As a clean energy with mature technology, hydropower has been widely applied in industry. The hydraulic turbine unit plays an important role in hydropower station. Since the fault of turbine unit will affect the normal operation of the whole hydropower station, this paper proposes a universal, fast and memory-efficient method trend for time-series prediction of hydraulic turbines. The proposed method adopts the expressive power of deep neural networks and the time characteristics of sequence-to-sequence structure (parallel convolution and recurrent neural network) to make time-series prediction. It also uses convolutional quantile loss and memory network to predict future extreme events. The experimental results show that the proposed method is fast, robust and accurate. It can reduce at least 34% in mean square error and 33% in convergence speed compared with the existing methods.

No abstract

The oil in hydropower station catchment wells is a source of water pollution which can cause the downstream river to become polluted. Timely detection of oil can effectively prevent the expansion of oil leakage and has important significance for protecting water sources. However, the poor environment and insufficient light on the water surface of catchment wells make oil pollution detection difficult, and the real-time performance is poor. To address these problems, this paper proposes a catchment well oil detection method based on the global relation-aware attention mechanism. By embedding the global relation-aware attention mechanism in the backbone network of Yolov5s, the main features of oil are highlighted and the minor information is suppressed at the spatial and channel levels, improving the detection accuracy. Additionally, to address the problem of partial loss of detail information in the dataset caused by the harsh environment of the catchment wells, such as dim light and limited area, single-scale retinex histogram equalization is used to improve the grayscale and contrast of the oil images, enhancing the details of the dataset images and suppressing the noise. The experimental results show that the accuracy of the proposed method achieves 94.1% and 89% in detecting engine oil and turbine oil pollution, respectively. Compared with the Yolov5s, Faster R-CNN, SSD, and FSSD detection algorithms, our method effectively reduces the problems of missing and false detection, and has certain reference significance for the detection of oil pollution on the water surface of catchment wells.

In order to make the grid-connected composite device (GCCD) controller meet the requirements of different operating modes and complex working conditions of power grid, this paper proposes to introduce sliding mode control (SMC) into GCCD controller. Firstly, the mathematical model of MMC converter is established, and the sliding mode controller is designed based on the SMC principle. Then, aiming at the problems of complex controller structure and difficult parameter tuning in multiple modes of the GCCD, this paper proposes a controller parameter optimization method based on improved Month Flame optimization (IMFO) algorithm. This method improves the MFO algorithm by introducing good point set (GPS) initialization and Levy flight strategy, which accelerates the convergence speed of the algorithm while avoiding falling into local optimization, and realizes the optimization of converter controller parameters. Under a variety of standard test functions, the advantages of the proposed IMFO algorithm are verified by comparing it with the traditional algorithm. Finally, in order to realize the automatic tuning of control parameters, the Python–PSCAD joint simulation method is studied and implemented. Taking the comprehensive integral of time and absolute error (CITAE) index as the objective function, the parameters of the sliding mode controller are optimized. The simulation results show that the controller parameters optimized by the IMFO algorithm can make the GCCD have better dynamic performance.

With the rapid development of artificial intelligence, machine vision and other information technologies in the construction of smart power plants, the requirements of power plants for the state monitoring of hydro-generator units (HGU) are becoming higher and higher. Based on this, this paper applies YOLOv5 to the state monitoring scenario of HGU, and proposes a method for rotor speed measurement (RSM) and operating state identification (OSI) of HGUs based on the YOLOv5. The proposed method is applied to the actual RSM and OSI of HGUs. The experimental results show that the Precision and Recall of the proposed method for rotor image are 99.5% and 100%, respectively. Compared with the traditional methods, the online image monitoring based on machine vision not only realizes high-precision RSM and the real-time and accurate determination of operating states, but also realizes video image monitoring of the rotor, the operation trend prediction of the rotor and the early warning of abnormal operating states, so that staff can find the hidden dangers in time and ensure the safe operation of the HGU.