Research On Explicit Model Predictive And Low Complexity Control Technology Of VSV-HVDC

With the development of power system,compared with traditional AC transmission technology,HVDC has been widely used because of its larger transmission capacity,suitable for long-distance transmission and fast power regulation.However,the traditional HVDC converter is a semi-controlled ordinary thyristor device,which can only control the switching on of the device,while the switching off of the device needs to be completed by using the connected AC system to provide the commutation voltage.In order to better supply power to weak AC and island power grids,Voltage Sourced Converter-based High Voltage Direct Current(VSC-HVDC)emerges as the times require.Compared with conventional HVDC transmission system,VSC-HVDC system can independently control active and reactive power.As a system form which can greatly improve transmission efficiency,VSC-HVDC system has gradually been recognized all over the world.Because of this,the control of VSC-HVDC has become a hot research topic.The device structure and operation principle of VSC-HVDC system are introduced.Mathematical models in ABC three-phase coordinate system on both sides of rectifier and inverter are established.The original model is transformed into a third-order mathematical model in dq coordinate system by using Park transformation matrix.Considering the obvious non-linear relationship between DC voltage and current in the system,the original model is piecewise linearized at a specific equilibrium point by using small deviation linearization method.Then,the method of explicit model predictive control(EMPC)developed on the basis of model predictive control(MPC)is introduced.The state space is convex partitioned by using the theory of multi-parameter quadratic programming(MP-QP).A large number of on-line control calculation processes are transferred to off-line stage for predictive calculation.At the same time,in order to further reduce the computational complexity of EMPC,the separation function method and the concept of misjudgement point are introduced in the offline part to reduce the offline partitions that need to be predicted.In the online part,the reachable partition method is used to narrow the search range of state variables.Finally,the validity of the reachable partition method and the separation function method is verified by simulation calculation.After embedding the separation function,the control performance of the explicit model predictive control is verified in such environments as load surge,active and reactive power mutation and system component parameter mutation.