| Positional control is an important part of tokamak plasma control.In recent years,various advanced plasma configurations have gradually been applied to many tokamak devices,such as ITER-like configuration,snowflake configuration,and super-X con?figuration.While the advanced configurations improve the plasma confinement and increase the plasma parameters,it also leads to pose higher requirements for position control.On the one hand,the position control needs to be more accurate,and on the other hand,due to the advanced configuration usually brings higher vertical displace-ment growth rate,so it also puts forward higher requirements on the dynamic response of vertical displacement control.In order to meet the requirements of advanced config-uration control ability,this dissertation optimizes the shape and position control ability of EAST from several aspects such as system response model,system identification,vertical displacement control optimization and multiple input multiple output control.In order to optimize the shape and position control system on EAST,a sufficiently accurate system response model must be available.At the same time,the response model must also be properly simplified to suit the needs of the controller design.In this paper,the RZIP plasma response model which is widely used in the field of plasma control is adopted.Based on the RZIP model.the plasma vertical displacement control response model and the shape control response model are obtained,and the obtained high-order model is simplified.And using the experimental data,the response.model was verified.While using the physical method to establish the response model,this paper also uses the system identification method to establish the response model for the system.System identification can use experimental data to create a control-oriented re-sponse model for a specific system.In this paper,using the ARMAX model in system identification,through the corresponding identification experiments,a system identifi-cation model is established for vertical displacement control and configuration control,and verified with the RZIP model.The results show that the system identification model has good accuracy.On this basis,this paper uses the system identification model to op-timize the vertical displacement control parameters.For better control of the high-growth rate plasma shape and optimize the vertical displacement control problem encountered in previous EAST experiments,this paper designs a time optimal controller and a velocity controller.The time optimal controller uses the optimal control principle and uses the control time as the optimal amount,so as to optimize the dynamic response.In a system that limits the maximum value of the control component,it can be shown that the control component of the time-optimal control is to switch back and forth between two boundary values,which will result in the noise effect near the equilibrium position will be amplified,so the time optimal controller design incorporates a PID controller to ensure that it can better adapt to system noise.The velocity controller is an optimization scheme proposed for the coupling of the vertical displacement control brought about by the separation of the EAST fast and slow Z control.The velocity controller no longer feedback the position of the vertical displacement,but to feedback vertical velocity and the IC current,which as a stability term,to achieve to stability control.In the vertical displacement control,the noise of the system will have a great impact on the entire control loop.Based on the experimental data,this article analyzes the signal noise level that EAST can currently use for vertical displacement control,and based on the experimental noise level,using the TOKSYS tool.The package simulates the control performance of time-optimized controllers and speed feedback controllers under different configuration conditions,and verifies that the two controllers can be used for plasma under different configurations.In one discharge,in order to maintain the plasma in the target configuration,in addition to controlling its vertical displacement.other configuration parameters need to be controlled to maintain its shape.In the ISOFLUX control algorithm adopted by EAST,the shape control is to control the shape and position accurately by controlling the difference between the magnetic flux at the X point and the magnetic flux at sev-eral control points on the plasma boundary by 12 independent PF coils.The previous shape control was a single-input single-output strategy.That is,the control points cor-responding to the PF coils closer to the control point were selected.Each control point also received only 1 to 2 sets of PF coils control.This kind of control is concise and intuitionistic,but because of the strong coupling between various control quantities,it takes a lot of experiment and rich experience to make the system reach the ideal state.Based on the plasma response model,this article using the statically decouple method,designs a Multi-Input Multi-Output controller,and uses a singular value decomposition method to optimize the decoupling matrix so that it does not exceed the power supply limited.This paper uses TOKSYS and TSC tools to verify the decoupling effect of the decoupled matrix.In the experiment,the MIMO control algorithm is used successfully to achieve complete discharge and accurate control. |
