Abstract

Autoencoders (AEs) are widely used in industrial gas turbines for fault diagnosis. However, traditional AEs often perform poorly with limited training data, causing frequent false alarms. Insufficient data can cause a mismatch between the model and the actual system, especially under operational conditions absent from the training samples. This paper introduces a novel multi-fidelity autoencoder (MFAE) model that integrates limited high-fidelity operational data with abundant low-fidelity simulation data. The MFAE model learns latent features from a pre-trained low-fidelity AE model and establishes correlations between low- and high-fidelity data via linear and nonlinear networks. The effectiveness of the proposed methods is evaluated on an industrial gas turbine unit. MFAE effectively captures essential process characteristics and adapts to various operating conditions with limited operational data, enabling accurate gas turbine monitoring.

Abstract

Hydrogen supply system needs to provide sufficient mass flow and pressure of hydrogen to meet the requirement of fuel cell stack in PEMFC system. Ejector and hydrogen circulating pump (HRP) are commonly used to recycle unreacted hydrogen to improve energy utilization and system efficiency. However, ejector can not achieve active control and has narrow efficient working range, while HRP leads to extra parasitic power. This paper establishs an accurate hydrogen supply system model and proposes single recycle component scheme and combination of ejector and HRP scheme, discussing the working performance of different circulation structure scheme. The simulation results indicate that single ejector can not recycle enough flow rate and even counterflow occurs with fuel cell system output power less than 60kW. Single HRP controls the hydrogen recycle flow rate accurately but an average of 280.1W parasitic power is consumped. In ejector and HRP parallel scheme, matching control strategy is proposed and an average of parasitic power can be reduced 190W than single HRP scheme in whole fuel cell stack working operation. In ejector and HRP serial scheme, matching strategy controls HRP establish high pressure difference to help overcome flow resistance from ejector with maximun 0.45kPa in low current loading condition. When system output power exceeds 60kW, HRP is shut off and ejector overcomes a maximun flow resistance of 22.25kPa from HRP. An average of parasitic power is reduced 200.8W than single HRP scheme.

Abstract

Attributes is highly heterogeneous along horizontal wells and the water breakthrough mechanism is complex in the bottom-water oil reservoir. The conventional numerical and physical simulations are laborious and have a limited contribution to on-site water plugging operations. This paper presents an effective method for clarifying the water breakthrough rule of horizontal wells. Firstly, the water drive mathematical model of the heterogeneous production area was constructed by analyzing well-logging curves. Secondly, the adaptive iterative algorithm was adopted to realize the prediction of the water breakthrough parameters along horizontal wells, based on the non-piston water drive theory. Finally, the validity of the innovative model is verified through comparison with a rigorous physical simulation experiment. The results show that the water breakthrough time difference along the horizontal well and the water cut stage height of the single well are positively correlated with the permeability differential. The permeability differential is negatively correlated with the utilization degree of low permeability area. The increase of the area of the high permeability section delays the water breakthrough time and has a great influence on the water cut of single well. With the increase of water avoidance height in high permeability area, the water breakthrough time of oil well is delayed, and the height of water cut stage of single well is reduced. Finally, taking a horizontal well of an oilfield in Bohai Bay as an example, this paper clarifies the water breakthrough rule of the 9th member of the horizontal well and identifies more remaining oil in the first and 4th members. This research has an important theoretical instructional significance to formulate reasonable single well control water plugging measures.

Abstract

The energy performance of 3D printed buildings has been of great interest. This study focuses on the layer-by-layer surface shapes of 3D printed buildings and simulates the effect on the heat transfer process. The results demonstrate that the layered surface leads to a periodic fluctuation in surface temperature, with an average surface temperature closer to the ambient temperature. Although the surface heat flow decreases, the total surface heat transfer increases. According to the results, the layered surface affects the surface temperature distribution and surface heat transfer, which cannot be directly ignored in the study.

Abstract

Adsorption-based carbon capture is a promising carbon reduction technology for combating climate change. However, the desorption of CO2 from the adsorbent demands a high energy input, which limits this technology‘s further development. In the desorption step of a temperature swing adsorption (TSA) cycle, the heat transfer from the heat source to the adsorbent and the mass transfer of CO2 from the adsorption phase to the gas phase inevitably leads to exergy losses. Moreover, both heat and mass transfer are typically initially fast and then slow down, as the imbalance potential differences starting large but decreasing over time. The initial large potential differences accelerate the process but cause significant exergy losses. A more uniform potential difference distribution over the entire time scale is expected to achieve better balance between the exergy losses and desorption time. In this regard, this study proposes an active control strategy, that is, regulating the potential difference by process temperature control to reduce exergy losses. A one-dimensional numerical model for carbon capture was established to validate the proposed strategy. The optimal temperature-time condition was identified for a case study to guide the process design. The exergy demand of desorption step is reduced effectively under the active temperature control strategy. This study aims to minimize the exergy losses thereby contribute to the large-scale deployment of adsorption-based carbon capture.

Abstract

This study addresses the critical challenges of thermal management in Li-ion batteries for electric vehicles (EVs) to ensure optimal performance. The study focuses on assessing two passive cooling methods: Phase Change Material (PCM) and passive immersion. The study compares their thermal performance against natural air cooling. The investigation involves mathematical modelling, numerical formulation, and validation of the electrothermal and PCM models, showcasing their efficacy in simulating Li-ion cell thermal dynamics. The results indicate that the PCM-based thermal management system performs better than others, achieving a remarkable temperature reduction of 11.02 °C and maintaining safe operating temperatures. Moreover, the PCM-based system demonstrates an impressive energy density of 136.88 Wh/kg, outperforming other passive immersion methods. This study contributes valuable insights for enhancing the thermal performance of Li-ion cells. It also contributes to understanding passive Li-ion cell thermal management strategies, emphasizing the pivotal role of PCM-based cooling solutions for improved safety, efficiency, and lifespan in various applications.

Abstract

Gas turbine propulsion systems are widely used in modern commercial ships due to their high power density, reliability, and operational flexibility. However, their efficiency often drops significantly under varying load conditions. To address this issue, a novel intercooled and recuperative gas turbine system integrated with transcritical organic Rankine cycle (ICR-GT-TORC) was proposed in the present work. The system performance was evaluated under both design and off-design conditions using thermodynamic and off-design models developed with Thermoflex software. Comparative analysis with conventional UGT25000 marine gas turbines revealed that the ICR-GT-TORC system achieved an efficiency of 49.4% and a power output of 38.1 MW, reflecting improvements of 29.3% and 37.6%, respectively. Moreover, the ICR-GT-TORC and ICR-GT system demonstrated significant advantages under part-load conditions. At 30% load, the efficiency of the ICR-GT-TORC system remained high at 43.7%, with a reduction of 11.5%. The ICR-GT system maintained an efficiency of 40.9%, with only a 7.9% reduction. In contrast, the efficiency of the simple cycle dropped significantly to 28.4%, with a 25.7% decrease. These findings could provide valuable insights into optimizing marine propulsion systems for off-design performance.

Abstract

The gas diffusion layer is one of the most important components of the proton exchange membrane fuel cell (PEMFC). Microstructure reconstruction is an efficient research strategy. The influence of the morphological presence of resin binder when applying a stochastic method to reconstruct the three-dimensional gas diffusion layer was studied. Finite element method (FEM) simulation was conducted to explore the mechanical response under compression load for both reconstruction models with the binder morphologically present or absent. The differences in pore structure of the two reconstructed models before and after compression were meticulously analyzed. The morphological appending of the binder made the overall equivalent modulus larger, but the deformation and stress concentration became severe. When subjected to a compression ratio of 20%, the porosity changed to 74.4% and 74.7% for structures with and without the presence of binder morphology, respectively. Binder morphological addition altered the pore space greatly, making it a more uniform pore size distribution and a smoother porosity variation. In the through-plane (TP) direction, the morphological presence of the binder decreased the tortuosity, while in the in-plane (IP) direction, the tortuosity increased.

Abstract

Block C of Sinopec Shengli Oilfield is a typical shallow-thin ultra-heavy oil reservoir. After multiple cycles of steam huff and puff, the reservoir faces issues such as low utilization, low well production, and high water cut. Based on the geological and production parameters of the target block, a three-dimensional physical simulation experiment was designed to transition from steam huff and puff to multi-component thermal composite flooding. The study examined the temperature field and production dynamics characteristics at different stages. The results indicate that during the steam huff and puff stage, the heating range of the reservoir was limited, resulting in a poor development effect, with an oil recovery rate of only 10.4%. In the multi-component thermal composite flooding stage, through the synergistic effect of viscosity reducers, nitrogen, and steam, the final oil recovery reached 57.1%. These findings provide useful insights for the development of shallow-thin ultra-heavy oil reservoirs.

Abstract

Understanding the impact of urban land use patterns on energy-related carbon emissions is critical for developing effective climate change mitigation strategies. This study employed machine learning techniques to model the relationship between multidimensional urban land use characteristics and city-scale carbon emissions. Urban land use was characterized across four dimensions: scale, structure, mixture, and intensity. Machine learning algorithms, including CART, Random Forest, and XGBoost, were trained to quantify the relative importance of these land use features in predicting carbon emissions. The machine learning models demonstrated strong predictive performance, outperforming traditional linear regression. The feature importance analysis revealed that urban land use indicators collectively account for over one-quarter of the models’ predictive power, with land use scale, structure, and intensity features exhibiting greater importance than socioeconomic variables. These findings underscore the value of data-driven, nonparametric modeling approaches in elucidating the complex, multifaceted links between urban form and greenhouse gas emissions.