四川轻化工大学

Abstract This study systematically characterized volatile organic compound (VOC) compositions across cuisines in China, examined how edible oils affect VOC/polycyclic aromatic hydrocarbon (PAH) emissions, and evaluated associated health risks using both deterministic and Monte Carlo probabilistic models, from which a tiered health-risk early-warning classification was derived. Results showed distinct cuisine-specific VOC profiles: high-temperature frying cuisines [barbecue (BB) and western fast food (KFC)] were dominated by alkanes, whereas cooking-wine cuisines [Hunan cuisine (XC) and western cuisine (WC)] had high alcohols (13.2%–29.2% ethanol); Sichuan cuisine (CC) showed high VOC compositional similarity with the other cuisines ( similarity > 0.6 ). Rapeseed oil emitted the most particulate matter, soybean oil emitted the most nonmethane hydrocarbons, and both produced the highest PAHs, notably anthracene ( 276.88 μg m − 3 ) and benzo[a]pyrene. Noncarcinogenic risk arose mainly from acrolein, and the lifetime carcinogenic risk (LCR) exceeded 1.0 × 10 − 6 for all cuisines. VOCs contributed 86.3%–94.1% of total cancer risk and PAHs 5.9%–13.7%, the latter dominated by benzo[a]pyrene. Applying the proposed five-level early-warning scheme, high-oil, high-temperature methods (barbecue and Sichuan cuisine) reached the highest warning levels, and occupational workers were consistently classified one to two levels above nearby residents. The results highlight the need for targeted control of key pollutants (e.g., acrolein and 1,3-butadiene, benzo[a]pyrene) and for occupational protection in catering kitchens.

Fermentation is a key process in food manufacturing that determines product quality, safety, and nutrition. Traditional offline and manual monitoring methods are inefficient and lack real-time adaptability. Advances in intelligent food processing have made real-time monitoring and control essential for precise regulation, improved consistency, and higher production efficiency. This review summarizes recent developments in intelligent fermentation monitoring, focusing on spectroscopic and sensing technologies. Artificial intelligence (AI), data-driven modeling, and the Internet of Things (IoT) are highlighted as enabling tools for data integration and intelligent decision-making. Applications in beverages, tea, coffee, and dairy systems are also discussed. These technologies improve monitoring accuracy and reduce response time, enhancing process consistency and efficiency. However, challenges remain, including limited model generalization, sensor instability, and fragmented data systems. Future work should focus on robust miniaturized sensors, interoperable platforms, and integrated IoT–cloud frameworks for scalable real-time fermentation management.

During Baijiu brewing, the physicochemical parameters of Daqu serve as crucial indicators for evaluating its quality. However, traditional testing methods require sample destruction and lack real-time monitoring capabilities, making it difficult to meet the demands of intelligent production in the modern brewing industry. This study was aimed at proposing a rapid, accurate, and nondestructive testing method for Daqu physicochemical parameters based on machine learning models. The environmental parameters and physicochemical data of Daqu were collected to extract the key environmental time period features using XGBoost, RF, and ReliefF algorithms. Also, a support vector regression (SVR) model was constructed to predict acidity and moisture content. The prediction accuracy of acidity was further enhanced by introducing a regression chain (RC) optimization algorithm combined with a backpropagation neural network (BPNN) to establish a weighted integrated prediction model. The XGBoost–SVR model demonstrated the best predictive performance for moisture content and acidity, with R2 values reaching 0.9686 and 0.8776, respectively. The optimization through the RC–BPNN integrated model improved the R2 for acidity prediction by 4.5

The precision diagnosis and treatment of tumors have long faced challenges, including difficulty in early detection and poor treatment specificity. At the same time, the separation of diagnosis and treatment in the traditional medical model further delays treatment decisions. It limits the ability to adjust flexibly to patients' individual circumstances. In recent years, covalent organic framework (COF), as an emerging class of crystalline porous materials with high surface area, customizable pore structures, excellent optoelectronic properties, and abundant functional sites, have provided an ideal platform for breaking through the bottlenecks of diagnosis and treatment and for building smart tumor diagnosis and treatment systems. This review systematically discusses the recent research progress and applications of COF materials in tumor diagnosis and treatment, focusing on how COF materials are applied in tumor diagnosis and therapy through structural design and functional modification to construct a tumor diagnosis and treatment system. Firstly, this article outlines the cutting-edge applications of COF sensors for detecting key biological markers, such as tumor metabolic markers. COF improves the sensitivity and specificity of the sensors, promising early and precise diagnosis of tumors, providing real-time data for subsequent treatments, and laying the 'perception' foundation for the tumor diagnosis and treatment system. Next, it focuses on how COF can serve as a drug carrier to achieve targeted drug delivery and synergize multiple treatment modalities. It achieves a seamless connection from diagnosis to treatment, breaking down the traditional separation between diagnosis and treatment. On this basis, it further explores how to turn the detection signal into a decision-making trigger, enabling real-time monitoring of treatment efficacy and constructing a tumor diagnosis and treatment system. Finally, the challenges currently faced by the COF tumor diagnosis and treatment system, as well as its future development and prospects for clinical translation, were discussed.

Natural fractures are regarded as important reservoir space and seepage channels in low-permeable sandstone reservoirs. Taking the complex multi-scale fractures of the Xujiahe Formation in the YB area as a case study, the characteristics and controlling factors are analyzed using data from multi-scale observations, including core samples, thin sections, and field outcrops. Multi-scale fracture distribution is predicted through a combination of geological analysis, fractal theory, and rock fracture theory. Specifically: (1) It is dominated by tectonic shear fractures of high angle (45°–75°) and vertical (75°–90°), the fracture density (D) in sandstone is higher than mudstone, negatively correlated with the thickness of the rock mechanical layer (RML). (2) A comprehensive fracture prediction method based on weight theory is proposed, where gray correlation analysis is used to determine the weight coefficients of different methods’ prediction results. The comprehensive fracture development index (IF) accurately delineates the fracture distribution range. The greater the IF value, the greater the fracture D. (3) The fracture development area is classified into three grades (Grade I, II, and III). A strong positive correlation exists between the IF value and fracture D. Grade I areas, with IF ≥ 2.40, are mainly distributed near faults and the high parts of nose structures. The comprehensive fracture evaluation method overcomes the limitations of individual methods and provides a more precise characterization of fracture development.