The fundamental operational nature of modern internal combustion engine (ICE) systems is intrinsically defined by an extremely strong degree of nonlinearity, as well as the intricate and multifaceted complex characteristics associated with high-dimensional interactions and multi-parameter coupling effects. The inherent physical complexities of the subject make it exceptionally difficult for traditional experimental teaching methodologies to successfully achieve high-precision combustion prediction or to realize the effective improvement of overall system energy efficiency in a consistent manner. One reason for this is that such methodologies often lack the necessary capacity for handling such massive amounts of multivariate data effectively. Within the specific context of contemporary engineering education, it has become imperative and critical to develop practical, technically effective, and systematic teaching frameworks. The overarching objective of these frameworks is to systematically foster and cultivate students’ essential capabilities in the fields of computational modeling and strategic decision-making. This enables them to successfully navigate the multifaceted challenges involved in dealing with the modeling and optimization of the complex engineering problems encountered in real-world scenarios.

This teaching reform is a direct and targeted response to identified pedagogical and technical challenges. It proposes an innovative, comprehensive, and integrated method specifically designed for the purposes of ICE combustion prediction and smart energy management. This method integrates advanced machine learning technologies into the existing curriculum in a seamless and deep way. The methodological framework is meticulously implemented through a rigorous, structured, and sequential process. First, a systematic sampling design strategy is meticulously executed for the specific test content to ensure comprehensive data coverage of the operating space. The extensive datasets obtained from these comprehensive experimental tests are then utilized as the foundational material to train sophisticated machine learning models. Following this preliminary data acquisition phase, a robust and high-fidelity ensemble tree surrogate model is computationally constructed to achieve high-precision fitting and accurate prediction of critical ICE performance indicators. This surrogate model serves as a reliable digital proxy for the physical engine system. Finally, advanced multi-objective optimization algorithms are integrated into the workflow to process these predictive models, with the aim of generating a Pareto optimal solution set that represents the ideal mathematical balance between conflicting operational goals.

The results of the quantitative optimization derived from the implementation of the present study provide compelling and robust empirical evidence that demonstrates the practical effectiveness of the framework. The experimental data unequivocally demonstrate that, under the stipulated trade-off operating condition, ICE’s fuel consumption is significantly reduced by a substantial margin of 14.5%. Concurrently, the environmental performance has been notably enhanced, with exhaust emissions of hydrocarbons (HC), carbon monoxide (CO), and nitrogen oxides (NOx) reduced by 7.0%, 5.7%, and 32.8%, respectively. This outcome validates the efficacy of the proposed optimization strategy.

In summary, the test teaching reform framework was successful in establishing a complete, comprehensive, and logical closed loop encompassing the iterative stages of “experimental design–data modeling–optimization feedback.” This structured pedagogical approach enables students to have a solid foundation in the scientific research process, extending seamlessly from the initial phases of test design to the final and critical stages of data analysis and parameter optimization. Moreover, it considerably augments their comprehensive capacity to solve complex scientific problems by efficaciously bridging the divide between theoretical academic knowledge and practical engineering application. This teaching reform offers a valuable, replicable, and promotable paradigm for the future teaching reform practice in energy and power-related majors.

With the rapid development of higher education in China, experiments are becoming increasingly diverse and new methods are constantly emerging. Experimental research is showing a trend toward interdisciplinary collaboration, and traditional safety management methods are proving inadequate for effective prevention of accidents, such as fires, explosions, and poisoning. In recent years, a series of laboratory accidents in universities have exposed shortcomings in the traditional management systems to identify and warn against dynamic risks. Consequently, integrating cutting-edge technologies, such as artificial intelligence, to establish an intelligent, proactive laboratory safety risk prevention and control system has become an urgent necessity for enhancing the effectiveness of laboratory safety management in higher education institutions.

To enhance the effectiveness of laboratory safety management and prevent various types of accidents, this study systematically reviews the current state of the application of the five-in-one technology—artificial intelligence, neural networks, image recognition, geographic information systems (GIS), and visualization—in laboratory safety, and proposes the construction of an integrated technological framework comprising “AI-neural network-image recognition-GIS-visualization” integrated technological framework. Building upon this, a multitiered intelligent laboratory safety management framework comprising the following specific steps is designed: (1) a laboratory safety knowledge graph is constructed based on multisource data. The graph deploys the ALBERT-BiLSTM-CRF model for entity recognition and uses the Neo4j graph database for knowledge storage and associative reasoning;(2) an intelligent agent is developed for risk perception and digital-intelligent control. A closed-loop “perception-identification-decision-making” system is constructed, and the Dempster–Shafer evidence theory is integrated with dynamic risk assessment models and a cloud computing platform for real-time monitoring and intelligent discrimination of multidimensional information on personnel, equipment, environment, and procedures;(3) A university laboratory safety risk assessment and intelligent monitoring/early warning system is established, integrating multimodal sensing networks, dynamic knowledge graphs, and agent-driven decision-making to form a complete technical chain spanning data collection, risk modeling, and early-warning response.

The five-in-one technology proposed in this paper offers an innovative approach to laboratory safety management in higher education institutions, reflecting the trend toward the deep integration of modern information technology and safety management. Although the proposed technology still faces challenges such as high implementation costs, data security risks, and a shortage of specialized personnel during roll-out, with continuous advancements in technologies, such as artificial intelligence, digital twins, and edge computing, safety inspection robots with enhanced intelligence, visualization, and automation capabilities are expected to be developed in the future. By integrating multimodal sensing, intelligent algorithms, and visual interaction, these robots will be capable of real-time monitoring of environmental parameters, equipment status, and personnel behavior, automatically performing risk assessments and issuing tiered alerts. This will ensure round-the-clock intelligent inspection support for laboratories.

This research not only drives the continuous development of laboratory safety management toward greater precision, proactivity, and universal accessibility but also provides a solid guarantee for the high-quality advancement of higher education and the safety of faculty and students during experimental research.

Laboratory safety is a fundamental prerequisite for scientific researcher to conduct research experiments effectively. Environmental laboratories, which support interdisciplinary research in areas such as wastewater treatment, air pollution control, environmental microbiology, and environmental chemistry, are particularly critical. These facilities play a vital role in developing graduate students’ practical, innovative, and scientific research competencies. In recent years, the continuous expansion of graduate enrollment has led to a corresponding increase in the scale of environmental research laboratories. Concurrently, the variety and quantity of hazardous reagents, such as trinitrophenol, trinitrobenzene, and nitrocellulose, have risen significantly. Improper management or non-compliance with operational protocols in these laboratories can result in serious safety and regulatory violations, potentially leading to accidents such as explosions, poisoning, burns, or fires. Given these risks, this study examines key issues in safety standardization and management within environmental graduate research laboratories, aiming to enhance operational safety and regulatory compliance.

There are many existed problems for the safety specification and management of environmental graduate scientific research laboratories in Guangxi University, including the lack of early laboratories planning, aging facilities and insufficient space. In addition, there are short of sufficient and dedicated laboratory safety staffs, which cause the graduate student tutors and graduate students are insufficient awareness of dynamic safety challenges in interdisciplinary research fields. Moreover, both these graduate student tutors and graduate students face heavy research tasks and cannot take sufficient time to rectify the safety problems in research laboratories. These above-mentioned reasons have brought various types, large quantities and difficult rectification of safety hidden troubles. Therefore, there are major problems in environmental graduate scientific research laboratories such as chaotic spatial zoning, incomplete safety management systems and poor effectiveness of safety education. In combination with the specific safety issues existing in the environmental laboratories of Guangxi University, three strategies and a series of practical safety management standards are proposed from the aspects of laboratory safety education and training. It mainly includes institutional guarantees, experimental reagents & equipment management, waste supervision & management. These specific measures include: safety regulation management Firstly, Guangxi University pays more attention to the flexibility and multi-functionality of laboratory layout; Secondly, our university continuously optimize safety measures including fire prevention, explosion prevention, and gas defense. Thirdly, the professional laboratory staffs regularly conduct laboratory safety regulations inspections every month, supervise teachers and students to strictly abide by the laboratory safety regulations management system.

Now Guangxi University has form a “standardization-implementation-inspection-rectification-recheck” safety management mode. Finally, we introduce digital supported management methods in the background of rapid technological development. Guangxi University has comprehensively improved the safety management level of environmental graduate scientific research laboratories: optimizing the functional space design of laboratories, establishing the digital safety standard management system and introducing high-quality hardware resources for scenario-based safety training such as intelligent management, virtual simulation, science popularization emergency scenario exercises and VR training.

The study provides a theoretical reference for improving the safety management system of research laboratories of environmental category, strengthening laboratory safety level and ensuring the property safety of teachers and students. These optimized safety standards management not only helps to ensure the safety of public property and teachers & students, but also serves as an important guarantee for promoting the in-depth development of scientific research. It can provide reference for the standardized construction and management of research laboratories in sister universities.