Search

Steam turbine blades may crack, break, or suffer other failures due to high temperatures, high pressures, and high-speed rotation, which seriously threatens the safety and reliability of the equipment. The signal characteristics of different fault types are slightly different, making it difficult to accurately classify the faults of rotating blades directly through vibration signals. This method combines a one-dimensional convolutional neural network (1DCNN) and a channel attention mechanism (CAM). 1DCNN can effectively extract local features of time series data, while CAM assigns different weights to each channel to highlight key features. To further enhance the efficacy of feature extraction and classification accuracy, a projection head is introduced in this paper to systematically map all sample features into a normalized space, thereby improving the model's capacity to distinguish between distinct fault types. Finally, through the optimization of a supervised contrastive learning (SCL) strategy, the model can better capture the subtle differences between different fault types. Experimental results show that the proposed method has an accuracy of 99.61%, 97.48%, and 96.22% in the classification task of multiple crack fault types at three speeds, which is significantly better than Multilayer Perceptron (MLP), Residual Network (ResNet), Momentum Contrast (MoCo), and Transformer methods. [ABSTRACT FROM AUTHOR]

In steam turbines, the operation of turbine blades can lead to vibrations, which may result in turbine accidents. Traditional calculation methods are difficult to use due to the complex structure of the torsional blade, and simulation analysis can be time-consuming. Therefore, it is necessary to use the equivalent model instead. Most of the previous studies have simplified the straight blade to a cantilever beam model without considering the shear effect, and used harmonic response analysis to study the vibration of the blade for simulation. The blade body of the torsional blade can be regarded as a torsional variable section beam, and the Timoshenko beam model is the same as the variable section beam. Therefore, the Timoshenko beam model can be regarded as the equivalent of the sheathed torsional blade. Considering the complexity of the calculation of the self-shrouded torsional blade model. According to the theory of Timoshenko beam, the model of the shrouded blade of Timoshenko beam with different torsion angles was established. The first sixth-order modes of the blade model and the vibration response of the adjacent blades under forced vibration are compared, respectively. The results show that the results obtained from the Timoshenko beam blade model considering the torsion angle are closer to the torsional blade model. In addition, the blade vibration response of the same model under different excitation force amplitudes was further compared, and the results showed that the vibration amplitude, acceleration and velocity of the shrouded blade increased with the increase of the excitation force amplitude. However, the magnitude of the inter-shroud collision force and the number of collisions between adjacent blade shrouds are non-linearly related to the amplitude of the excitation force. [ABSTRACT FROM AUTHOR]

Digital twin (DT) is a key enabling technology to realize cyber-physical system (CPS), which can truly perceive, map, and predict the operating state of physical entity. Through analyzing the generality of the testing process of large and complex equipment components (LCEC), a five-dimensional framework of DT-based test process management (DTTPM) is proposed, which comprises physical layer, network layer, data layer, model layer, and service layer. In order to realize the visualization and enhance the controllability, security, and information transparency of LCEC in the testing process, three key technologies are elaborated in detail as follows: (1) the construction of the twin semantic model of the testing process in model layer, (2) the synchronization method of twin model and physical entity based on real-time data, (3) and human–computer interaction–based visual monitoring of authenticity and safety coordination. Through the case of the vibration test for the crowned blade of a steam turbine by blade tip-timing measurement, the feasibility and flexibility of DTTPM are demonstrated. [ABSTRACT FROM AUTHOR]

In steam turbine, turbine blades are prone to vibrate during operation, resulting in steam turbine accidents. The most common method for reducing the vibration of steam turbine blades is to design an integral shroud for blade which is termed as integral shroud blade. Most previous studies simplified straight integral blades into cantilever beam and used harmonic response analysis method to simulate the vibration response of blades. This method is suitable for simulating straight blade vibration under harmonic force conditions. For twisted blades, accurate results are hard to acquire and the specific collision process cannot be simulated. In order to observe the collision process on a microscopic scale and explore its collision damping mechanism, this study evaluated the collision process of twisted blades with different integral shroud clearance distance based on LS-DYNA software. The collision process for a two-blade system and a three-blade system with integral shroud clearance distance from 0.1 mm to 0.5 mm has been simulated. The results indicated that integral shroud clearance distance have opposite vibration damping effect when the blade under the condition of forced vibration and free vibration. For the two-blade system, the optimal integral shroud clearance distance is 0.4 mm for forced vibration condition and 0.1 mm for free vibration condition. For the three-blade system, the optimal integral shroud clearance distance is 0.1 mm for forced vibration condition and 0.5 mm for free vibration condition. [ABSTRACT FROM AUTHOR]