Advanced Structural Damage Detection Methods for Normal and Extreme Loads

点击翻译按钮获取中文摘要

AbstractThe objectives of this project are to develop damage diagnosis algorithms based on distributed decision-making whereby signals collected locally at each sensing unit within a dense wireless sensor network are processed, the information is fused at various network levels and only assessments are transmitted to network endpoints. The approach includes the development of data preprocessing and structural damage detection algorithm development. The damage detection algorithms will utilize autoregressive and wavelet based analysis for signal modeling. Damage detection will be achieved by first defining features within signals and then tracking changes in the features using statistical pattern classification methods. Localization and quantification of damage will be addressed by using sensor fusion and feature change statistics correlation to physical damage levels. Algorithm calibration and verification will be performed with data obtained from recent laboratory and field tests. These will include measurements from vibration, strain, temperature, humidity and crack propagation experiments. Sample algorithms will be embedded on microprocessors of existing wireless multi-sensor structural sensing units and utilized in ongoing tests. The results of this research will greatly increase the capabilities of wireless structural monitoring by providing diagnosis at the local component, subassembly and overall structural levels. Furthermore, the algorithms will significantly reduce the need for direct data transmission and thus reduce data communications rates. The lower data rates and local embedded data processing will also greatly decrease power consumption increasing potential field deployment of wireless sensor networks. Graduate and undergraduate students will benefit from the research conducted under this research. The inclusion of one Hispanic graduate student who has a diversity fellowship from Stanford University and a female student to be supported by this project will increase the doctoral pool from these underrepresented groups. A new MS/PhD level course will be initiated from this research. The research team will involve students from local high schools currently under the Stanford Chartered School system. Results from the research will be disseminated through publications in journals, conferences and symposia.