Signal Processing Issues Related to Structural Health Monitoring

Structural health monitoring (SHM) refers to automated methods for determining adverse changes in the integrity of mechanical systems. Key components of the SHM process include data acquisition and normalization, feature extraction and information condensation, and statistical model development. Data acquisition includes optimizing the number and placement of sensors on the structure and ensuring that the sensors are robust enough to enable accurate measurements over the life of the structure. Normalization is necessary to account for undesired effects such as changes in environmental conditions and sensor-to-sensor variability. Once data has been collected, features - portions of the data with the potential to discriminate between different structural damage states - must be extracted. A subset of the most useful features extracted from the measured data is selected for use in the statistical models. Finally, statistical models are developed to enable automatic identification of the structural damage state. In this paper, issues relating to the key SHM components will be discussed and specific examples of efforts to address these issues will be presented.