Simulation Research On Dynamic Power Characteristics Of Wind Turbine Driven By Fusion Of Physical Model And Operational Data Fusion

A high percentage of wind power connected to the grid is an inevitable feature of future power systems.Wind turbine operation control,wind farm operation control and optimal scheduling of power systems containing large-scale wind power are the key technologies for realizing high-ratio wind power grid connection.Most of the existing wind turbine simulation operation methods are based on physical models,but this method cannot quickly and accurately simulate the operating characteristics of wind turbines under complex dynamic conditions,especially when considering complex dynamic conditions such as time series,the simulation operation accuracy is poor.In view of the above problems,this paper focuses on the simulation method of wind turbine dynamic characteristics based on physics-data fusion drive.The main research contents are as follows:(1)Construction of a full-scene sample set based on GH Bladed simulation and actual operation data.High-quality data set is an important basis for improving the model training quality and model simulation accuracy.Based on GH Bladed software,the WTG simulation model is built to calculate the operating status,characteristics and load data of WTGs under different operating conditions,which can effectively compensate for the lack of data due to insufficient operating conditions in real situations.Combined with the actual operation data of WTGs,a full-scene WTG operation sample set is built to provide a solid data base for the simulation of dynamic characteristics of WTGs.(2)Constructing a sample set of wind turbine operation under typical scenarios based on orthogonal tests.Considering the different interconnections between the operating parameters and operating states of wind turbines under different typical operating scenarios,a sample set of wind turbine operation based on orthogonal tests is constructed.In order to further improve the accuracy of the dynamic characteristic model for wind turbine operation.based on the actual operating situation of wind turbines,three typical operating scenarios are considered,and orthogonal experiments are introduced to analyze the primary and secondary impact of wind turbine operation parameters on the operating status under different typical operating scenarios.Based on the range analysis and variance analysis results in the orthogonal experiment,the most suitable parameter input quantity and input sequence for the dynamic characteristics simulation model of wind turbine operation are selected.(3)Building a dynamic characteristics model of wind turbine operation driven by physics-data fusion.Considering that the existing wind turbine simulation model cannot simulate the operating characteristics of wind turbines quickly and accurately when considering complex dynamic conditions such as time series,and the simulation accuracy is poor,a wind turbine operating dynamic characteristics model based on physical-data fusion is established.The mapping relationship between each operating parameter and operating characteristics of WTGs in different typical scenarios under time series,i.e.,the nonlinear mapping model between wind condition parameters,WTG operating condition parameters and WTG output power and tower root load,is introduced by the fusion of sequence learning and attention mechanism of long and short term memory neural network.Using the actual operating data of wind turbine for arithmetic analysis,the method proposed in this paper can accurately mine and learn the operating dynamic characteristics of wind turbine time series under different typical scenarios,so as to improve the simulation accuracy and computational efficiency of the dynamic characteristics of wind turbine output power and tower root load.