With the accelerated urbanization, rising grid integration of new energy, and widespread application of flexible regulation resources represented by mobile energy storage (MES), load fluctuations in regional power systems have grown increasingly complex. Accurate electric load forecasting is the core technology to ensure stable grid operation, optimize energy allocation, and enable efficient MES dispatch. However, regional load is jointly driven by meteorological conditions, time cycles, user behavior, and MES charging-discharging behaviors, featuring strong nonlinearity, multi-scale fluctuations, and coexisting short- and long-term temporal dependencies. MES dynamic access further amplifies load uncertainty, posing severe challenges to traditional forecasting methods. The proposed model has good engineering adaptability in park grid scenarios with MES access, and can provide a reusable technical framework for load forecasting in similar regional power grids.
In order to explore the influence of earth resistivity and coast dip angle on coastal effect in geomagnetic storm process, this paper selects the typical New Scotia area as the research object, the corresponding earth resistivity model is established by using the finite element method. The induced electric field on the earth surface is calculated by changing coast dip angle and the surface and subsurface earth resistivity. From the results, when the resistivity of surface layer is almost the same as that of sea water, the coastal effect is not obvious, and with the increase of coast dip angle, the surface electric field in offshore area increases slightly. When the surface resistivity is larger than the seawater resistivity to a certain extent, the coastal effect will make the surface electric field in the offshore area larger than that in area away from coast, and the smaller the coast dip angle is, the more obvious this trend is. The surface induced electric field in area away from coast will not be affected by the coastal effect, so it will not be affected by the coast dip angle. The influence of subsurface resistivity on the surface electric field in both offshore area and area away from coast is the same: with the increase of subsurface resistivity, the surface induced electric field gradually decreases and finally tends to be stable. When the surface resistivity is determined, the influence of the coast dip angle on the surface electric field is small. The results of this paper can provide reference for coastal effect calculation and GIC hazard assessment during geomagnetic storms.
In order to reduce the risk of the power grid in the process of transformation, this paper proposes to optimize the sequence of the exit of the regional thermal power generator group. This method calculates the risk parameters of the power grid in the aspects of adequacy and safety after the exit of different thermal power units in the region by building the power flow mode of the regional power grid. On this basis, combined with emergency control measures and scheduling schemes, the risk parameters of different dimensions are transformed into economic losses, so as to generate a risk assessment matrix, and determine the exit sequence of the optimized thermal power unit by layer optimization method. Finally, taking a provincial power grid in China as an example, a comprehensive evaluation of the decommissioning sequence of thermal power units in this region is given, and the feasibility of this method is verified.
In this paper, a video based objection detection method is proposed for traceability system with deep learning method. The surveillance video is collected first, from which an annotated image database of target object such as people or vehicle was constructed to train convolutional neural network model off-line. With the trained model, a real-time target detection and recognition system is designed and implemented. The proposed method mainly includes three aspects: video processing, target detection and object recognition. It provides a variety of video interfaces to support the downloaded video and real-time video stream. The experimental results indicate that the proposed deep learning based detection method is efficient for the traceability application.
The performance of bucket-wheel excavators used for the selective mining of multiple-layered lignite deposits is mainly determined by the physical and mechanical properties of the excavated material (digability parameters), as well as by the thickness and the inclination of the excavated layer. The lack of sufficient information regarding the diggability of the excavated layers during mine design stage results in a poor estimation of the bucket-wheel excavator's performance. Such inaccurate estimation could lead to inappropriate equipment selection. The possibility of prediction of the bucket-wheel excavator performance by using an expert system was investigated in the present study. The development of the expert system was based on the existing boreholes' data related to physical and mechanical properties of waste and lignite layers, on their geological descriptions and on the existing experiential knowledge about selective mining by bucketwheel excavators. The developed expert system was applied for the prediction of the performance of the bucket-wheel excavators used for the exploitation of Mavropigi Lignite Mine in Ptolemais area (Northern Greece).
