The accurate prediction of gas turbine output power is critical for flexible scheduling and shipboard microgrid resilience. However, purely data-driven models suffer from poor generalization and physical inconsistency in complex marine environments, especially under unseen operation conditions. This paper proposes a condition-adaptive physics-informed long short-term memory (CAPI-LSTM) framework to ensure physical consistency across the full operation envelope. In the proposed framework, an MLP-based condition-adaptive regulator is developed to dynamically adjust the compressor air flow rate within the embedded physics-informed loss function. The proposed CAPI-LSTM model is verified using the operation data from an LM2500+ gas turbine. The comparison results demonstrate the superiority of the proposed method over traditional architectures. The CAPI-LSTM model achieves the lowest root mean square error of 0.177 MW, and its error distribution is the most concentrated near zero among all compared models. The robustness of the CAPI-LSTM model is further verified under the unseen operation conditions. The CAPI-LSTM still maintains excellent generalization capability compared to both purely data-driven models and standard physics-informed models, with an average error of only 0.218 MW and a narrow interquartile range of [0.058, 0.363]. The paired t-test results confirm that the improvement of the CAPI-LSTM model is statistically significant. The CAPI-LSTM model achieves competitive computational efficiency despite the integration of the physics-informed loss function with a condition-adaptive regulator. Furthermore, the CAPI-LSTM model achieves superior performance in noise immunity and transferability to other types of gas turbines. In summary, the proposed CAPI-LSTM model provides an effective and practical solution for marine gas turbine output power prediction.

Data-driven algorithms provide a powerful method for rapid gas path analysis (GPA) of gas turbines. The conventional GPA models based on only sensor data could be improved in the generalization capability. Incorporation of knowledge offers a potential solution to enhance the diagnosis performance of data-driven GPA models. This paper proposes a novel Quantified-Knowledge Guided Graph Convolutional Network (QK-GCN) model to enhance the physical interpretability and generalization. The QK-GCN model embeds dual forms of knowledge: firstly, the graph topology is constructed from the physical relations among 9 gas path sensors, identified through reachability analysis of the thermodynamic equations; secondly, the corresponding edge weights in the graph are assigned by quantifying the relation similarities between sensors via knowledge representation learning technique. The proposed QK-GCN model is developed and verified using the annual field data from a GE 9F-class gas turbine. The proposed model achieves a lower overall error value and a narrower error distribution in all comparative studies. Embedding relation similarity as the GCN edge weight improves the root mean square error by 15.3 % for compressor efficiency degradation (DEC) and 33.5 % for flow rate degradation (DGC). Furthermore, the proposed QK-GCN model maintains strong generalization capability even with a 30 % training data size. The results indicate that the proposed QK-GCN model provides a reliable, generalizable, and interpretable solution for gas turbine GPA. The dual-knowledge method is promising for application to the performance prediction and fault diagnosis of other energy and power systems.

Exchange bias (EB) in two-dimensional van der Waals (vdW) ferromagnetic (FM)/antiferromagnetic (AFM) heterostructures holds great potential for advancing the applications of spintronic devices thanks to their defect-free and atomically flat interfaces. Normally, a field-cooling process is needed to either trigger or sustain the EB effect. Here we report a sizable EB effect in the Fe3GaTe2/CrSBr vdW heterostructure in a zero-field cooling process. Remarkably, an exceptionally large EB field (HEB) of 130.1 mT was achieved in Fe3GaTe2/CrSBr at 5 K, even though the spin configuration in Fe3GaTe2 and CrSBr is orthogonally arranged. Additionally, the HEB of Fe3GaTe2/CrSBr exhibits pronounced nonmonotonic and asymmetric dependence on the cooling field, with the maximum values appearing at intermediate field strength. The EB is effectively tuned by the thickness of the FM layer relative to that of the AFM layer, identifying this ratio as an additional important governing parameter. Our work suggests an unconventional mechanism of EB in vdW heterostructures, providing an innovative route for fabricating low-power and robust 2D spintronic devices.

SOFC-GT system is often combined with other power generation devices due to its exhaust gas still possessing high-quality waste heat, which enhances electricity generation. Currently, the emphasis of research is mainly on the conceptualization and thermodynamic assessment of the SOFC triple cycle system. In reality, appropriate low-temperature cycles can be introduced according to the characteristics of exhaust gas temperature distribution to achieve staged utilization of heat. This paper couples an ORC and a LNG system with the SOFC-GT-SRSCO2BC system, establishing a comprehensive energy system integrating power generation and carbon capture. The exergoeconomic performance under varying subsystem parameters was analyzed, followed by bi-objective optimization implemented through the Particle Swarm Optimization (PSO) method. The system optimal solution was determined using the Bidimensional Preference Analysis Linear Programming Decision Method (BPALPDM). The result indicates that incorporating two cycles for recovering waste heat enhances the maximum exergy efficiency of the SOFC-GT system by 13.28%. The LNG subsystem contributes just 4% to the cost of the system, improving efficiency and aiding in carbon capture. The entire system reaches an exergy efficiency of 70.57% at the ideal operating point, while the LCOE is 0.06335 $/kWh.

Digital Engineering Asset (DEA) refers to the digital artifacts, information and data involved in the lifecycle of engineering systems. It is challenging to maintain the interoperability and to ensure the security and immutability of DEAs during the exchange process among different stakeholders. Model-based systems engineering (MBSE) provides an efficient tool for improving cross-domain interoperability of DEAs. Distributed Ledger Technology (DLT) provides a potential solution to the security and immutability problem for DEA exchanges owing to its unique advantages such as tamper-resistance and decentralization. This study proposes a decentralized DEA exchange method by integrating the advantages of MBSE and DLT technologies. A definition of DEA is first presented from perspectives of digital engineering objects, development processes and system architectures. Based on this definition, the MBSE meta-model approach is used to formalize the DEAs. A conceptual DEA marketplace framework is then proposed based on a directed acyclic graph (DAG) structured DLT solution. A prototype of the proposed DEA marketplace is developed and a case study is conducted to verify its feasibility. The experiment results demonstrate that the proposed method can enable low-cost and high-efficiency DEA exchange.

The exhaust gas temperature (EGT) of the gas generator is a critical indicator for the health management system of a marine gas turbine engine. Therefore, EGT prediction can not only support predictive maintenance decision-making but also serves as a reliable virtual sensor for EGT measurement. However, the engine EGT exhibits strongly nonlinear coupling relationships with other gas path variables, which causes challenges for data-driven prediction. Graph neural networks (GNNs) are particularly effective in capturing the coupling relationships among gas path sensor variables. However, conventional static graph structures fail to characterize the varying coupling strengths under different operating conditions. In this study, a thermodynamic knowledge-driven graph attention network (TKD-GAT) method is proposed for accurate and robust EGT prediction. First, a physics-guided graph topology is constructed based on the gas turbine thermodynamic equations. Subsequently, a multi-head attention mechanism is introduced to generate edge weights that capture the varying thermodynamic coupling strengths under different operation conditions. The proposed model is evaluated on a real-world LM2500 gas turbine, which is widely used in modern propulsion systems of commercial and military ships. The ablation study confirms that the thermodynamic knowledge-driven graph topology and the attention mechanism-based edge weights are both necessary to enhance the EGT prediction performance. The TKD-GAT model shows the best performance with an RMSE of 0.446% and an R2 of 0.971 compared with state-of-the-art models. The paired t-test and effect size measurement (Cohen’s d) statistically confirm the significance of performance improvements. The statistical results from multiple independent experiments prove the stability of the TKD-GAT model. Additionally, the model achieves a competitive computational cost despite the integration of a physics-guided graph topology and attention mechanisms. Crucially, an interpretability analysis confirms that the learned attention weights adhere to thermodynamic principles under different operation conditions. The proposed TKD-GAT model provides an effective solution for EGT prediction in health management systems.

Abstract Gas turbine (GT) rapid performance prediction is a key task in health management and operation optimization. Existing methods for degradation parameters estimation either tend to become trapped in local optima or require a substantial computational effort. To tackle these problems, a stage-wise fast solution algorithm for GT degradation parameters based on a hybrid Levenberg-Marquardt (LM) and particle swarm optimization (PSO) method was proposed in this study. Firstly, a component-level simulation model was established, and the corresponding performance degradation parameters were defined to accurately characterize the health states of primary components. Secondly, a physics-guided stage-wise hybrid LM–PSO algorithm was proposed to balance solution accuracy and computational efficiency. Finally, comparative studies with other solution algorithms were conducted using both actual operational data and simulation data to demonstrate the effectiveness of the proposed method. The comparative results demonstrate that the proposed method significantly improves the solution speed while maintaining high accuracy, making it well-suited for practical GT degradation assessment and health management applications.

With the growing share of renewables, gas turbines (GTs) are increasingly operating under part-load conditions. Although existing models can accurately predict performance at both design-and off-design conditions, they often overlook part-load conditions. Besides, during full-load modeling, the mismatch between ideal operational data (with fully opened IGV) and actual operational data (with nearly fully opened IGV) may introduce additional modeling errors. Moreover, existing part-load performance prediction methods are typically implemented based on specific assumptions that only work under certain preconditions. To tackle these problems, a novel GT full-and part-load performance prediction model is proposed. First, the baseline model is established, serving as a solid foundation for subsequent correcting. Then, the IGV correction coefficients are calculated to quantify the baseline model's prediction bias caused by varying IGV positions. Subsequently, a multi-layer perceptron (MLP) enhanced with residual connection and feature engineering techniques is developed to accurately predict the IGV correction coefficients. Due to the small amount of input variables and training samples, feature generation and feature analysis methods are adopted to improve the prediction accuracy. Moreover, the residual connection is incorporated to partially mitigate the gradient vanishing phenomenon. Finally, the trained model is integrated into the physics model to calculate the corrected performance. The actual data adopted from various operating conditions are utilized to verify its efficiency and accuracy. The average prediction error is less than 1.59 %, indicating that the proposed method can accurately predict GT performance under both full- and part-load conditions, thereby providing a solid foundation for fault diagnosis.

Aero engines are complex coupled systems in which faults in one subsystem can propagate and affect the performance of others. Relying on single-source performance parameters is often insufficient for accurately assessing component degradation. Although multi-source fusion diagnosis methods, such as those based on Bayesian networks, have been widely applied, their diagnostic performance remains limited when prior knowledge is scarce. To address this challenge, this paper proposes a multi-source information fusion diagnosis method for aero engine fault detection based on Dempster–Shafer (D-S) evidence theory. Data from gas path and vibration subsystems are separately processed to extract fault features, and a decision-level fusion strategy is employed to achieve comprehensive diagnoses. A case study based on real operational data from a two-shaft aero engine demonstrates that the proposed method significantly improves diagnostic performance. Specifically, the Bayesian-network-based fusion method achieves a diagnostic confidence of 87.2% without prior knowledge and 91.2% with prior knowledge incorporated, whereas D-S evidence theory attains a higher fault confidence of 99.6% without requiring any prior information.

Lean-premixed combustion technologies have been widely adopted in advanced civil turbofan engine to reduce NOx emission. There exist hot streak (HS) and swirl simultaneously at the exit of lean-premixed combustor. Current paper presents a numerical investigation on the aerothermal performances of film cooled high pressure turbine (HPT) nozzle guide vane (NGV) subjected to HS and swirl. Current investigation was carried on the stage one film cooled NGV of GE-E3 HPT, which took into consideration the realistic clocking position of HPT NGV relative to the fuel injector in combustor. The effects of swirl orientations on the migration of HS and film coolant and the aerothermal performances on NGV surface were examined. Results demonstrates that, swirl and its induced incidence angle effect turn over some film coolant from pressure side (PS) to suction side (SS), or vice versa. Such effects also dominate the radial migration of film coolant. The redistributions of film coolant show significant effect on the film cooling effectiveness on NGV surface. Compared with no-swirl case, the greater film cooling effectiveness appears at the region where film coolant accumulates, and as expected, the smaller film cooling effectiveness arises at the region covered by less film coolant. As for heat transfer coefficient (HTC), swirl affects HTC on NGV surface mainly through redistributing film coolant and the hot fluid from HS. The migrations of film coolant and hot fluid keep step with each other on the NGV surfaces directly impinged by inlet HS and swirl, so that the HTC distributions on these surfaces are not significantly affected by swirl. The radial momentum of film coolant endowed by the film hole with radial incidence angle can partly offset the swirl's induced incidence angle effect, which reduces the variations in film cooling effectiveness and heat transfer coefficient due to swirl.

For aircraft engine predictive maintenance (PdM) programs, accurate remaining useful life (RUL) predictions can significantly reduce unscheduled maintenance downtime and ensure engine safety. To this end, we introduce a novel dual-channel degradation monitoring (DCDM) algorithm designed to minimize RUL prediction errors. Unlike traditional RUL prediction algorithms based on graph neural networks (GNNs), the proposed DCDM model extracts fault features from both node embeddings and changes in graph structures. This dual-channel approach allows for the fusion of fault information, improving the model's ability to extract features across different fault patterns for RUL prediction. During the graph structure learning (GSL) process, domain-specific knowledge and a dynamic graph learning algorithm are integrated to generate graphs, enhancing the interpretability of graph representation. In addition, by introducing node sparse encoding as a model input, the DCDM model's capability to discern critical features is significantly improved. The predictive performance of the DCDM model and the effectiveness of its individual components are validated using the C-MAPSS dataset. The results demonstrate the superior accuracy of the proposed method compared to existing approaches.

Gas path analysis methods impose strict requirements on the number of measurement points. The multioperating point method somewhat addresses this by using adjacent operating data, yet it relies on the assumption of uniform degradation levels and may not converge when applied to data from before and after abrupt faults and is time-intensive. To overcome these limitations, this paper introduces a feature fusion cascade neural network to diagnose both gradual and abrupt faults. Firstly, a component-level model was used to generate fault data. The multi-operating point method is then used to support fault diagnosis and provide training and testing data for the proposed method. Lastly, a novel cascade neural network architecture was proposed and the feature fusion methods including feature analysis and feature generation were adopted and incorporated into the architecture to improve the fault diagnosis accuracy. Compared with other data-driven models, it exhibits higher prediction accuracy, notably in SEC. The proportion of diagnosis error under 2 x 10- 4 among all diagnosed values for gradual and abrupt faults diagnosis is 52.56 % and 51.18 %, respectively. These findings support the high accuracy and efficiency of the proposed method.

Data-driven models utilizing machine learning algorithms provide an effective approach for predicting power in heavy-duty gas turbines, extracting valuable insights from large-scale operational datasets. However, global unified models often struggle to meet the accuracy requirements of all data when dealing with complex and variable operating conditions, leading to limited prediction accuracy for local conditions. To address this problem, a clustering modification method is introduced to develop a novel power prediction model for heavy-duty gas turbines. In this study, the Support Vector Regression (SVR) prediction model is combined with a k-means clustering modification model, enabling the model to adapt to different operational conditions. Operational data from an E-class gas turbine are carefully preprocessed, including filtering, noise reduction, and steady-state selection, to enhance data quality. Then, the k-means algorithm is employed to classify operational conditions, with tailored modification models trained for each category. These modification models refine predictions to accommodate variations in specific operating states. Experimental results demonstrate that the composite model achieves a 32.66% reduction in MAPE and an increase in R2 to 0.9982 compared to single-model approaches. The analysis further highlights that training the model with 70% of the annual data achieves optimal prediction accuracy and stability. Additionally, the model significantly reduces high-error occurrences, with 75% of predictions having errors below 0.2946%. This method improves the precision and adaptability of power prediction for gas turbines, providing a practical framework that enhances the reliability of real-world applications and supports the advancement of data-driven energy systems.

Gas turbine systems operate in high temperature and high pressure conditions, which easily lead to performance degradation or various serious failures. The reliability is one of the most important attributes to monitor the gas turbine safety. Failure Mode, Effects, and Criticality Analysis (FMECA) is a well-known approach to systematically analyze the gas turbine reliability. However, traditional document-centric implementation of FMECA requires manual recording of fault related information. Such implementation approach is prone to human error and tedious. As an evolving approach, Model-based Systems Engineering (MBSE) has been applied in many industries to support system design, analysis and management. In this study, an automatic FMECA analysis approach for gas turbine systems is proposed based on MBSE. The FMECA architecture model is established according to the SysML diagrams of internal block diagram and block definition diagram. The topology between FMECA related concepts is characterized through the constructed architecture model. Then, the Design Structure Matrices is used to realize the risk analysis automatically. A case study of a real gas turbine is conducted to demonstrate the effectiveness of the proposed approach. The case results suggest that proposed approach can significantly enhance the efficiency of FMECA analysis.

The fabrication of MnBi2Te4 family of materials has been demonstrated that intercalating magnetic MnTe layer into topological insulator Bi2Te3 is an efficient way to explore magnetic topological insulator. A natural question is that if multiple MnTe layers can be intercalated into Bi2Te3, meanwhile if the topological properties are well preserved. Based on first-principles calculations, we first confirm that multiple MnTe layers can be inserted into Bi2Te3 by forming antiferromagnetic MnnBi2Te3+n (n = 1–5) crystals, where each system is energetically stable and can be readily exfoliated down to a monolayer. Specifically, the structural phase transition from ABC to ABAC stacking happens when n ≥ 2. Further electronic band structure calculations and topological identification reveal that the magnetic MnnBi2Te3+n undergoes a topologically nontrivial to trivial phase transition when n > 2, which is attributed to that multiple MnTe layers destroy the topological property of Bi2Te3. In addition to applying compressed strain, the topological states can be restored by intercalating Bi2Te3 layers by forming (MnTe)n(Bi2Te3)m heterostructures. Most strikingly, it is found that the topological properties of (MnTe)n(Bi2Te3)m are universally determined by the ratio of n/m, once n/m ≤ 2.8, the desired topological states are recovered. Our notable findings enrich the family of magnetic topological insulator, providing opportunities to explore more intriguing topological phenomena.

The whorls of localized moments in chiral magnetic structures, such as skyrmions, lead to a quantized topological charge, which may make them useful as next-generation information bits. So far, the most reliable way to detect the existence of skyrmions is by using the topological Hall effect, which stems from electron scattering by the emergent magnetic field manifesting the topological charge. Here we employ two-dimensional magnets to establish a magneto-optical hallmark of skyrmions, which we call the topological Kerr effect, using the recently discovered ferromagnet CrVI6 as a material platform. The Kerr angle hysteresis loop of this non-centrosymmetric system exhibits two antisymmetric bumps that are absent in the centrosymmetric CrI3 and VI3. We develop a minimal model to further identify the bumps as direct manifestations of the topological charge, thereby providing a magneto-optical fingerprint of skyrmions with broader applicability. The ferromagnet CrVI6 serves as a material platform to demonstrate the topological Kerr effect in two-dimensional magnets. This can be used to identify skyrmions by magneto-optical means.

A SOFC temperature control system based on adaptive multimodel predictive control (MMPC) method is designed for a solid oxide fuel cell-gas turbine (SOFC-GT) hybrid system with anode and cathode ejectors. Two multi-input and multi-output MPCs (under 100% and 90% load) are designed to control the anode and cathode inlet temperatures. The accuracy of the identified linear models are both more than 95%. The control performance of the designed MMPC is compared with a single MPC and traditional PI. The comparison results demonstrate that the proposed MMPC is most effective and competitive in SOFC thermal management. During the load following, the controller overshoot is less than 1.19K. The settling time is about 2000s, and the integral of time-weighted absolute error is less than 472.

Hydrogen can alleviate the increasing environmental pollution and has good development prospects in power generation due to its high calorific value and low environmental impact. The previously designed hydrogen-fired combined cycle ignored water recycling, which led to an inefficient application of hydrogen and the wastage of water. This paper proposes the concept of a hydrogen-fired combined cycle with water recovery to reuse the condensed water as an industrial heat supply. It was applied to an F-class combined cycle power plant. The results demonstrate that the efficiency of hydrogen-fired combined cycles with and without water recovery increased by 1.92% and 1.35%, respectively, compared to that of the natural-gas-fired combined cycle under full working conditions. In addition, an economic comparison of the three cycles was conducted. The levelized cost of energy of the hydrogen-fired combined cycle with water recovery will be 52.22% lower than that of the natural-gas-fired combined cycle in 2050. This comparative study suggested that water recovery supplementation could improve the gas turbine efficiency. The proposed hydrogen-fired combined cycle with water recovery would provide both environmental and economic benefits.