The health of the hydroelectric generator determines the safe, stable, and reliable operation of the hydropower station. In order to keep the hydroelectric generator in a better state of health and avoid accidents, it is crucial to detect its faults. In recent years, fault detection methods based on sound and vibration signals have gradually become research hotspots due to their high sensitivity, achievable continuous dynamic monitoring, and easy adaptation to complex environments. Therefore, this paper is a supplement to the existing state monitoring and fault diagnosis system of the hydroelectric generator; it divides the hydroelectric generator into two significant parts: hydro-generator and hydro-turbine, and summarizes the research and application of fault detect technology based on sound signal vibration in hydroelectric generator and introduces some new technology developments in recent years, and puts forward the existing problems in the current research and future development directions, and it is expected to provides some reference for the research on fault diagnosis of the hydroelectric generator.

To enhance the operational efficiency and fault detection accuracy of hydroelectric units, this paper proposes a parallel convolutional neural network model that integrates the Gramian angular summation field (GASF) with an Improved coati optimization algorithm–parallel convolutional neural network (ICOA-PCNN). Additionally, to further improve the model’s accuracy in fault identification, a multi-head self-attention mechanism (MSA) and support vector machine (SVM) are introduced for a secondary optimization of the model. Initially, the GASF technique converts one-dimensional time series signals into two-dimensional images, and a COA-CNN dual-branch model is established for feature extraction. To address the issues of uneven population distribution and susceptibility to local optima in the COA algorithm, various optimization strategies are implemented to improve its global search capability. Experimental results indicate that the accuracy of this model reaches 100%, significantly surpassing other nonoptimized models. This research provides a valuable addition to fault diagnosis technology for modern hydroelectric units.

Diagnosing the vibration signals of hydropower units is crucial for safe and stable operation. This paper proposes a fault diagnosis method for hydropower units based on Gramian Angular Summation Fields (GASF) and parallel convolutional neural networks-gated recurrent unit-multi-headed self-attention (CNN-GRU-MSA). The original data forms a double branch, and the first branch selects the original timing signal for feature extraction using GRU. The second branch converts the timing signal into a 2D image using GASF for feature extraction using CNN, and the merged signal is enhanced with MSA for feature values. The experimental results show that the accuracy of the method reaches 97.2%. In order to explore the generalization and practicability of the proposed model, the public dataset of Jiangnan University is introduced for re-analysis. The diagnostic result of 600 rpm is 98.5%, and the diagnostic result of 800 rpm and 1000 rpm is 100%, significantly better than the other comparative models. This study can be valuable to the hydropower unit’s fault diagnosis methods.

Studies show that sediment erosion is one of the main factors attributing to hydraulic turbine failure. The present paper represents an investigation into acoustic vibration signals generated by the water flow impacting the hydraulic turbine runner under three different operating conditions. Collected signals were denoised using the ICEEMDAN-wavelet threshold method, and then the spectral characteristics and sample entropy characteristics of the signals for the three operating conditions were analyzed. The results show that when clean water flows through the hydraulic turbine, the sample entropy reaches its smallest values and the dominant frequency component in the spectrogram is 59.39 Hz. When transitioning from clean water to the flood flow containing 2–4 mm sediment particles, the sample entropy is increasing and a high-frequency component higher than 59.39 Hz becomes the prominent frequency of the spectrogram. Meanwhile, the formation of high-frequency components increases with the sand-containing particle size. Based on the spectral characteristics and sample entropy characteristics of the acoustic vibration signals under different operating conditions, it can provide a reference for the sand avoidance operation of the hydraulic turbine during flood season. In addition, it provides a supplement to the existing hydraulic turbine condition’s monitoring systems and a new avenue for subsequent research on early warning of hydraulic turbine failure.

The power oscillation induced by pressure fluctuation in the draft tube of the hydraulic turbine is one of the limiting factors preventing the Francis turbine from operating in the vibration zone. At the present power grid with a high proportion of renewable energy resources, we try to improve the load regulation ability of the hydropower units by extending the stable operation zone to the vibration zone. By the mathematical modelling of pressure fluctuation, this paper gives an analytical expression of the power oscillation. We derive the extended Hamiltonian model of the hydropower unit where power oscillation is external excitation. Secondly, the damping injection method introduces some desired interconnection and damping matrices as the Hamiltonian damping factor into the additional damping control. Finally, through theoretical analysis and experimental simulation, this research discusses the resonance characteristics of pressure fluctuation and power oscillation, the equivalent analysis between the damping factor and equivalent damping coefficient, and the control design of vibration zone crossing during the start-up. Simulation results show that when r₂₅ = 1.3, the minimum power oscillation amplitude is 0.5466, which is equivalent to an increase in D by 20. The maximum oscillation amplitude decreases by 4.6%, and the operation limited zone is reduced by 10.1%. The proposed additional damping control can effectively suppress the power oscillation and expand the regulation range.

The feature extraction and state recognition of vibration signals of hydroelectric generating units are of great significance for effectively ensuring the safety and lifespan of unit operation. This article proposes a fault diagnosis method for hydroelectr ic units using wavelet threshold algorithm preprocessing and improved (Gravitational Search Algorithm) GSA optimizer optimized (Radial Basis Function) RBF neural network. Using the (Wavelet Transform) WT algorithm for denoising preprocessing of the origina l signal, the processed signal is decomposed using (Complete Ensemble Empirical Mode Decomposition With Adaptive Noise) CEEMDAN, and IMF components are selected using the correlation coefficient method for sample entropy calculation. Finally, a GSA RBF fault diagnosis model for hydroelectric generating units is established. To improve convergence speed and accuracy, Lévy flight and chaotic sequence optimization GSA are introduced to reestablish the LCGSA RBF diagnosis model. The experimental results show th at the accuracy of this method reaches 100%, which is superior to other comparative models. Compared with CEEMDAN RBF, the accuracy of this method has been improved by 12.5%, and the diagnostic results have the highest consistency with the actual results. This study is of great significance for promoting the safe and reliable operation of hydroelectric units.