Research And Application On Key Technique For Asynchronous Based On Three-level Dual-PWM Converters

Intelligent manufacturing and green development will bring a significant change in the energy generation and use.It will also promote the upgrading of transmission equipment.Traditional diode rectifier of the converter makes the quality of power grid worse and the motor cannot be achieved energy feedback in the mine hoist applications.But,dual-PWM converter with three level topology structure,which can realize not only "green" energy transformation,also can realize four quadrant operation of ac asynchronous motor.Therefore,the converters have been more and more put into the production fields.In this dissertation,regard the three-level dual-PWM inverter drive three-phase asynchronous motor control system as the research object,the key techniques such as the inhibition of startup inrush current,neutral voltage balance,dead zone compensation,sensor-less speed estimation system,and motor parameter identification are studied.(1)Aiming at neutral-point unbalance and the dead zone effect causes the output current distortion,a hybrid neutral-point balancing algorithm based on fuzzy optimization and three closed-loop dynamic without dead zone compensation control strategy are proposed.The hybrid neutral-point balancing algorithm clarifies the sector selection according to whether or not contains medium vector.In the area of no the medium vector,Two-level space vector control algorithm based on seven segment switching sequence is used to solve voltage deviation caused by the redundancy small vector.In the rest area,time controlling element of the redundant small vector is optimized by the fuzzy control to achieve the meticulous control of the neutral voltage.A dynamic three closed-loop without dead zone compensation adds the closed-loop unit of the dead zone to the dual-closed loop control system and forms a three closed-loop dynamic dead-zone compensation system.The inaccurate problem of passing zero current detection is solved by setting the threshold value.(2)For the issues of large startup inrush current in the process of starting and dynamic load mutation and the current tracking delay problem in digital control system of the three-level PWM rectifier,a new double closed-loop control strategy is proposed based on the outer square voltage loop and current model predictive control inner loop.The large inrush current is effectively suppressed by the method of the square voltage feedback control and the load current and reactive current feed-forward compensation.Aiming at the considerable computation and inconvenient realization of the traditional model predictive control,an improved model predictive control based on first order differential equation is proposed.The current value of the future two sampling time is calculated and optimized by the performance function optimization algorithm which contains delay time of rectifier.The strategy can better suppress the impact of the startup current and improve dynamic current tracking capability in the load mutates.(3)For the optimal solution of feedback adaptive rate parameters of full order state observer is difficult to find,an improved particle swarm offline optimization algorithm is proposed based on a given part of the population parameters is given.The feedback adaptive rate parameter design criterion is given by the frequency domain design method.Combined with the optimal advantages of particle swarm algorithm,some designed and encoded parameter values are mixed to the random initial population to increase the number of the fine individuals.The method can improve the convergence speed and search efficiency and achieve good optimization effect.Aiming at the problem of error and the amount of calculation,an improved method of discretization is proposed.The euler method is adopted in low speed and the trapezoidal method is adopted to discrete the full order state observer.At the same time,the trapezoidal method are reasonable approximate simplified.Compared with the euler method,the discretization precision of improved method is greatly increased in the high speed and there is no obvious increase amount of calculation.Compared with trapezoidal method,the amount of calculation is greatly reduced to realize the full order state observer of accurate digital control with the premise of guarantee accuracy advantages.Simulation and experimental results show that the speed estimation system has good dynamic response speed and steady-state accuracy based on the given algorithm(4)In the motor parameter identification based on least square method,in order to improve the sampling signal filtering accuracy and reduce amount of calculation,a three-level low-pass filter is presented to filtering the sampling signal.At the same time,a forecast correction method is used to solve the equation of state of tertiary low-pass filter and can receive the signals directly when the parameter identification.In this way,it avoids the discretization of the first derivative and the second derivative of the sampling signal.Both the effect of derivative discretization error and the recognition of two-dimensional matrix inversion problem are avoided.Reducing the amount of calculation and improving the calculation accuracy are achieved in the identification.The simulation results are presented to demonstrate the stability,consistency and accuracy of the proposed method.(5)The prototypes of small power and high-power three-level dual-PWM converter are developed.Phase-lock experiment,rectifier start experiment,dynamic performance experiment,and disturbance resistance experiment of rectifier are carried out respectively.Starting,braking experiment and load experiment of inverter are also carried out respectively.The related full order state observer experiments are also carried out in low frequency.The experimental results show that the two prototypes have good static characteristic and dynamic characteristic.