify the modes with high accuracy and save more energy, which achieve the optimization goals mentioned above. The optimal scheme of the uniform scheme is obtained by solving the objective function. By comparing between the two optimal schemes, discrete particle swarm optimization algorithm is verified with reliability.For a certain heterogeneous wireless sensor networks, clustering issue is studied further to reduce the energy consumption. A cluster based energy consumption function is obtained by analyzing energy consumption within the cluster. And a typical extremum method is used to determine the optimal number of clusters.Theories that all mentioned above are simulated to monitor the main bridge of Ping Hu marshalling station bridge. And optimal scheme of heterogeneous wireless sensor networks is obtained. The implementation of the heterogeneous wireless sensor network monitoring system has some engineering significance.