Research On Intelligent Dispatch Of Power Grid With Wind Power Based On Cognitive Theory

Facing the new vision of carbon peak and carbon neutrality,the new energy industry represented by wind power will usher in a period of accelerating development and become a key point to change the structure of energy supply in China.With the continuous increase of the share of new energy and system uncertainty,the consumption of a high proportion new energy is still a challenge to the flexible operation of the system,so it is necessary to reconsider the modeling and operation of the power system.In response to the above problems,this paper introduces the theory of granular cognition into the economic dispatching of power systems with wind power,and proposes an adaptive dispatch scheme that takes into account the gradient of the net load.The main theoretical issues to be studied are as follows:Firstly,through the research on the basic notions and typical models of cognitive computing and granular computing,based on data-driven theory,a data-driven granular cognitive computing is proposed.Combined with the bottom-up information computer system driven by traditional data and the top-down human cognitive law,The expression spectrum construction of power domain and the establishment of power-time two-layer granulation model based on unsupervised machine learning are completed,and a multi-level and multi-granularity granulation structure is proposed.Secondly,in order to characterize the gradient changes in the hourly scale of the net load,a three-layer(cognitive layer-decision layer-feedback layer)and multi-period(day-ahead-real-time multi-time scale)adaptive dispatch scheme based on the expression spectrum is established.Among them,the cognitive layer realizes allweather environmental state awareness,and establishes an adaptive time period division mechanism for the net load curve;the decision layer realizes the formulation of day-ahead scheduling schemes;the feedback layer performs error control in realtime scheduling to adjust power deviation.Taking the improved IEEE-78 node interconnection system as an example,the feasibility of the scheduling scheme is verified.Finally,an adaptive scheduling scheme with wind power fuzzy particles is proposed.Based on the study of the discrete-time approximate distribution error,the forecast error is analyzed.The membership function of the fuzzy parameter is introduced through the analysis of historical data to construct the wind power fuzzy particle,thereby establishing an adaptive dispatch model that takes the dual error into account.The dual particle swarm optimization algorithm is used to solve the fuzzy chance constrained model.The results of calculation examples show that compared with traditional dispatch,the proposed method has better economy and calculation efficiency.