This study addresses the critical safety concern of pit defects in crane booms, which occur under complex working conditions and prolonged operation, potentially leading to loss of stability and catastrophic collapse. Our research focuses on a 1200-ton all-terrain crane boom, systematically investigating how geometric pit defects affect structural integrity. We precisely quantify the extent of load-bearing capacity degradation at various defect positions and identify the locations most susceptible to the maximum weakening effect. To enable proactive maintenance, this paper proposes an advanced fault diagnosis method to address the lack of fault diagnosis approaches for crane boom pit defects based on a hybrid Particle Swarm Optimization-Backpropagation (PSO-BP) Neural Network. The model utilizes simulated data, where the PSO algorithm first efficiently searches for and optimizes the network’s thresholds and weights. These optimized parameters are then assigned to the BP neural network, creating a robust diagnostic system capable of high-precision fault identification. We conducted comprehensive experiments to investigate both single-pit and multi-pit faults at different locations on the boom. The results demonstrate the model’s exceptional performance, achieving a maximum recognition accuracy of 100% in single-pit fault scenarios, with a comprehensive accuracy of 99.02% via 5-fold cross-validation. This high level of accuracy provides a reliable and novel methodology for the practical, engineering-based fault diagnosis of cranes, enhancing operational safety and enabling timely interventions, providing a data-driven tool for predictive maintenance and reducing structural failure risks in heavy-lift operations.
Zhao et al. (Sun,) studied this question.