Optimization of Embedded Sensor Placement for Structural Health Monitoring of Composite Laminates

This dissertation develops an optimization method specifically for embedded sensors for structural health monitoring of a composite laminated aircraft structure. The chosen cost function is the component lifetime, balancing both the positive benefits of the condition based monitoring enabled by the sensor information with the negative costs of the structural performance degradation. As a result, for a given application the sensor placement optimization may not yield a solution with a positive benefit, unlike the case for surface mounted sensor networks. The optical fiber spacing is optimized, rather than the placement of individual sensors. When sensors are embedded into the laminated material system, their interaction with damage is highly influenced by the relative location of the optical fibers to local microstructural features. Therefore the material lifetime degradation is not deterministic and cannot be well modeled by a predictive computational model. These effects are therefore included into the optimization problem through an experimentally derived, probabilistic model. Due to the same effect, the response of the embedded sensor is also not deterministic and depends highly on its relative location to the material microstructure. For this same reason, the sensor feature extraction method is also experimentally driven under similar conditions. The optimization method is demonstrated for the example of a composite laminate subjected to multiple, low-velocity impact events.