Search

This paper proposes a fault diagnosis method for rotating machinery that integrates transfer learning with the ConvNeXt model (TL-CoCNN), addressing challenges such as small sample sizes and varying operating conditions. To meet the input requirements of the model while minimizing feature loss, an alternative approach to visualizing vibration data is introduced. Specifically, RGB images are synthesized from time-domain, frequency-domain, and time-frequency domain representations of the original signal, which are subsequently used as the input dataset. The fault diagnosis process leverages a pre-trained ConvNeXt model, initially trained on the ImageNet dataset, and fine-tunes its parameters using the synthesized RGB images to perform the fault classification task. Experimental results demonstrate that this data visualization method extracts more fault-related information compared to traditional time-domain and frequency-domain techniques, without the need to augment the sample size. The TL-CoCNN model achieves superior recognition accuracy when evaluated in terms of training time and model size across multiple test datasets. As an end-to-end fault diagnosis system, TL-CoCNN significantly enhances the feature representation capability of complex signals, showing promising potential for practical applications in fault detection and diagnosis. [ABSTRACT FROM AUTHOR]

In this article, an innovative online adaptive high-efficiency control strategy is proposed to improve the power generation efficiency of a marine micro gas turbine under partial load. Firstly, a mathematical model of the micro-gas turbine is established, and a control strategy consisting of an on-board prediction model and an online update model is proposed. To evaluate the performance changes of the gas turbine, we applied deep learning techniques to enhance the extreme learning machine (ELM) algorithm, resulting in the development of a high-precision, high-real-time deep extreme learning machine (DL_ELM) prediction model. This model effectively monitors changes in the gas turbine's performance. Furthermore, an online time-series deep extreme learning machine with a dynamic forgetting factor (DFF_DL_OSELM) model is designed to achieve the real-time tracking of performance variations. When the DL_ELM model detects a gas turbine's performance change, a particle swarm optimization (PSO) algorithm is employed to iteratively calculate the DFF_DL_OSELM model, determining the optimal speed control scheme to ensure the gas turbine operates at maximum efficiency. To validate the superiority of the proposed control strategy, a comparison is made with traditional high-efficiency control strategies based on polynomial fitting and BP neural networks. The results demonstrate that although all three strategies can achieve efficient operation under constant conditions, traditional strategies fail to identify and adjust to performance changes in real time, leading to decreased control performance and potential engine damage as engine characteristics degrade. In contrast, the proposed online adaptive control strategy dynamically adjusts the speed control plan based on performance degradation, ensuring that the gas turbine operates efficiently while keeping the turbine inlet and exhaust temperatures within safe limits. [ABSTRACT FROM AUTHOR]

To ensure the continuous high-efficiency operation of fuel cell systems, it is essential to perform real-time estimation of the maximum efficiency point and maximum power point for multi-stack fuel cell systems. The region between these two power points is commonly referred to as the "high-efficiency operating region." Initially, a transformation of the general expression for hydrogen consumption in multi-stack fuel cell systems is conducted to obtain an algebraic expression for the efficiency curve of multi-stack fuel cells. Utilizing a polynomial differentiation approach, the parameter equation for the maximum system efficiency is computed. Subsequently, a reverse deduction is carried out using the maximum efficiency and its corresponding power of underperforming subsystems to enhance the maximum efficiency of multi-stack fuel cell systems.Furthermore, an equivalent hydrogen consumption minimization method is introduced for real-time optimization of hybrid energy systems. The state machine control method serves as an auxiliary strategy, imposing the high-efficiency operating region as a boundary constraint for the equivalent hydrogen consumption minimization strategy's results. This ensures that the multi-stack fuel cell system operates as much as possible within the high-efficiency operating region.Through simulation validation using MATLAB/Simulink, the proposed approach comprehensively leverages the advantages of the state machine and equivalent hydrogen consumption. This approach enables effective identification of the high-efficiency operating region of fuel cells, while concurrently enhancing the operational range efficiency of the system, reducing hydrogen consumption, and elevating system stability. [ABSTRACT FROM AUTHOR]

Titanium alloys are widely used in the manufacture of gas turbines' compressor blades. Elucidating their mechanical behavior and strength under damaged conditions is the key to evaluating the equipment's reliability. However, the conventional Johnson-Cook (J-C) constitutive model has limitations in describing the dynamic response of titanium alloy materials under the impact of a high strain rate. In order to solve this problem, the mechanical behavior of a TC4 titanium alloy under high strain rate and different temperature conditions was analyzed by combining experiments and numerical simulations. In this study, the parameters of the J-C model were analyzed in detail, and an improved J-C constitutive model is proposed, based on the new mechanism of the strain rate strengthening effect and the temperature softening effect, which improves the accuracy of the description of strain sensitivity and temperature dependence. Finally, the VUMAT subroutine of ABAQUS software was used for numerical simulation, and the predictive ability of the improved model was verified. The simulation results showed that the maximum prediction error of the traditional J-C model was 23.6%, while the maximum error of the improved model was reduced to 5.6%. This indicates that the improved J-C constitutive model can more accurately predict the mechanical response of a titanium alloy under an impact load and provides a theoretical basis for the study of the mechanical properties of titanium alloy blades under subsequent conditions of foreign object damage. [ABSTRACT FROM AUTHOR]

To address the challenge of measuring the dynamic characteristic parameters of the gas turbine rotor system affected by hull excitation, a vibration transmission model integrating a ship model slice, test data, and a three-dimensional entity is proposed, based on the two-dimensional slice theory, scaled ship model, and finite element model of the turbine rotor system. The transient dynamic responses of the front and rear bearing points were calculated and analyzed. Vibration response tests with significant wave heights of 0.5 m, 1.25 m, 2.5 m, and 4 m were carried out in the towing tank of the ship model to obtain the dynamic characteristic parameters of the deck position. Techniques including wavelet denoising, Fast Fourier Transform (FFT), and signal resampling were employed to filter out and reconstruct high-frequency noise, overcoming the technical challenges of a high sampling frequency and a low computational efficiency. The experimental data and simulation results were compared and analyzed, validating the accuracy of the vibration transmission model of the turbine rotor system with data and entity integration. By comparing the vibration signal values in the X and Z directions at the front and rear bearing points after vibration transmission, it is evident that the effective values of the vibration signals at the front bearing point are 0.03% to 0.1% greater than those at the rear bearing point. This model provides a theoretical basis and reference for the design of the gas turbine rotor system. [ABSTRACT FROM AUTHOR]

Vibration analysis is crucial for studying rotor dynamics. The gas turbine rotor system is subjected to complex alternating loads during navigation, resulting in vibrations transmitted to the bearings that alter the system's dynamic characteristics. Based on the similarity law of the wave resistance test, a hull model was established. Beam sea and head sea tests were conducted in the towing pool to measure the acceleration response at the key positions. A finite element model of the turbine rotor system was established, and the test data were imported into the model after wavelet noise reduction and resampling to calculate the vibration response at the front and rear bearing points. The vibration responses transmitted to different locations and directions caused by beam sea and head sea conditions were analyzed. A comparison and analysis were conducted on the acceleration responses in various locations and directions under beam sea or head sea conditions. The equivalent von Mises stress distribution of the gas turbine rotor system under beam sea and head sea loads was obtained. The vibration transfer model was verified for accuracy and can be used to quickly analyze the vibration response of bearings under wave load transfer. This study provides a theoretical basis and reference for enhancing the stability of the gas turbine rotor system. [ABSTRACT FROM AUTHOR]

This paper takes the support structure of a low-pressure rotor system of a marine gas turbine as the background.Considering the heel angle of the rotor system,a dynamic model of the rotor supported by asymmetric rolling bearing is established. The variable-step fourth-order Runge-Kutta method was used to solve the system’ s dynamic equations,and the nonlinear theory were used to analyze the simulation results. The study shows that the heel angle has no effect on the stability of the system at low rotational speeds. At high rotational speed,the heel angle mainly affects the 1/2 harmonic of the system,and the system has a single chaotic island or two chaotic islands. When the left or right tilt angle of the rotor system is the same,the heel angle has the same effect on the nonlinear dynamics of the system. The research results can provide a theoretical basis for the design,monitoring and fault diagnosis of the rotor system.

The performance of ammonia hydrogen fuel cells (AHFCs) with the same activated area but different aspect ratio is quite different. In this paper, based on the diffusion characteristics of hydrogen nitrogen mixture, five different aspect ratios of AHFCs were designed, and the influence of aspect ratio on the performance of AHFCs was studied by combining experiment and simulation analysis. The results reveal that the power of AHFCs reaches the maximum when the aspect ratio is 1, that is, the shortest hypotenuse of the plate. With the extension of the hypotenuse, the output power decreases, and the output power decreases by 19.3% when the aspect ratio is 4. Due to the strong ability of hydrogen diffusion, when the aspect ratio of AHFCs is 4, the proportion of hydrogen and oxygen starvation area at the interface reaches the maximum and minimum, which are 11.8% and 0.3% respectively, which is contrary to the aspect ratio of 0.25. A larger aspect ratio is conducive to the uniform distribution of gas-water-heat. Although the AHFCs with aspect ratio of 0.64 has advantages in power generation efficiency and fuel utilization, the pressure drop loss increases by 43.7%, and the pressure drop loss with aspect ratio of 0.25 increases by 141.2%. • Experimental data of AHFCs with different aspect ratios are obtained. • Power generation efficiency of AHFCs with different aspect ratios is compared. • AHFC with the shortest hypotenuse has the largest output power. • CV is used to characterize the uniformity of gas-heat-water distribution in AHFC. [ABSTRACT FROM AUTHOR]

In mini-grids and marine-isolated grids, power generation gas turbines are subjected to rapid start-up, shutdown, and acceleration/deceleration. This sudden load change can pose a significant impact on the power grid, severely affecting the operational characteristics of gas turbines. To understand the dynamic characteristics of the gas turbine in the transitional processes, this testing takes twin-shaft medium-sized power generation gas turbines as the test object, and goes through the process of startup, acceleration, deceleration, acceleration, shutdown in one hour, and repeats this process 40 times continuously. With fuel flow as the control parameter and power turbine outlet temperature and high-pressure turbine speed as the controlled parameters, the parameter response rate of the gas turbine under various transition processes is analyzed and the effect of thermal inertia on the gas turbine mass temperature as well as speed is studied. Research findings: During the transition processes, the gas temperature exhibited an axial gradient distribution in the channel. In both the acceleration and deceleration processes, the working fluid temperature gradually decreased along the flow direction. And thermal inertia posed different extents of impact on the dynamic characteristics of the gas turbine under different transitional processes. In the same transition process, the impacts of thermal inertia on the response speeds of temperature and rotational speed varied. The results of this study help to more accurately predict the operating state of the gas turbine during the transition process and lay the foundation for the dynamic simulation model of the non-adiabatic gas turbine. [ABSTRACT FROM AUTHOR]

Due to high greenhouse gas emissions, countries worldwide are stepping up their emission reduction efforts, and the global demand for new, carbon-free fuels is growing. Ammonia (NH3) fuels are popular due to their high production volume, high energy efficiency, ease of storage and transportation, and increased application in power equipment. However, their physical characteristics (e.g., unstable combustion, slow flame speed, and difficult ignition) limit their use in power equipment. Based on the structural properties of the power equipment, NH3 fuel application and emissions characteristics were analyzed in detail. Combustion of NH3 fuels and reduction measures for NOx emissions (spark plug ignition, compression ignition, and gas turbines) were analyzed from various aspects of operating conditions (e.g., mixed fuel, fuel-to-exhaust ratio, and equivalence ratio), structure and strategy (e.g., number of spark plugs, compression ratio (CR), fuel injection, and ignition mode), and auxiliary combustion techniques (e.g., preheating, humidification, exhaust gas recirculation, and secondary air supply). The performance of various NH3 fuel cell (FC) types was analyzed, with a focus on the maximum power achievable for different electrolyte systems. Additionally, the application and NOx emissions of indirect NH3 FCs were evaluated under flame and catalytic combustion conditions. The system efficiency of providing heat sources by burning pure NH3, anode tail gas, and NH3 decomposition gas was also compared. Based on a comprehensive literature review, the key factors influencing the performance and emissions of NH3-powered equipment were identified. The challenges and limitations of NH3-powered equipment were summarized, and potential strategies for improving efficiency and reducing emissions were proposed. These findings provide valuable insights for the future development and application of NH3 FCs. [ABSTRACT FROM AUTHOR]

In this paper, based on the combination of comprehensive sampling and one-dimensional convolutional neural network, a bearing fault intelligent diagnosis technique is proposed for the classification of rolling bearing vibration data. At first, the fault data set is expanded by ADASYN method. Then, the data is cleaned up by Tomek link under sampling technique, the risk of overfitting caused by overlap of different classes is reduced and the data of different categories is more apparent, and finally the normal data set and fault data set after comprehensive sampling are classified by one-dimensional convolutional neural network algorithm. Compared with random forests and support vector machines, the results show that the method has a high accuracy in identifying classifications and can effectively solve the classification problem of unbalanced bearing data.

The review was done in order to make clear the problem of the hot corrosion of the Thermal Barrier Coatings(TBCs)during gas turbine serving. This paper summarizes the factors resulting from the hot corrosion of TBCs during turbine service and classifies methods for enhancing the corrosive resistance of TBCs. A prospective methodology for improving corrosion resistance is also formulated. The main types of corrosion coating include phase reaction, oxidizing of the bond coating, salt-fog corrosion, CMAS corrosion and fuel impurity corrosion. So far, methods for improving the corrosion resistance of TBCs include developing new coating materials, anticorrosive treatment on the surface of TBCs, modifying the stacking configuration and improving the cleansing functions of the gas turbines. In the future, developing new materials with excellent performance will still be the main direction for boosting the improvement of the hot corrosion resistance of TBCs. Simultaneously, improving the tacking configuration and nanotechnology of TBC coatings are potential approaches for improving corrosion resistance. With the development of a Ceramic Matrix Composite (CMC), the focus of the hot corrosion of TBCs may turn to that of Environmental Barrier Coatings (EBCs).

Residual stress is an important parameter to evaluate and predict the interfacial peeling and failure of a thermal barrier coating (TBC) system in the thermal cycle. Considering a non-uniform temperature field, an analytical model for predicting residual stress of the TBC-film cooling system was established. Based on the double-layer TBC-film cooling system, the analytical solutions were validated by comparing with the previous results. Then, the four-layer system, including the superalloy substrate (SUB), the metallic bond coat (BC), the thermally grown oxide (TGO), and the ceramic top coat (TC), was analyzed. The distribution of residual stress was discontinuous at the interface among the layers. The maximum peeling moment occurred at the BC/SUB interface, and the maximum shear stress occurred at the TGO/BC interface. Therefore, the probabilities, of which opening edge cracks (mode I) appeared at the BC/SUB interface and shearing edge cracks (mode II) appeared at the TGO/BC interface, were higher than that at other interfaces. The deflection and bending direction of the system were affected by the temperature difference and coefficient of thermal expansions. The large thickness of the coating could effectively improve the thermal insulation capacity of the system, but the peeling moment and shear stress also increased. Therefore, preventing the thermal growth of the TGO layer and reducing the thickness of the TC layer can improve the stability of the system and extend service lifetime of the system while ensuring that the cooling requirements for hot section components are met. [ABSTRACT FROM AUTHOR]