Transient Voltage Stability Assessment Of Wind Power Grid-connected System Based On GrcForest Model

With the continuous access of large-scale wind power to the power grid,the study of power system transient voltage stability is becoming more and more important in ensuring the safe and stable operation of the power system.In view of the shortcomings of machine learning in the rapidity and accuracy of transient voltage stability assessment of wind power grid-connected systems at the present stage,a method of transient voltage stability assessment of wind power grid-connected systems based on grcForest model is proposed.Firstly,guided by energy shortages and the rapid development of wind power technology in recent years,a dynamic equivalent modeling of wind farms based on the Canopy Kmeans algorithm was completed on the basis of the mathematical model of doubly fed wind turbines.The Canopy algorithm can optimize the number of clustering centers in the Kmeans algorithm,which needs to be manually set,and shorten the calculation time;Select operating parameters such as wind speed,rotor speed,electromagnetic torque,and active power as clustering indicators;Compare the simulation results of equivalent models with detailed models to evaluate the accuracy of equivalent results verified by indicators.Secondly,analyze the key influencing factors of transient voltage in wind power grid connected systems,and construct and select input features based on transient faults.A total of 25 feature vectors are selected;The random forest algorithm is used to rank the importance of the input features,and 13 features with the highest importance are selected,so that the subsequent transient voltage stability assessment can be completed faster and more accurately.Finally,in order to further ensure the rapidity and accuracy of the transient voltage stability assessment of the power system,the grcForest model is applied to complete the transient voltage stability assessment of the wind power grid-connected system.Aiming at the problem that the number of input features may increase or decrease with the increase of the number of cascaded forest layers,the residual network is used to optimize it to ensure that the model after the increase of the number of layers can still maintain the original learning ability;Complete the parameter setting and performance optimization of the model through offline training of the evaluation model;The constructed input characteristics are applied to the grcForest wind power grid-connected system transient voltage stability assessment model,and the accuracy rate,recall rate and leakage rate are used as assessment indicators to complete the wind power grid-connected system transient voltage stability assessment.The simulation analysis of IEEE 10-machine 39-bus system and the comparison with different machine learning methods verify the accuracy and effectiveness of this method.