Abstract
<title>Abstract</title> <p>Tower-crane construction safety depends on equipment reliability, foundation and attachment stability, installation/dismantling/jacking control, personnel coordination, maintenance rectification and the operating environment. Conventional checklists and single monitoring indicators cannot fully describe whether a tower-crane safety system can absorb disturbances, recover safe operation and adapt to changing site conditions. This study develops a safety-resilience assessment framework that integrates a cloud model with an intuitionistic fuzzy Bayesian network. Twelve influencing factors are identified from resilience theory, tower-crane standards and previous studies, and are organized into absorptive, recovery and adaptive capacities. Interpretive structural modeling is used to derive a four-level factor hierarchy and to support Bayesian-network topology construction. Expert linguistic assessments are represented by intuitionistic fuzzy numbers and transformed into local support probabilities through the cloud model. The framework is applied to a QTZ125 tower crane in the Hongling Jiayuan resettlement housing project. The comprehensive resilience score is 0.785, and the high-resilience probability is 0.706. Sensitivity analysis identifies standard updating and continuous improvement, organizational management and responsibility closure, installation/dismantling/jacking control, foundation and attachment stability, and inspection-maintenance-hazard rectification as the main improvement factors. The proposed framework provides a traceable decision-support tool for tower-crane safety-resilience assessment under uncertain expert information.</p>