In the port domain, quayside container cranes are an indispensable component of maritime freight transport. These cranes are not only costly but also associated with safety accidents that often result in casualties and property loss, severely impacting port operations and the surrounding environment. Given their complex operational environment, rapid technological updates, high dependency on human factors, and the challenges of maintenance and inspection, the safety of quayside container cranes is a significant concern for port enterprises and managers. This paper, based on the operational modes and structural characteristics of the cranes, divides them into five main systems and identifies twenty-eight safety evaluation indicators, covering a comprehensive range of risk factors from equipment integrity to operator behavior, as well as environmental factors. However, numerous pain points exist in the safety risk evaluation process of quayside container cranes, such as fuzziness, uncertainty, and complex multi-criteria decision-making (MCDM) environments. These issues make traditional safety evaluation methods inadequate in accurately reflecting the actual safety conditions. Therefore, this paper proposes a safety evaluation method for quayside container cranes based on the Best–Worst Method (BWM) and Pythagorean hesitant fuzzy VIKOR. This method effectively overcomes the uncertainties and fuzziness of traditional safety evaluation methods by integrating the decision maker’s preference information from the BWM and the fuzzy handling capability of Pythagorean hesitant fuzzy sets, enhancing the accuracy and reliability of the evaluation results. A case study was conducted on a quayside container crane at a specific port. Through empirical analysis, the feasibility of the proposed method was validated. Overall, the safety evaluation method for quayside container cranes based on the BWM and Pythagorean hesitant fuzzy VIKOR proposed in this paper enriches the theoretical research on the safety risk assessment of quayside container cranes and offers a new approach and tool for port enterprises and managers in practice.

Conventional data-driven networks for gearbox fault diagnosis process multi-sensor streams as isolated sequences, failing to capture spatial-topological kinetic correlations and structural energy propagation pathways governed by multi-stage gearbox dynamics. To address these limitations, this study proposes a graph neural network-based fault diagnosis methodology integrating multi-dimensional attention and dynamic topological priors (KT-GNN-CBAM). Mesh stiffness characteristics are analytically evaluated to initialize physical topology edge weights, while a convolutional block attention module filters spatio-temporal features to suppress background noise. Node features are subsequently aggregated through a parallel dual-stream architecture comprising a physics-prior kinetic stream and a data-driven attention stream, which are dynamically fused via an adaptive gated mechanism. Experimental validation on the HP-GBS-2023 testbed under mixed operations and 6 dB noise shows that the proposed framework achieves an optimal diagnostic accuracy of 98.87%. Ablation evaluations confirm that omitting the mechanics-driven prior branch induces a 282.30% relative surge in the model’s misclassification rate. Ultimately, embedding mechanical invariants as a physical inductive bias mitigates purely data-driven black-box constraints, offering an interpretable and robust solution for advanced intelligent fault diagnosis in complex gearbox systems.