The high penetration of distributed renewable energy aggravates the power fluctuation and uncertainty of distribution network. In this paper, a two-stage optimal operation strategy of distribution network based on multi-objective reduction and multi-step reconfiguration was proposed. In the day-ahead stage, multiple operational flexibility index models of voltage deviation, power loss, total operating cost, net load peak-valley difference and net load coefficient were constructed. On the basis of the model, a multi-flexibility index dimension reduction method based on target correlation was proposed, which aims to balance the relationship between various indicators, reduce the number of targets and improve computational efficiency. In the intraday phase, a distribution network reconfiguration method based on multi-step switch sequence exchange was proposed by cooperatively controlling the sectionalizing switches and tie switches, which fully exploits the available flexibility resources of the system. Finally, the simulation of arithmetic cases based on the IEEE-33 node system, the IEEE-84 node system, the IEEE-119 node system and the IEEE-136 node system were carried out, and the results show the effectiveness of the proposed two-stage distribution network flexibility enhancement strategy.

The calculation of maximum power supply capacity is an important issue in optimal operation of UHV complex large power grid. With the large-scale access of high-proportion renewable energy to the power grid, the problem of power and energy balance is becoming more and more prominent. Considering the stability spinning constraint, coupling section constraint and unit regulation capacity constraint, the calculation problem of maximum power supply capacity presents high-dimensional characteristics, which is difficult to model and solve. According to the limitation of transmission section and the characteristics of partition operation of UHV complex large power grid, an optimal dispatching model of maximum power supply capacity of UHV complex large power grid considering section coupling control was established. Then, the Simplex method (SM), Interior Point method (IP), Simulated Annealing (SA), Beluga Whale Optimization (BWO) and Artificial bee colony algorithm (ABC) were applied to solve the model respectively, and the maximum power supply capacity of the large power grid in the next 15 minutes was calculated. The coordinated optimization scheme of various types of power supply and reserve was proposed. Finally, a regional power grid was taken as an example to verify the model and the solution method. The simulation results show that because of the large scale of the power system and the complex coupling relationship between the sections, the heuristic intelligent optimization algorithm has a long running time, slow convergence speed and low search accuracy. The interior point method converges rapidly, has strong robustness, is insensitive to the selection of initial values, has better stability and computational efficiency, which can provide effective and practical support for power supply and power balance of complex UHV power.

Under the background of building a new-type power systems, the hydro-thermal-wind-solar-storage system with a high proportion of new energy has become the main research object of power system planning. It is necessary to perform high-frequency iterative calculations on the long-term supply-demand balance of the hydro-thermal-wind-solarstorage system to achieve quantitative analysis of sensitive boundaries, which puts forward higher requirements for the calculation efficiency of the analysis method. Considering the working characteristics of various power sources and the output probability characteristics of new energy sources, this paper proposed a heuristic unit commitment algorithm based on operating rules, which realizes the daily determination of the unit start-up mode throughout the year, and further constructed a supply-demand balance model. Finally, a case study based on a regional power grid example proved that the proposed model and method were feasible and effective, which provides reference for analyzing long-term supply-demand balance of the hydro-thermal-wind-solar-storage system.

The aim of this paper is to describe the drought events in the Zhanghe upstream comprehensively. Firstly, monthly precipitation data were selected to calculate standardized precipitation index. And then, both drought duration and drought severity were identified according to runoff theory. Finally, Copula function was adopted to establish joint distribution between drought duration and drought severity to explore the spatial and temporal distribution of drought characteristics. The results show that the highest frequent drought events in the Zhanghe upstream from 2001 to 2020 is 3 months drought, and the drought severity ranges from 2.24 to 4.92; There is a strong positive correlation between drought duration and drought severity, and Frank Copula function is the most optimal joint distribution function of them; The joint return period of drought characteristics ranges from 18.56-23.94 years, and the northwest and southeast parts of the Zhanghe upstream basin are the two key drought risk areas; The co-occurrence return period of drought characteristics ranges from 22.25-31.94 years, and the northwest, northeast and southeast parts of the Zhanghe upstream basin are the three key drought risk areas. Consequently, it can be obtained that the joint distribution can analyze drought events comprehensively, and thus provide scientific guidance for thorough evaluation of drought.

In view of the relatively serious drought in the Feiyun River basin in recent years, the inability to effectively utilize the regulating capacity of the counter-regulating reservoir, and the inability to effectively guarantee the timeliness and scientificity of the operation, which affect the water supply safety of nearly 6 million people in Wenzhou, the anti-drought operation model and data center of the Feiyun River basin are established. The real-time data of rainfall, water level, water supply flow and ecological flow of multiple reservoirs in the basin are quoted. The meteorological data including 1 h, 3 h, 6 h short-term and imminent grid forecast data, and 1 d, 3 d, 7 d, 15 d, 30 d, 90 d and other long-term rainfall forecast data are introduced. Combined with water and soil moisture in the basin, reservoir and basin evaporation, support vector regression (SVR) calculation method is used to develop a distributed digital twinning business system based on SPARK components, which realizes early warning of available water supply, drought-resistant water level early warning, drought-resistant days early warning and ecological flow early warning, increases the use of the anti-regulation storage capacity of 2.8 million cubic meters and the reserve storage capacity of 2.93 million cubic meters. Thus, it provides data basis and technical support for the drought-resistant decision-making, scientific scheduling and "four pre-" realization in the Feiyun River basin.

Based on the agricultural drought-affected rate of 13 major grain producing provinces in China from 1980 to 2019 and the daily rainfall data, the climate-caucusing loss intensity (CCLI) were extracted using H-P filtering, and its responses to dry-days indices (DDI) and large-scale atmospheric-ocean indices (LAOI) were explored. The correlation analysis between CCLI of each province and DDI in different months indicated that summer drought had the higher contribution to the disaster situation, especially the total number of dry days in July was identified as the primary drought-causing factors (PDCF) in 10 provinces. The PDCF in Henan, Jiangsu, Hubei, Inner Mongolia, Jilin, and Liaoning all showed a significant strengthening trends, indicating the intensification of drought risk in the key month of summer; The fluctuation of PDCF in each province had a main cycle of 2-3 years; At the time-lag of 0-12 months, circulation signals such as North Atlantic Oscillation, Pacific Warm Pool, and Arctic Oscillation had the significant impacts on the drought situation. This study can provide scientific basis for regional drought monitoring and early warning.

In order to solve the problem that the current existing precipitation trend analysis methods cannot simultaneously detect precipitation trends in low and high value areas, this article introduces innovative trend analysis (ITA) and combines it with traditional trend analysis methods for analysis. Taking the station data of Yili River basin from 1961 to 2020 as an example, the ITA and Mann-Kendall (MK) series, Theil Sen and other seven methods are compared and analyzed. The results show that the annual precipitation of stations in Yili River basin, except Huocheng Station, Yining County Station and Zhaosu Station, increases significantly in other stations in high value areas, which is the same as the trend of annual extreme precipitation; The extreme precipitation intensity in the low and median areas of other stations shows an increasing trend, while the high value areas show a decreasing trend, and the overall trend is increasing; The ITA can reflect the range of rainfall changes and detect the trend changes of rainfall in low, medium, and high value areas. Therefore, the combination of ITA and traditional trend method to conduct trend analysis on the overall rainfall and rainfall in high and low regions simultaneously will help to more comprehensively analyze the precipitation change law of Yili River basin.

Due to the existence of snow, melting snow and frozen soil, the hydrological cycle in cold regions has its particularity. In order to probe into the hydrological law of small watershed in cold region and simulate the process of runoff formation, the hydrological data of Yongcuihe River basin from 1994 to 2015 were selected as the research area. The SWAT model for runoff simulation in cold region was constructed, and the applicability of DEM with different resolution, as well as three kinds of meteorological data, CMADS data set and CFSR data set, in Yongcuihe River basin was analyzed. The results show that the SWAT model can simulate the runoff process of Yongcuihe River. DEM resolution has a great impact on the extraction of watershed features, but has no obvious effect on runoff simulation. The applicability of different meteorological data has significant difference, the simulation effect of measured meteorological data is the best, the simulation effect of CMADS data set is better, and the simulation effect of CFSR data set is worse, which provides a reference for further study of small watershed in cold region.

In order to improve the accuracy and reliability of runoff prediction, the advantages of EMD in dealing with non-stationary time series are introduced, and a BP neural network prediction model based improved adaptive noise complete set empirical mode decomposition (ICEEMDAN) and whale algorithm (WOA) optimization is established. Taking the inflow runoff prediction of Jinpen Reservoir in Heihe, Shaanxi Province as an example, a simulation model based on multiple intelligent optimization algorithms is established to predict the inflow runoff of the reservoir. At the same time, historical data of different time series, such as precipitation and runoff, are selected as input factors to compare the prediction ability and results of BP, WOA-BP, ICEEMDAN-BP and ICEEMDAN-WOA-BP models under the same input factor conditions. The results show that as far as the input sequence is concerned, the prediction effect of the model with precipitation as the input factor is better than that of the model with runoff as the input factor; For different algorithms, ICEEMDAN-WOA-BP model has good stability, Nash coefficient can reach 80%-90%, and the prediction accuracy is higher. The proposed ICEEMDAN-WOA-BP model can provide technical support for river runoff prediction, reservoir hydrological prediction and watershed water resources management.

To study the trend of runoff evolution under the influence of climate change and human activities, the Qinhuai River Basin was taken as the research area based on the hydrometeorological data of representative stations. Firstly, the Mann-Kendall method was used to study the variation trend of hydrometeorological elements in the basin. Secondly, the abrupt change points of the runoff series were discussed by orderly cluster method and double cumulative curve method. Thirdly, the key factors affecting runoff were screened from numerous hydrometeorological factors, and the Wavelet Analysis-BP neural network coupling model was established to predict the runoff. Finally, attribution analysis method was used to identify contribution of climate change and human activities to the runoff variation. The results show that the rainfall of the Xinhe river sluice presents a downward trend, while the Wudingmen sluice and the whole basin show an upward trend, but none of the trends were significant. The runoff of the basin shows an upward trend, and the increase trend of Xinhe river sluice is not significant, both the Wudingmen sluice and the whole basin increase trend pass the test of 0.01 confidence level, and the annual evaporation tends to decrease. There is a great consistency between the Wudingmen sluice and the whole basin in the variation trend of rainfall, runoff, and the time of abrupt change point. The contribution rates of human activities and climate change to runoff increase were 73.92% and 26.08%, respectively, the human activities were an important factor that led to the runoff variation. The conclusion provides a scientific basis for flood and drought prevention and water resources planning in this basin.