Zhiqian Hu, Jie Hou, Yunsheng Su, Yuqing Wang, Wei Dai, Jie Yang
As a highly integrated and increasingly complex high-risk process industry, the petrochemical sector plays a critical role in industrial continuity and social stability, yet faces significant governance adaptability challenges under normalized public health emergencies. Taking a Chinese petrochemical enterprise as a case study, this paper develops an integrated framework combining STAMP/STPA, complex network analysis, and robustness analysis. Based on a reconstructed four-level hierarchical control and feedback structure, STPA was applied to identify 20 unsafe control actions (UCAs). These UCAs and their precursor factors were further abstracted into a relational network of control deficiencies for topological analysis and Monte Carlo-based robustness testing under random failure and targeted attack. The results show pronounced small-world and core–periphery structural characteristics, with vulnerability concentrated in a limited number of high-centrality source and hub nodes. Systemic resilience constraints mainly arise from governmental deficiencies in response experience and training, enterprise-level amplification at hub nodes, and pressure accumulation at frontline execution nodes. Accordingly, three resilience protocols are proposed: distributed authorization for source nodes; digitized dual-channel feedback for hub nodes; and minimum operational redundancy with cross-replacement for terminal nodes. This study provides theoretical basis and strategies for high-risk industrial systems to enhance resilience and sustainable development in uncertain environments.