Korea Testing Laboratory

Abstract Quantitative knowledge of competitive adsorption isotherms is essential for the design and optimization of adsorption-based separation processes. To adjust the retention times of the components to be separated in liquid phase applications, mixtures of two or more solvents are often selected as the mobile phase. The solvent composition thus has an influence on the courses of the adsorption isotherms. The paper presents the results of breakthrough curve (BTC) experiments considering the separation of three benzene derivatives characterized by different aliphatic carbon chains, exploiting conventional reversed-phase chromatography (RP18). As the mobile phase, different mixtures of acetonitrile (ACN) and water (W) were used. Equilibrium loadings were extracted from characteristic features of the observed BTCs for single components, binary mixtures, and ternary mixtures. The analysis ignored kinetic effects and exploited classical equilibrium theory. For an initially selected solvent composition, the courses of the isotherms revealed an anti-Langmuir behavior for the two longer retained components. All single-component and mixture equilibrium data could be well described with a newly derived widely applicable multi-component single-site BET adsorption model, assuming a finite number of adsorbed layers. The analysis was applied again to determine and analyze the courses of BTCs and the corresponding multi-component adsorption isotherms for three other solvent compositions. The slopes of the single-component isotherms changed significantly for the different components. The derived BET-type equation could again be used to describe rather well the isotherms for increasing amounts of ACN with specific numbers of adsorbed layers. For ACN/W = 90/10, the layer number approached unity and, thus, an almost perfect Langmuir behavior was identified. The identified trends between the solvent compositions and the estimated isotherms model parameters could be well described with the LSS model. Finally, suggestions were made for further improving the agreement between measured and predicted isotherms.

Abstract Pre-trained vision-language models ( e.g ., CLIP) have shown impressive success in various computer vision tasks with their generalization capability. Recently, parameter-efficient fine-tuning (PEFT) approaches have been actively explored to effectively and efficiently adapt the pre-trained vision-language models to a variety of downstream tasks. However, most existing PEFT approaches suffer from a task overfitting issue since the general knowledge of the pre-trained models is forgotten while a small number of learnable parameters in soft prompts/adapters are fine-tuned on a small data set from a specific target task. Thus, we propose a P arameter- E fficient F ine- T uning via Meta - R egularization (PEFT-MetaR) to improve the generalizability of parameter-efficient fine-tuning methods for vision-language models. Specifically, PEFT-MetaR meta-learns both the regularizer and learnable parameters to harness the task-specific knowledge from the downstream tasks and task-agnostic general knowledge from the pretrained models. Further, PEFT-MetaR augments the task to generate multiple virtual tasks to alleviate the meta-overfitting. In addition, we provide the analysis to comprehend how PEFT-MetaR improves the generalizability from the perspective of the gradient alignment. Our experiments demonstrate that PEFT-MetaR improves the generalizability of parameter-efficient fine-tuning methods on various datasets.

This study proposes a reinforcement learning-based link selection method for Low Earth Orbit satellite networks, aiming to reduce handover frequency while extending link duration under highly dynamic orbital environments. The proposed approach relies solely on basic satellite positional information, namely latitude, longitude, and altitude, to construct compact state representations without requiring complex sensing or prediction mechanisms. Using relative satellite and terminal geometry, each state is represented as a vector consisting of azimuth, elevation, range, and direction difference. To validate the feasibility of policy learning under realistic conditions, a total of 871,105 orbit based data samples were generated through simulations of 300 LEO satellite orbits. The reinforcement learning environment was implemented using the OpenAI Gym framework, in which an agent selects an optimal communication target from a prefiltered set of candidate satellites at each time step. Three reinforcement learning algorithms, namely SARSA, Q-Learning, and Deep Q-Network, were evaluated under identical experimental conditions. Performance was assessed in terms of smoothed total reward per episode, average handover count, and average link duration. The results show that the Deep Q-Network-based approach achieves approximately 77.4% fewer handovers than SARSA and 49.9% fewer than Q-Learning, while providing the longest average link duration. These findings demonstrate that effective handover control can be achieved using lightweight state information and indicate the potential of deep reinforcement learning for future LEO satellite communication systems.

We report a solvent-free melt polycondensation strategy for synthesizing a partially biodegradable isosorbide-based polycarbonate (ISB-based PC) incorporating ethylene oxide (EO)-functionalized comonomers.

The effects of color coating and insulation on the power generation performance of buildingintegrated photovoltaic (BIPV) systems were experimentally analyzed.A mock-up test bed was constructed, and the irradiance and power generation were measured for one year using south-facing modules.The analysis included colorless (reference), red (Color A), and dark blue (Color B) modules under both insulation (I) and noninsulation (NI) conditions.The output of the Color B module was higher than that of Color A owing to the higher light absorption efficiency of the shorter blue wavelengths.Under insulation conditions, the performance ratio (PR) of Color A was lower by -12% to -24% and that of Color B was lower by -2% to -10% than that of the reference, mainly owing to heat accumulation from suppressed rear heat dissipation.The power generated by BIPV modules is strongly influenced by building envelope factors, such as the coating color and insulation structure.

Korea has a high population density, considering the size of its territory. Therefore, the importance of convenient and comfortable apartment buildings and high-rise residential–commercial complex buildings has been rising. In addition, because of the improvement in the standard of living along with continuous national economic growth, the interest in well-being and the expectation of a quiet life with a comfortable and pleasant residential environment have also been increasing. However, Koreans have a lifestyle involving sitting on the floor, so floor impact noise has been occurring more and more frequently. Because of this, neighborly disputes have been a serious social problem. And lately, damage and disputes from noise between floors have been increasing much more. The present work, therefore, used waste tire chips as a resilient material for reducing floor impact noise in order to recycle waste tires effectively. Also, a compounded resilient material, which combines EPS (expanded polystyrene), a flat resilient material on the upper part, with waste tire chips for the lower part, was developed. After constructing waste tire chips at a standardized test building, experiments with both light-weight and heavy-weight floor impact noise were performed. The tests confirmed that waste tire chips, when used as a resilient material, can effectively reduce both light-weight and heavy-weight floor impact noise.