A Support Vector Machine-Based Intelligent System for Real-Time Structural Health Monitoring of Port Tower Cranes

Ensuring the structural integrity of port tower cranes is paramount for safety and maximizing their operational lifespan. Conventional monitoring methods often lack real-time capabilities and struggle with continuous damage assessment. This research addresses this gap by proposing a novel, real-time structural health monitoring (SHM) system for port tower cranes that utilizes a support vector machine (SVM)-based intelligent system. The proposed system leverages sensor data to specifically detect and localize cracks, a critical damage type in these cranes. It employs experimental modal analysis, a technique that extracts the inherent vibrational characteristics of a structure. By measuring the first three natural frequencies, the system gathers valuable information about the crane health. These natural frequencies are then fed into the SVM algorithm, which is trained to recognize patterns that correspond to the presence and location of cracks. To validate the effectiveness of this intelligent system, a scaled model of a real port tower crane is constructed. This model undergoes numerical modal analysis, simulating various crack configurations. The resulting natural frequencies are used to train the SVM algorithm. Subsequently, the system crack detection and localization capabilities are tested through experimental measurements on the fabricated model. The results demonstrate a high degree of accuracy between the SVM-predicted crack locations and the actual measurements. This success confirms the potential of the proposed SVM-based system for real-time SHM applications in port tower cranes, paving the way for a more proactive approach to ensuring their structural safety.