Optimal sensor placement and model parameter identification for distributed- parameter structures

For the discrete-coordinate systems, the sensors are confined to be placed at only a very limited number of discrete locations. In this paper, for the purpose of model parameter identification, a sensor/actuator configuration methodology is developed for the distributed-parameter structures, where the sensor locations are identified by their coordinates that are continuous within the region of interest, having both the theoretical and practical significance. For the structural model parameters, the information entropy is employed as a scalar measure to quantify the uncertainties, and the Bayesian statistical approach is adopted to identify the optimal values and associated uncertainties. The problem of optimal sensor/actuator placement is then formulated as a continuous optimization problem by minimizing the information entropy using the binary-coded genetic algorithm. By investigating the optimal sensor/actuator placement as well as model parameter identification, the validity of the proposed methodology is verified through a set of numerical cases.