Research On Online Decesion-making Method For Distributed/Centralized Undervoltage Load Shedding

With the expansion of the scale of power system in our country,the voltage level of transmission line is improving gradually,intermittent power like wind power's proportion is increasing year by year,the voltage stability problem has become more and more prominent.Voltage instability has become a major threat to the operation of power system.As the last line of defense to prevent voltage instability,load shedding is drawing more and more attention.However,the traditional undervoltage load shedding method has been difficult to meet the demand of an actual system,so it's necessary to find an economic and efficient undervoltage load shedding control method.Due to voltage problems' local characteristics,load shedding control should be executed based on voltage partioning.Whether voltage partioning result is reasonable or not determines the undervoltage load shedding control method' effective.Based on principle and objective function of the spectral clustering algorithm,this paper turns the partitioning problem into mathematical problems by means of complex network theory,using k-means algorithm to solve the problem.At the same time,an evaluation method of partition validity is proposed in this paper.The simulation results show that the voltage partitioning method based on spectral clustering can divide power system into several parts reasonably.The optimization of the partition number was realized by analyzing the partition results obtained by selecting different partition numbers of the k-means clustering algorithm.Traditional undervoltage load shedding method generates control strategy outline and the voltage amplitude is used to select the dangerous nodes.When used in a real system,operation error is common.This paper improves the traditional undervoltage load shedding method.Theoretical derivation shows that some factors such as dangerous node selection and location of load shedding have an affect on load shedding control effect.According to the above questions,this paper presents a dangerous node selection index and a load sensitivity index,undervoltage load shedding control method is generated online based on voltage partitioning.Simulation results show that distributed undervoltage load shedding control method based on online decesion-making can fliter dangerous nodes effectively and choose location of load shedding reasonably.It has very strong robustness for different forms of disturbance in power system.The control target of undervoltage load shedding is to realize the optimal load shedding in the premise of guarantee the safe and stable operation of power system.Distributed undervoltage load shedding control method lacks coordination between regionals,which is difficult to achieve the target.This paper presents a distributed/centralized undervoltage load shedding method,the control strategy is generated in a centralized way and carried out in a decentralized way.Using information obtained by PMU,a two-level load shedding scheme is designed.Simulation results show that distributed/centralized undervoltage load shedding method can prevent the occurrence of voltage collapse phenomenon effectively,is adaptable to different forms of disturbance.It has a better voltage regulation effect and smaller economic loss compared with the distributed undervoltage load shedding method.