A Typical Reduced-scenario Analysis Method For Wind, Photoelectric, Load Scene And Its Application In Power System Planning

The middle / long-term planning and dispatching of power system involves the analysis and evaluation for the vast variety of scenarios and solutions.The Output Power of wind and photovoltaic power station varies with wind energy and solar radiation,the influence of weather,season,region make the outputs have the characteristics of strong randomness,volatility and intermittent.Because the increasing capacity of wind and photovoltaic farm will influence real-time balancing of power,severe challenges are proposed,the power system requires the integration capability of wind and photovoltaic power when planning and operation.The operation of scenario analysis will be affected by the randomness and uncertainty of wind or photovoltaic power and load demand,which further increasing the computations' complexity and time.In order to capture typical scenarios with a desirable efficiency and precision,a typical reduced-scenario set generating method is proposed to reduce the number of scenarios dramatically and establish the time-saving evaluation and analysis.The method is implemented by clustering analysis of the wind farm output as well as load profile during the computing period,then a set of typical reduced-scenarios is produced which can reflect the characteristics of the history data.Studying the typical reduced-scenario method has a great significance for power network planning and design,evaluation of wind or photoelectric power grid Integration,power supply optimal configuration,energy storage planning and operation scheduling.In this paper,the wind,photoelectric and load scene characteristics in grid system and clustering analysis of planning and dispatching of power system are analyzed first.Then,the randomness,volatility and intermittent of the wind / photoelectric / load's inherent characteristic are separately studied.Meanwhile,the wind / photoelectric / load transformation law of the daily,days and seasonal data are analyzed.In this paper,we summarize the scene analysis methods in the power system research field,including the scene generation methods and the scene reduction methods.And by researching the deterministic scene reduction,a typical reduced-scenario method is proposed.Also,the key technique of clustering analysis algorithm in the typical reduced-scenario method is introduced.By studying the improved k-means clustering algorithm,the clustering algorithm is utilized to reduce and merge the scenes.Meanwhile,in order to test the effectiveness of the improved k-means clustering algorithm based typical reduced-scenario method,the annual wind and load comprehensive scene set data and the monthly wind,photovoltaic and load comprehensive scene set data are selected separately to conduct a validity check of the algorithm.Finally,the typical reduction scene set data are used in the evaluation of wind power grid integration and the computation of Economic Operation Optimization of Micro-grids.The results show the high accuracy and efficiency in the computation when utilized the typical reduced-scenario set generating method,thus,the substantial engineering value of the algorithm is presented.