To address the low efficiency in developing catalysts for CO2 hydrogenation to methanol, this study constructs and validates an intelligent performance prediction model based on large language model (LLM) and deep learning. First, a Large Language Model (LLM) to design structured prompts, achieving semi⁃automated and high⁃efficiency extraction of multi⁃dimensional catalyst data from literature. Subsequently, a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN⁃GP) is employed to augment the sparse original dataset, effectively overcoming the bottleneck of data scarcity. Following data cleaning, feature engineering, and dimensionality reduction, a hyperparameter⁃optimized Multi⁃Layer Perceptron (MLP) is constructed as the prediction model. The results show that the optimized MLP model achieves high prediction accuracy on an independent test set, with R² values for CO2 conversion and methanol selectivity reaching as high as 0.972 3 and 0.969 3, respectively. SHAP⁃based feature analysis reveals that BET surface area and Cu⁃based catalysts are the dominant factors affecting catalytic performance, and also uncovered the unique dependency of In⁃based catalysts on metal content. This data⁃driven model, integrating LLM and WGAN⁃GP, provides a powerful tool for the rapid screening and rational design of novel catalysts, demonstrating the great potential of AI in catalysis research.

Card sleeve joints are widely used in the connection of hydraulic and pneumatic equipment such as oil and gas pipelines, and its connection reliability has an important impact on the safety of oil and gas pipelines. However, there is no report on the influence of external working conditions on the stress characteristics of the thread of card sleeve joints in oilfield ground pipelines. A three?dimensional model of the double card sleeve joint was established using SolidWorks software; The maximum equivalent stress (Von Mises stress) of the sleeve joint was numerically analyzed using ANSYS finite element software. The results indicate that within the yield limit, the greater the axial force, the better the sealing performance of the sleeve joint, while the internal pressure has little effect on the sealing of the pipe joint, and the risk of thread sticking due to excessive stress can be ignored; The influence of thread parameters on sealing performance is significant; The optimal pitch and number of threads for a sleeve joint with an outer diameter of 12 mm and an inner diameter of 9 mm are 1.5 mm and 7, respectively. The research results can provide theoretical basis and reference for the optimization of structural performance and scientific operation in the assembly process of card sleeve joints, which has important engineering significance.

With the decline of conventional crude oil production, the exploitation and application of heavy oil has gradually attracted attention. However, heavy oil with high density, large viscosity and difficult mining is the problem faced by heavy oil production. There are many ways to reduce the viscosity of heavy oil, among which upgrading visbreaking reaction and non-upgrading visbreaking reaction are two important ways to reduce the viscosity of heavy oil. The latter include heating and viscosity reduction, thinning and viscosity reduction, and chemical viscosity reducing and so on. There are two methods for upgrading visbreaking reaction, namely mild thermal cracking and aquathermolysis. The aquathermolysis viscosity reduction technique is a more efficient method of viscosity reduction based on thermal cracking. Under the appropriate reaction conditions and suitable catalyst conditions, the viscosity of heavy crude oil can be greatly reduced, which is a heavy oil mining technology with good application prospects.

Using a high-temperature and high-pressure ultrasonic reactor, the hydrothermal cracking reaction of vacuum residue of Huizhou Refinery. The gas product composition, liquid product composition, reaction coke SEM morphology, and element analysis were investigated to discuss the mechanism of hydrothermal cracking reaction under the effect of ultrasonic wave. The results show that there is no significant difference in the yield of gas products in hydrothermal cracking reaction mechanism of the residue under the effect of ultrasonic wave and the production rate of light oil has been increased slightly while that of coke has been slightly reduced. The angularity of the coke particles is relatively smooth and the pore structure appears, indicating that under the effect of ultrasonic wave, the hydrothermal cracking reaction of residual oil is mainly based on the radical thermal reaction mechanism and the cavatition of ultrasonic wave can promote the residuum to have cracking reaction so that the production rate of light oil has been slightly increased. The mechanical effect of ultrasonic wave causes the raw coke precursor suspended in the residue to oscillate violently, so as to prevent its polymerization, thus having reduced coking, the large increase of Ni in the coke indicates that the catalyst provides a coking center for coking.

 

Differential operator elimination method was applied to solve the Maxwell's system of partial differential equations obtain to the general expressions of transverse permeability for magnetic ribbons. The expressions of transverse permeability for magnetic ribbons were found out in driving current full frequency section with the Landau-Lifshite equation. The results show that under the alternating magnetization the transverse permeability of ribbons relate to external magnetic field and driver electric current frequency，and has be orientation-related relativity as well.

hermal decomposition behaviours of no woody biomass (straw) and waste plastic (agricultural film) were investigated using TGA．The results showe that the biomass is decomposed at a wider temperature range than plastics, and the transfer efficiency of biomass is lowest because of the high content of ash and fixcarbon．It is exhibited significant synergistic effect created more the light component between biomass and plastic during co-pyrolysis. The kinetic analysis indicates that the pyrolytic processes can be described as first order reactions model, a quite good fitting of experimental data was obtained for all samples studied．The only plastic can be described as the one model, and the only biomass be described as the two consecutive models, than the biomass/plastic co-pyrolysis need be described as the four consecutive models. The activation energies were found to be in the rang of 64.6 ~ 306.6 kJ/mol, and the pre-exponential factors were 1.1×104～3.0×1022. 

The solubility of p-aminophenol in water, ethanol and the mixture of water and ethanol was determined respectively by the test apparatus established in the lab. The experimental water-solubility of p-aminophenol is in good consistency with the literature data, which confirms the reliability of the experimental apparatus and method. Using solubility model and S-H activity coefficient equation, the parameters lij of the solubility equation were obtained by correlating the solubility data. The thermodynamic model of the calculating solubility of p-aminophenol was obtained, which prove to be useful in predicting the solubility of p-aminophenol in the mixture solvent at higher temperature. The correlated solubility model and calculated results are consistent well with the experimental data.