Probabilistic and Reliability-Based Health Monitoring Strategies for High-Speed Naval Vessels

摘要

In this project, a strategy is proposed for fatigue life estimation of a ship hull using a wireless sensor network installed in the hull for autonomous health monitoring. First, a rainflow counting algorithm is implemented as a continuous time-domain approach to fatigue estimation. Cycles are defined by a rainflow counting procedure and are kept on-board the wireless sensing unit in a histogram structure for long-term storage. Fatigue damage is accumulated by the Palmgren-Miner linear summation method. Second, a statistical approach called the Dirlik procedure is used. It relies on spectral moments of the stress-time history power spectral density (PSD) function. The Dirlik procedure outputs a probability density function (PDF) of stress ranges tailored to resemble the rainflow counting results. The PDF is converted to damage using an S-N curve and again accumulated by the Palmgren-Miner method. Experimental tests are conducted on an aluminum hull stiffened element specimen as part of the Monitored Aluminum Hull Integrity (MAHI) test program to verify the embedded fatigue life estimation procedures.