Commutation failure is inevitable in high-voltage DC transmission systems,usually triggered by faults in the AC system on the receiving end,and the relevant mechanisms and controls have been well researched.Due to the strong coupling between the AC and DC systems,when a disturbance occurs on the AC systems of the sending end,if the DC control system does not respond properly,the operating state of the receiving end is affected by the interaction of the electrical and control quantities,which can also lead to commutation failure on the receiving end.The research on commutation failure caused by disturbances in the AC system at the sending end is still at a preliminary stage.This paper analyzed the mechanism of commutation failure on the receiving end caused by disturbances on the sending end,and carries out research work in terms of both small and large disturbances on the sending end.Firstly,for the commutation failure on the receiving end caused by small disturbances on the sending end,a small signal analysis of the CIGRE standard DC system is carried out and it is found that there are inherent frequencies in the DC system that are in the low and sub/supersynchronous frequency bands,and there is a risk of resonance when oscillations of similar frequencies occur in the AC grid at the sending end after the disturbance.On this basis,the process and mechanism of the commutation failure and subsequent commutation failure triggered by the oscillation of the AC system on the sending end are analyzed respectively,and the expressions for the DC current are derived through the quasi-steady-state equations of HVDC to analyze the causes of the rise in DC current after resonance leading to a fall in the extinction angle.The effect of the whole process of control switching on the inverter side on the commutation failure is then investigated from the control system at both ends.Following this,based on the results of commutation failure and small signal analysis,commutation failure is prevented by adjusting control parameters of the DC system.Then,for the commutation failure on the receiving end caused by small disturbances on the sending end,based on the existing research,the fault is divided into non-serious and serious faults,and the fault is divided into four stages before and after the fault.The quasi-steady-state model and control equations of the DC transmission system are used to analyze the change state of the electrical and control quantities and the interaction between the electrical quantities and control quantities caused by the fault at different stages,and to study the interactive effects of electrical quantities and control quantities at different stages before and after the fault in the process of fault information transmission of the AC system at the sending end,so as to reveal the mechanism of commutation failure on the receiving end caused by the large disturbance on the sending.Finally,on the basis of the mechanism analysis,the influencing factors of the commutation failure are analyzed,focusing on the influencing factors that lead to the reduction of the extinction angle,and the prediction indicators of commutation failure are summarised.Then a machine learning-based commutation failure prediction method is proposed,in which the measurable electrical quantities,control quantities and prediction indicators of the system are taken as the input feature set,and the input feature set is filtered with features to form a prediction model of commutation failure through offline training to realise the machine learning-based commutation failure prediction.As the coupling between AC and DC systems deepens,the interaction between electrical and control quantities in HVDC systems increases and the risk of commutation failure increases significantly.This paper investigates the process and mechanism of commutation failure on the receiving end triggered by disturbances on the sending end,which can provide a reference for the design and operation of large new energy base and sending HVDC systems in the Shagor desert region. |