中国石油大学（华东）

Achieving social sustainability becomes a key debate after the introduction of sustainable development goals globally. However, the existing strategies like green manufacturing, and circular economy proposed for industries more focussed on economic and environment sustainability and have a little direct impact on social aspects. Considering the fact, studies started to correlate the impact of corporate social responsibility as one of the key strategies for achieving social sustainability goals. There are many existing studies exists with tools and strategies which effectively implement corporate social responsibility. Most of these studies stressed the importance of responsible investment for effective implementation of corporate social responsibility. This study seeks the opportunity to explore and understand the relationship among responsible investment practices which could drives the effective implementation of corporate social responsibility for achieving social sustainability goals. Totally 16 common responsible investment practices were considered in this study, which further analysed through DEMATEL-ANP with the assistance of case study methodology. The findings revealed that among 16 considered common practices, “investment on communications” (RP13) holds the top position followed by “Invest on socially responsible mutual funds and exchange traded funds” (RP12). Contrary to this, “providing micro finance to local startups within the mining ecosystem” (RP16) holds the least influence responsible investment practice, which is followed by investment on training and education of workers to understand their social rights (RP14). This study contributes to strengthen the theoretical discussions by including social aspects in company’s long term investment practices. The obtained results were used to provide key managerial implications including finding different strategies to enhance (RP13), which could further be considered by industrial managers for implementation of corporate social responsibility in their organization in effective way for long term goal achievement.

Raman spectroscopy is a non-destructive analytical technique based on molecular vibrational properties. However, its practical application is often challenged by weak scattering signals, complex spectra, and the high-dimensional nature of the data, which complicates accurate interpretation. Traditional chemometric methods are limited in handling complex, nonlinear Raman data and rely on tedious, expert-knowledge-based feature engineering. The fusion of data-driven Machine Learning (ML) and Deep Learning (DL) methods offers a robust solution, enabling the automatic learning of complex features from raw data and achieving high-accuracy classification and prediction. The present study employed a structured narrative review methodology to capture the research progress, current trends, and future directions in the field of ML-assisted Raman spectral classification. This review provides a comprehensive overview of the application of traditional ML models and advanced DL architectures in Raman spectral analysis. It highlights the latest applications of this technology across several key domains, including biomedical diagnostics, food safety and authentication, mineralogical classification, and plastic and microplastic identification. Despite recent progress, several challenges remain: limited training data, weak cross-dataset generalization, poor reproducibility, and limited interpretability of deep models. We also outline practical directions for future research.

Traditional methods for ensuring security and privacy face challenges in safeguarding multimedia data within the IoT-edge continuum, as their significant computational demands render them unsuitable for IoT devices with limited resources. Next, we find that the federated learning techniques can naturally adapt to edge frameworks and provide effective data security and privacy protection. In this paper, we propose FLiForest, an innovative anomaly detection approach that integrates federated learning with the isolation forest algorithm, tailored for the IoT-edge continuum. Specifically, our method designs a three-stage process, including data collection and sampling, model training, and data testing, to train an isolation forest among clients and edge servers jointly. In the training of each layer, all clients upload parameters to the central server for aggregation. FLiForest facilitates decentralized model training across IoT devices, enhancing data privacy and reducing computational burden, without necessitating the exchange of multimedia data. Through extensive experiments on a variety of multimedia datasets, the efficacy of our method is benchmarked against the state-of-the-art anomaly detection methods, showcasing its superior detection accuracy and robustness in ensuring data privacy and security.

Controlled shrinkage of open pores during carbonization enables closed pore formation in hard carbons as the anode for sodium ion batteries, though excessive open pores resist conversion and degrade electrochemical performances. Employing Mg2+ as a pore-forming agent chelated by humic acid, we constructed tailored closed pore architectures through pre-carbonization at 600 degrees C followed by 1500 degrees C treatment. The resulting hard carbons exhibit tunable interlayer spacing and disorder, with closed pores of uniform size (1.10-1.18 nm) yet distinct surface areas (307.4-408.3 m2/g). The optimized hard carbons deliver a high reversible capacity (262 mAh/g at 20 mA/g), excellent rate capability (52 % retention at 1000 mA/g), and cycling stability (75 % after 1000 cycles at 500 mA/g). Intercalation capacity correlates with pseudo-graphite carbon content, while porefilling capacity scales with closed pore surface area. This study paves the way for rational engineering of closed pores in hard carbons.

This study systematically investigates the hydrogen embrittlement (HE) behavior of X65 pipeline steel base metal (BM) and weld metal (WM) in gaseous hydrogen environments. The hydrogen permeation characteristics, fatigue properties and fracture behaviour of both the BM and WM under various hydrogen partial pressure environments were considered intensively. Electron backscatter scanning diffraction (EBSD) was employed to characterize crystallographic features of the BM and WM. The results indicate that the BM exhibits no pronounced texture and contains a low proportion of high angle grain boundaries (HAGBs). The WM shows distinct texture and possesses relatively low dislocation density. Under in-situ gaseous hydrogen charging conditions, the hydrogen diffusivity for various regions of WM are about half an order of magnitude greater than that of the BM. With the increase of hydrogen partial pressure, the correlation between partial pressure and the fatigue crack growth rate (FCGR) weakens. At the hydrogen partial pressure of 1.26 MPa, the FCGRR of WM is approximately 1.4 times that of the BM, which is more susceptible to hydrogen effects. Macroscopic and microscopic analyzes of the specimen fracture surfaces were conducted using a 3D super depth of field microscope and a scanning electron microscope (SEM). Comparing with BM specimens, WM specimens accompany more abundant brittle fracture characteristics and exhibit lower fracture toughness. Under the identical hydrogen partial pressure, the embrittlement index (EI) of BM and WM are comparable, whereas the fracture toughness of WM was approximately 37.7 % and 14.6 % lower than those of BM respectively.