Research On Modeling And Control Method Of Superconducting Magnetic Field DC Power Supply

With the continuous advancement of society,the level of industrial technology in the world is also developing rapidly.The application of magnetic fields in industrial manufacturing,engineering technology,scientific research,aerospace,medicine,military and other fields is increasing.For example,in the preparation process of Czochralski single crystal silicon,the method of applying Lorentz force to the melt by an applied magnetic field to suppress convection has become an effective method for improving the quality of the crystal.Due to the rapid development of superconducting technology,superconducting magnets with higher magnetic field strength,smaller volume and power consumption are gradually applied in the field of single crystal growth,which satisfying the production requirements of high magnetic field strength and high stability.The control of the superconducting magnetic field is essentially a switching power supply control problem.It is a system consisting of a superconducting magnet and a magnetic field power supply and can be represented by an inductively loaded switching power supply circuit.Due to the special nature of the superconducting magnetic field,small current changes cause large fluctuations in the magnetic field strength.Therefore,the superconducting magnetic field used in single crystal preparation has very high requirements on the accuracy,stability and reliability of the power supply.Since the traditional switching power supply modeling methods are mostly based on the idea of averaging and approximation,the nonlinear characteristics of the switching circuit cannot be essentially embodied.Therefore,this paper will focus on the control problem of superconducting magnetic field DC power supply,and use Buck type switching power supply topology to study the hybrid system modeling and control methods.Based on the hybrid theory,the nonlinear characteristics of the Buck-type circuit are analyzed.Firstly,the Hybrid Automata(HA)model of the Buck circuit is established and the control strategy is designed based on the discrete characteristics of the circuit.The boundary conditions are designed to control the strategy and the critical conditions of the two modes of operation.Since the method is essentially open-loop structure,the parameter change cannot be adjusted accordingly.Therefore,the parameter identification method based on hybrid system is proposed to obtain the circuit parameter estimation value,and feedback to the controller to update the control parameters,thus forming a closed loop.The control structure improves system stability.In view of the characteristics of the inductive load of the superconducting magnetic field,the control method of the hybrid automaton has a drawback that it is difficult to directly obtain the switching condition.Therefore,this paper proposes a Mixed Logical Dynamic(MLD)model and control method for inductive load Buck circuits.Firstly,the hybrid logic dynamic model is established according to the system switching state.The discrete state in the process is transformed into a linear inequality through the equivalence relation of the logical proposition.For the control problem of the model,this paper uses predictive control to solve.The model is transformed into the Mixed Integer Quadratic Programming(MIQP)form,and the branch and bound method is used to solve the planning problem to realize predictive control.The problem of steady-state deviation due to parameter mismatch resulting in model mismatch,this paper uses Model-Free Adaptive Control(MFAC)as the target value of the outer loop to adjust the inner loop predictive control.When the model is consistent with the system parameters,the control effect good;this method can make the system quickly stabilize at the target value when the model and system parameters are mismatched.In this paper,two modeling methods of mainstream hybrid theory are studied.Firstly,the hybrid automata model of Buck circuit and the closed-loop feedback control method based on parameter identification are simulated and verified by hardware circuit experiments.Secondly,the hybrid logic dynamic model predictive control method for inductive load Buck circuit is verified,and the double loop control structure is designed for the problem of model mismatch caused by parameter deviation.The effectiveness of the method is verified by simulation experiments.The problem that the computational complexity of the predictive control optimization is too large can not directly carry out the hardware circuit experiment.A simplified verification strategy is used to verify the steady-state performance of the proposed method.Through the above simulations and experiments,the effectiveness of the two hybrid modeling methods and their improvements on DC switching converters is demonstrated.