Huzhou Vocational and Technical College

Accurate State-of-Health (SOH) estimation for second-life batteries is a critical challenge due to unavailable first-life historical data. This work introduces the Prior-Informed Hierarchical Fusion Network (PI-HFN), a flexible deep learning framework to address this information asymmetry. The model leverages a historical degradation vector from the battery’s first life as a prior condition to guide the interpretation of its current electrochemical state. This prior-conditioning is achieved by initializing a recurrent neural network, which processes instantaneous second-life data, with the historical information. The framework’s efficacy was validated across two comprehensive case studies. In Case Study 1 on NMC cells using multimodal EIS and DCIR inputs, the PI-HFN achieved a state-of-the-art average RMSE of 1.91%. Case Study 2 on a large-scale LFP dataset, using only DCIR inputs, further demonstrated the framework’s robustness and flexibility. Both studies confirmed through systematic ablation studies that the historical prior (BFN) and the hierarchical fusion strategy (PFN) are crucial for high accuracy, with the historical prior proving to be the dominant factor in mitigating errors in data-sparse SOH regions. The model also shows excellent deployment potential, with an inference latency of 15.8 ms on an RK3568 edge platform after INT8 quantization. This study validates that conditioning deep learning models with historical priors offers a robust, flexible, and practical pathway for evaluating second-life batteries, supporting their integration into the circular economy.

Abstract Background The rapid advancement and widespread adoption of Industrial Robotics Technology (IRT) in manufacturing have significantly enhanced the automation and intelligence of production processes. While technological innovation boosts efficiency, it simultaneously raises concerns among factory workers regarding workplace safety, risks of skill obsolescence, and job role uncertainty, potentially increasing Workplace Stress (WS). Current research predominantly focuses on the organizational performance benefits of IRT, yet systematic evidence remains scarce regarding employees' psychological stress responses to technological innovation and its influencing factors. Methods To construct an evaluation framework for WS in the context of IRT and analyze the changing trends and related impact mechanisms of employee stress levels. The study is based on two manufacturing companies that are implementing IRT upgrades, and a total of 642 frontline employees were recruited to participate in the survey. The main measurement tools include Job Uncertainty Perception Scale (JUPS), Skill Replacement Threat Questionnaire (SRTQ), Workload Index (WI), and Workplace Stress Assessment Scale (WSAS). The investigation will be conducted in two stages, corresponding to 6 months before and after the implementation of IRT in the enterprise. Using a mixed effects model to analyze changes in stress levels and applying a structural equation model to analyze stress generation pathways. Results After six months of implementing IRT, employees' WSAS scores showed a 27% increase from baseline (p= .003), with job uncertainty perception rising by 31% (p=.004) and perceived threat of skill replacement increasing by 34% (p=.006). Although the workload index remained stable, operational complexity-related items improved by 18% (p=.03), indicating that heightened skill adaptation demands became a significant stress source. Structural equation modeling revealed that job uncertainty perception had the strongest predictive effect on WS, while perceived threat of skill replacement contributed 0.41 (p=.01). Together, these factors explained 62% of workplace stress variance. Notably, companies providing training support achieved a 16% increase in employee WS, significantly lower than the 32% observed in non-training groups (p=.02), demonstrating a clear organizational buffering effect. Further analysis showed that highly engaged employees improved their skill mastery scores by 28%, with their WSAS scores showing a significant decline during follow-up (p=.03), indicating that skill confidence serves as a crucial moderating factor in stress recovery. Discussion The study reveals that IRT technological innovation has significantly increased workplace stress levels among factory workers, with job uncertainty and skill substitution threats serving as primary psychological mechanisms. Although workloads remained stable, operational complexity caused by evolving skill requirements further exacerbated stress accumulation. Training support and employee engagement in skill enhancement proved crucial for stress relief, highlighting the need for organizations to address staff psychological adaptation during technological transformation. This research provides empirical evidence for understanding stress dynamics in IRT environments and offers practical guidance for developing employee support systems during technological upgrades. Future studies should conduct long-term tracking across different manufacturing sectors to verify the sustained impact of industrial robot innovation on workplace stress and the long-term buffering effects of training support mechanisms.