Yingkou Institute of Technology

To investigate the effect of cooling rate on the phase transformation behavior and mechanical properties of 9Ni steel, a 7 mm thick industrial 9Ni steel plate was selected as the research object in this study. The JMatPro software was employed to simulate and calculate key parameters, including the thermodynamic phase diagram, CCT curve, and mechanical properties. Meanwhile, static thermal simulation experiments at cooling rates ranging from 0.5 to 30 °C/s were conducted on a Gleeble-3500 thermal simulation testing machine. Microstructure characterization and property tests were carried out using a metallographic microscope, scanning electron microscope (SEM), and Vickers hardness tester, and the experimental CCT curve was subsequently plotted and compared with the simulation results. The results revealed that the microstructure of 9Ni steel changed regularly with the cooling rate. With the increase in cooling rate, the ferrite content decreased continuously, the bainite content increased initially and then decreased, and the martensite content increased continuously. At a cooling rate of 30 °C/s, the martensite content reached approximately 90%. The microhardness of 9Ni steel initially sharply increased and then stabilized with the increase in cooling rate, stabilizing at 359 HV1 at a cooling rate of 30 °C/s. The phase transformation law of the measured CCT curve is highly consistent with the simulation results, verifying the reliability and accuracy of JMatPro for predicting the phase transformation behavior and mechanical properties of 9Ni steel. This study provides a theoretical basis and data support for the precise optimization of the heat treatment process of 9Ni steel and has important practical significance for enhancing its service performance in cryogenic engineering applications.

Background Paediatric cancer survivors often experience cognitive late effects, including impairments in working memory, processing speed, and executive function, which can adversely affect academic performance, psychosocial development, and long-term functional outcomes. Traditional educational interventions may be limited by stigma, accessibility, and inconsistent engagement. Artificial intelligence (AI)-driven virtual tutors represent a promising, personalised approach to cognitive rehabilitation in this population. Methods A mixed-methods RCT with 150 survivors compared AI Tutor, Traditional Tutoring, and Control groups. Assessments included neurocognitive tests (WISC-V, Flanker, N-Back), academic metrics, psychosocial scales, and participant/parent/educator interviews. Subgroup analyses and ethical considerations (bias, equity) were explored. The study aimed to support and measure cognitive rehabilitation outcomes in this population. Results Participants in the AI Tutor group demonstrated significantly greater cognitive rehabilitation gains, particularly in working memory, processing speed, and maths performance compared to both the Traditional Tutoring and Control groups (p < 0.001). Flanker task error reduction and N-Back task performance indicated enhanced executive function, particularly among adolescents. Engagement metrics (e.g., session completion rates) were higher in the AI group, with younger children showing increased consistency and participation. Qualitative findings highlighted the AI tutor’s non-judgmental nature, reduced perceived stigma, and effective use of gamification to foster motivation and routine. Conclusion AI-driven virtual tutors provide a scalable, personalised, and stigma-reducing approach for cognitive rehabilitation in paediatric cancer survivors, effectively complementing traditional methods across settings. Addressing ethical concerns like algorithmic fairness and access is crucial for equitable implementation. This study supports integrating AI tutors into survivorship care.

ABSTRACT In order to reflect the actual production situation more comprehensively and optimize the production cost, this paper solves the short‐term scheduling optimization problem for a single pipeline containing high melting point crude oil. Based on the refining plan given by the upper layer, a multi‐objective optimization model with high melting point crude oil constraints is established to optimize five common cost objectives in the scheduling process. For the model characteristics, the NSGA‐III algorithm is used to solve the problem. The reference points of the algorithm can be flexibly set according to the number of objectives, which allows for more efficient coverage of the high‐dimensional objective space. The results show that the NSGA‐III algorithm performs better in the solution performance, and the hypervolume (HV) indicator is improved by more than 10% compared with other algorithms. Comparison of the solution sets obtained using the NSGA‐III algorithm with the existing literature shows that the optimization results on different objectives are improved by 4.6%–13.3%, and the comprehensive cost index is significantly reduced, which proves the effectiveness of the method. This study provides a more cost‐effective solution to the practical scheduling problem in refineries by incorporating high melting point crude oil constraints and improving the efficiency of the algorithm.

Fault diagnosis of drainage pumps is essential for ensuring the safe and stable operation of drainage systems. However, vibration signals acquired from drainage pumps under practical operating conditions are often characterized by strong non-stationarity, unavoidable environmental interference, and fault features distributed across multiple temporal scales, which poses significant challenges to accurate multi-class fault identification. To tackle these challenges, this paper proposes a multi-scale residual attention convolutional neural network (MRANet) for drainage pump fault diagnosis. A multi-scale residual attention feature extraction (MRAFE) module is designed to jointly model fault-related information at different temporal scales by means of parallel convolutions and residual connections, while attention mechanisms are incorporated within each scale branch to enhance discriminative fault features. Furthermore, a complementary multi-level feature integration (CMFI) strategy is developed to effectively integrate features extracted at different network depths, enabling end-to-end fault identification directly from raw one-dimensional vibration signals. Extensive experiments are conducted on a drainage pump vibration dataset containing eight operating conditions. Experimental results show that MRANet achieves an accuracy of 95.43%, outperforming several benchmark models. These results verify the effectiveness and robustness of the proposed method for multi-class fault diagnosis of drainage pumps.

With the development of the digital economy, livestream marketing for agricultural products has become an important channel for promoting the digital transformation of rural industries. Drawing on user value theory, this study examines how social presence influences customer engagement in agricultural livestream marketing by incorporating customer satisfaction as a mediating variable and trust atmosphere as a moderating variable, and proposes a conceptual model. Based on data from 312 valid questionnaires, the results show that social presence positively affects customer engagement both directly and indirectly through customer satisfaction. Moreover, trust atmosphere strengthens the direct effect of social presence on customer engagement as well as the mediating role of customer satisfaction. This study extends consumer behavior research by integrating livestream marketing into the theoretical framework and highlighting the role of livestream-specific relational factors in shaping customer behavior. The findings also provide actionable insights for optimizing agricultural livestream marketing, including fostering trust, enhancing interactivity, and supporting farmers and e-commerce platforms.

As the energy transition accelerates toward low-carbon development, smart microgrids, serving as the core infrastructure for efficient distributed energy integration, face dual challenges of renewable energy volatility and dynamic load complexity. Energy storage systems, through their power balance and energy time-shifting capabilities, form the foundation for stable system operation, with their performance relying on precise monitoring by battery management systems (BMS) and supported by information-based data communication technologies. This paper focuses on BMS security management, multi-objective dynamic scheduling, and cross-protocol communication technologies, aiming to address the coordination challenges of economic viability, reliability, and environmental sustainability under high renewable energy integration, thereby providing theoretical support and practical pathways for next-generation power systems.