Process parameters critically affect the flow, heat transfer, and mixing behaviors of rubber melts during single-screw extrusion. However, direct observation of the internal flow field is difficult, and high-fidelity numerical simulations are computationally expensive, limiting rapid analysis under multiple operating conditions. To address these challenges, a three-dimensional non-isothermal CFD model validated against experimental outlet temperature data was developed and coupled with an IPSO-BP surrogate model to establish a rapid prediction framework. The effects of screw speed, inlet temperature, and barrel temperature on flow, heat transfer, and local mixing characteristics were systematically investigated. The results indicate that screw speed predominantly affects melt flow and mixing behavior by enhancing drag flow, shear action, and internal heat transfer. Inlet temperature exerts the greatest influence on the mean outlet temperature, whereas barrel temperature primarily regulates the thermal state of the melt through wall heat transfer. Compared with the BP, GA-BP, PSO-BP, GPR, and XGBoost surrogate models, the IPSO-BP model demonstrated superior overall predictive performance, particularly achieving higher prediction accuracy and stability in predicting complex temperature responses. The proposed CFD–IPSO-BP framework provides an efficient tool for rapid prediction and optimization of EPDM single-screw extrusion processes.

Cross-condition fault diagnosis is important for ensuring the reliable operation of mechanical systems. However, most existing methods assume that the source and target domains share identical fault label spaces and data distributions, limiting adaptation to unknown faults and cross-condition shifts in practical industrial scenarios. To address this issue, a structured, reliable, domain-specific open-set domain adaptation method is proposed. The proposed method first constructs a domain-specific batch normalization-based feature extraction network, in which independent normalization branches model statistical discrepancies under different operating conditions; it then designs a cross-domain structured representation consolidation module to enhance feature discriminability through source-domain anchor compactness, target-domain multi-view contrastive, and prototype entropy regularization constraints; an open-set boundary learning mechanism is further introduced to establish a discriminative boundary between known and unknown classes; finally, a reliability-aware pseudo-label propagation strategy refines target-domain pseudo-labels and imposes separate prediction-consistency constraints on known and unknown classes. Experimental results on the CWRU bearing dataset and the self-built rolling bearing dataset show that the proposed method achieves average H-scores of 93.32% and 97.65%, respectively, on open-set transfer tasks. Compared with several baseline methods, the proposed method achieves a better balance between known-class recognition and unknown-class detection, thereby improving cross-condition open-set fault diagnosis performance.