High-voltage variable frequency motor starting and control method based on model predictive control
Through the model prediction control method, the start-up and operation control of high-voltage frequency converter motors is optimized, which solves the problem that traditional methods cannot cope with load changes and dynamic environments, and achieves safe and stable operation and efficient energy saving of the motor and the power grid.
Patent Information
- Application Number
- CN202510265582.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-07
- Publication Date
- 2025-04-18
AI Technical Summary
The traditional high-voltage frequency converter motor start control method cannot flexibly respond to load changes and dynamic environments, resulting in a large impact on the starting current, affecting the motor life and grid power quality.
A dynamic model of a high-voltage frequency converter motor is established by using a model predictive control (MPC) method, and the dynamic behavior and load changes of the motor are predicted in real time, the startup process and operation control are optimized, and the inverter output is accurately adjusted.
It achieves smooth start, reduces the risk of current shock and overload during the motor start-up process, ensures the safe and stable operation of the motor and the power grid, and improves dynamic response performance and energy-saving effects.
Smart Images
Figure CN119853539A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of high-voltage variable-frequency motor control, and in particular to a high-voltage variable-frequency motor starting and control method based on model predictive control. Background Art
[0002] High-voltage variable-frequency motors are widely used in the modern industrial field, covering many important industries such as petrochemicals, metallurgy, mining, water supply and power supply. These industries often have extremely high requirements for the continuity, stability and accuracy of the production process. As the core power equipment in the production system, the startup and control process of high-voltage variable-frequency motors directly affects the operating efficiency and stability of the equipment, and thus affects the production efficiency and product quality of the entire production system.
[0003] Traditional motor starting control methods usually rely on a constant starting mode and cannot flexibly respond to load changes and dynamic environments. For example, direct starting, star-delta starting, conventional variable frequency speed control, etc., gradually expose many defects when facing complex and changeable industrial conditions. For example, the direct starting method will produce a large starting current shock, which may cause the grid voltage to drop suddenly, affecting the normal operation of other equipment in the same grid; if the star-delta starting method is adopted, although the starting current is reduced to a certain extent, the impact of the current shock on other equipment is reduced, but there are problems such as inflexible control and difficulty in meeting the fine starting requirements under different load characteristics; if the conventional variable frequency speed control method is adopted, it is also insufficient in responding to rapidly changing load conditions and in the comprehensive optimization of motor operating efficiency and stability. For example, in scenarios where the motor needs to accelerate and decelerate frequently or the load changes suddenly, it is difficult to accurately adjust the motor parameters in real time, which is easy to cause energy waste and increased equipment wear.
[0004] It can be seen that during the starting process of the high-voltage variable-frequency motor, if the starting current and starting torque cannot be effectively controlled, the excessive starting current impact will not only damage the insulation, winding and other components of the motor itself, shorten the service life of the motor, but also interfere with the power quality of the power grid and affect the normal operation of surrounding electrical equipment.
[0005] With the continuous advancement of control theory and computer technology, model predictive control (MPC) as an advanced control strategy has gradually attracted widespread attention. Based on the mathematical model of the system, it can predict the future dynamic behavior of the system and determine the optimal control action in real time through online rolling optimization and feedback correction mechanism.
[0006] In summary, in view of the limitations of existing control methods and the urgent need for high-performance starting and control of high-voltage variable-frequency motors in industrial production, introducing MPC into this field to carry out relevant research has important practical significance and practical value. Summary of the invention
[0007] The object of the present invention is to provide a high-voltage variable-frequency motor starting and control method based on model predictive control to solve the problems raised in the above background technology.
[0008] In order to achieve the above object, the present invention adopts the following technical solutions: A method for starting and controlling a high-voltage variable-frequency motor based on model predictive control comprises the following steps: S1. Establish high voltage variable frequency motor The equation of state in the rotating coordinate system; S2, discretize the state equation to obtain the state prediction matrix of the high-voltage variable frequency motor; S3. Based on the state prediction matrix, a comprehensive objective function of the high-voltage variable frequency motor predictive control model is constructed; S4, determining the constraint conditions of the high-voltage variable-frequency motor predictive control model; S5. Solving the high-voltage variable-frequency motor predictive control model based on a quadratic programming algorithm; S6. Apply the optimal voltage to the high-voltage variable-frequency motor based on the solution result.
[0009] Furthermore, the step S1 includes the following steps: By establishing the state equation of the high-voltage variable-frequency motor, we can obtain: ; in, A is the system matrix, B is the input matrix, yes The derivative of is the state variable of the motor, yes d The stator current in the axial direction, yes q The stator current in the axial direction, oh is the motor angular velocity, yes d Input voltage in the axial direction, yes q Input voltage in the axial direction, T is the transposed matrix; Under continuous time conditions, the motor is k The state equation is expressed as follows: ; ; in, yes d The rotor current in the axial direction, yes q The rotor current in the axial direction, is the resistance of the stator winding, is the resistance of the rotor winding, is the stator inductance, is the rotor inductance, It is mutual induction. is the number of pole pairs of the motor, J is the moment of inertia of the motor, B is the coefficient of friction.
[0010] It should be noted that the system matrix A and the input matrix B The elements can be determined by motor parameters such as stator resistance , rotor resistance , stator inductance , rotor inductance , mutual induction M Wait to confirm; is a coefficient related to the motor inductance parameter and is used for the calculation of the motor model. oh Reflects the rotation speed of the motor rotor, It will affect the stator current and motor losses. represents the coupling inductance between the stator winding and the rotor winding, Determines the relationship between the motor's magnetic field distribution and speed and frequency. J Describes the inertia of the motor rotor, which will affect the dynamic response of the motor. B Used to take into account the friction loss during motor rotation.
[0011] Furthermore, the step S2 includes the following steps: Set the prediction time domain N p and control time domain N c ,and N c ≤ N p , the state equation is discretized using the first-order Euler discretization method, and we can get: ; in, T s is the sampling time; In the prediction time domain, the discretized state equation is iterated to obtain: ; Writing the above formula in matrix form, we can get: ; in, It is the current moment k For the future moment k+i The predicted value of the motor state, is the state transition matrix, is the input matrix, is the control input sequence, which includes k To the future k + N c The predicted value of the input voltage is -1.
[0012] It should be noted that the state transfer matrix Describes the natural transition law of the motor state when there is no input change, which can be obtained by iterative derivation of the discretized state equation. Its elements are similar to the system matrix of the motor A , Sampling time T s etc., used to calculate the future state based on the current state of the transfer part, the input matrix It reflects the influence of control input on the prediction of future motor state. Its elements are related to the system matrix of the motor. A , input matrix B , Sampling time T s etc., used to calculate the control input sequence U ( k )’s contribution to future state predictions, U ( k ) is the decision variable in the optimization problem, and the optimal control input sequence is determined by solving the optimization problem.
[0013] Furthermore, the step S3 includes the following steps: The comprehensive objective function J Including tracking error J 1 and control input change J 2, which means as follows: ; in, l is the weight coefficient that weighs the tracking performance and control input stability. l When it is large, the penalty term of the control input change accounts for a large proportion in the comprehensive objective function, and more emphasis is placed on the stability of motor control. l When it is small, the penalty term of the control input change accounts for a smaller proportion in the comprehensive objective function, and more emphasis is placed on the tracking performance of the motor state to the reference state; The tracking error term J 1 means the following: ; in, oh ref ( i |k ) is the current moment k For the future moment k + i The target setting value of the angular velocity, It is the current moment k For the future moment k + i The predicted value of the angular velocity, i dref ( i | k ) is the current moment k For the future moment k + i of d The target setting value of the stator current in the axial direction, It is the current moment k For the future moment k + i of d The predicted value of the stator current in the axial direction, i qref ( i | k ) is the current moment k For the future moment k + i of q The target setting value of the stator current in the axial direction, It is the current moment k For the future moment k + i of q The predicted value of the stator current in the axial direction, α1, α2 and α3 are weight coefficients used to adjust the angular velocity, d The stator current in the axial direction and q The importance of the tracking error of the axial stator current in the comprehensive objective function; The control input variation term J 2 means the following: ; in, and The current moment k For the future moment k + i and k + i -1 d Predicted value of input voltage in the axial direction, and The current moment k For the future moment k + i and k + i -1q Predicted value of input voltage in the axial direction, β 1 and β 2 is the weight coefficient, used to adjust d Axis and q The importance of axis direction control input variation in the comprehensive objective function.
[0014] Further, the step S4 includes the following steps: The constraint conditions include current constraint and voltage constraint; The current constraint is expressed as follows: ; in, i dmax yes d The maximum value of the stator current in the axial direction, i qmax yes q The maximum value of the stator current in the axial direction; The voltage constraint is expressed as follows: ; in, u dmax yes d The maximum value of the input voltage in the axial direction, u qmax yes q Maximum value of the input voltage in the axial direction.
[0015] It should be noted that i dmax and i qmax It is usually determined by the rated current of the motor and the safe operation requirements. For example, the rated current of the motor is I rated In order to ensure the safe operation of the motor, it is usually set i dmax = i qmax = I rated Or according to a certain safety factor (such as 0.8 I rated ) to set; u dmax and u qmax Usually determined by the output capacity of the inverter and the insulation requirements of the motor.
[0016] Furthermore, the step S5 includes the following steps: First, the comprehensive objective function is linearly transformed, and the control input sequence get: ; Expand J 1. J 2 About z The expression of the state variable oh ( i | k )and z The linear relationship is expressed as follows: ; in, P ij_1 are state variables in different time domains With input variables The correlation coefficient, Q ij_1 are state variables in different time domains With input variables The correlation coefficient of R i_1 is a state variable The independence coefficient of Then we have: ; For state variables Do the same processing as above to get: ; in, P ij_2 are state variables in different time domains With input variables The correlation coefficient, Q ij_2 are state variables in different time domains With input variables The correlation coefficient of R i_2 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ; in, P ij_3 are state variables in different time domains With input variables The correlation coefficient, Q ij_3 are state variables in different time domains With input variables The correlation coefficient of R i_3 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ; in, P ij_4 are state variables in different time domains With input variables The correlation coefficient, Q ij_4 are state variables in different time domains With input variables The correlation coefficient of R i_4 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ; in, P ij_5 are state variables in different time domains With input variables The correlation coefficient, Q ij_5 are state variables in different time domains With input variables The correlation coefficient of R i_5 is a state variable The independence coefficient of Then the comprehensive objective function can be written as: ; in, H is a symmetric positive definite matrix whose elements are J In the expanded form z The quadratic coefficient of the element is determined, f is a vector whose elements are J In the expanded form z The coefficient of the first-order term of the element is determined; Then, the current constraint and voltage constraint are linearly transformed respectively, and the following is obtained: ; in, A dj_1 is the state variable in the constraint i d ( k ) and input variables u d ( j | k ), B dj_1 is the state variable in the constraint i d ( k) and input variables u q ( j | k ), C d_1 is the state variable in the constraint i d ( k ) is the independent coefficient, A dj_2 is the state variable in the constraint i q ( k ) and input variables u d ( j | k ), B dj_2 is the state variable in the constraint i q ( k ) and input variables u q ( j | k ), C d_2 is the state variable in the constraint i q ( k ) is the independent coefficient, A dj_3 is the state variable in the constraint u d ( k ) and input variables u d ( j | k ), B dj_3 is the state variable in the constraint u d ( k ) and input variables u q ( j | k ), C d_3 is the state variable in the constraint u d ( k ) is the independent coefficient, A dj_4 is the state variable in the constraint u q ( k ) and input variables u d (j | k ), B dj_4 is the state variable in the constraint u q ( k ) and input variables u q ( j | k ), C d_4 is the state variable in the constraint u q ( k )’s independence coefficient; Then the constraints can be written as: ; Writing the above formula in matrix form, we can get: ; in, A m is the constraint matrix, whose row vectors are transformed from constraints into z The coefficient composition of the linear inequality is: b is a constraint vector whose elements are determined by the constraints and constant terms in the predictive control model.
[0017] Furthermore, the step S5 further includes the following steps: 1) Set the initial feasible solution z 0 : z 0 is the control input sequence z = U ( k ) and satisfy all constraints ; 2) Set the convergence accuracy , step size parameter α and barrier parameters m 0 ; 3) Update solution: ; Among them, 0< α <1, , z k and z k+1 are three adjacent solutions during the iteration process; 4) Update barrier parameters: ; in, r is the update factor, 0< r <1; 5) Check whether the convergence conditions are met: Calculate the current solution z k+1 With the previous solution z k The difference between J ( z k+1 ) and the comprehensive objective function value of the previous iteration J ( z k ), when: ; When , the algorithm is considered to have converged and the iteration is stopped; otherwise, the next iteration is continued; 6) Obtain the optimal solution: When the algorithm converges, the solution This is the desired optimal control input sequence.
[0018] It should be noted that the feasible solution z 0 It can be determined based on the initial state of the motor, empirical values, etc. For example, when the motor is just starting, the initial voltage can be set to a smaller value. , the control inputs at other future moments can be set to initial values according to some simple ramp functions or fixed ratios to form z 0 ; Convergence accuracy is a positive number close to 0, for example , used to determine whether the algorithm converges. It can measure the degree of change of the objective function value or variable during the iteration process. The step size parameter α Used to adjust the step size in the search direction in each iteration, the obstacle parameter m 0 is a suitable positive number that gradually decreases during the iteration process, and is updated by the factor in each iteration. r Update, so that the barrier function has less and less influence on the optimal solution, and finally approaches the true optimal solution.
[0019] Further, the step S6 includes the following steps: At the current sampling time k , the optimal control input sequence The first element in Applied to the high voltage variable frequency motor as input voltage; At the next sampling momentk +1, get the actual state of the motor x ( k +1), repeat steps S1-S6, continuously update the predictive control model and the comprehensive objective function, solve the optimization problem and apply control input to achieve dynamic control of the motor operation process.
[0020] Compared with the prior art, the present invention has at least the following beneficial effects: The present invention proposes a method for starting and controlling a high-voltage variable-frequency motor based on model predictive control (MPC). The method takes advantage of the MPC and optimizes the starting process and operation control of the motor by predicting the dynamic behavior and load changes of the motor in real time. Specifically, the present invention uses the MPC algorithm to establish a dynamic model of the motor, and considers the constraints of the motor's current, torque, speed, etc., to accurately adjust the inverter output to achieve smooth starting and minimize the current shock and overload risks during the motor starting process, and ensure the safe and stable operation of the motor and the power grid; the present invention also verifies through simulation analysis that the above method is superior to the traditional starting method in terms of current fluctuation, vibration, and energy consumption during the starting process, and has better dynamic response performance and energy-saving effects, which provides a theoretical basis and practical guidance for the efficient and safe starting of high-voltage variable-frequency motors.
[0021] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. The specific implementation of the present invention is given in detail by the following embodiments and their accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 This is a flow chart of a method for starting and controlling a high-voltage variable-frequency motor based on model predictive control according to Embodiment 1 of the present invention; Figure 2 A graph showing the relationship between the motor speed and time in the simulation experiment in Example 2 of the present invention; Figure 3 A graph showing the relationship between the motor control input and time in the simulation experiment in Example 2 of the present invention; DETAILED DESCRIPTION
[0023] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention. The present invention is described in more detail by way of example with reference to the accompanying drawings in the following paragraphs. The advantages and features of the present invention will become clearer according to the following description and claims. It should be noted that the drawings are all in a very simplified form and are not in precise proportions, and are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.
[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. Example
[0025] Example 1
[0026] See also Figure 1 , a high-voltage variable-frequency motor starting and control method based on model predictive control, comprising the following steps: S1. Establish high voltage variable frequency motor The state equation in the rotating coordinate system: By establishing the state equation of the high-voltage variable-frequency motor, we can obtain: ; in, A is the system matrix, B is the input matrix, yes The derivative of is the state variable of the motor, yes d The stator current in the axial direction, yes q The stator current in the axial direction, oh is the motor angular velocity, yes d Input voltage in the axial direction, yes q Input voltage in the axial direction, T is the transposed matrix; Under continuous time conditions, the motor is k The state equation is expressed as follows: ; ; in, yes d The rotor current in the axial direction, yes q The rotor current in the axial direction, is the resistance of the stator winding, is the resistance of the rotor winding, is the stator inductance, is the rotor inductance, It is mutual induction. is the number of pole pairs of the motor, J is the moment of inertia of the motor, B is the coefficient of friction.
[0027] S2. Discretize the state equation to obtain the state prediction matrix of the high-voltage variable-frequency motor: Set the prediction time domain N p and control time domain N c ,and N c ≤ N p , the state equation is discretized using the first-order Euler discretization method, and we can get: ; in, T s is the sampling time; In the prediction time domain, the discretized state equation is iterated to obtain: ; Writing the above formula in matrix form, we can get: ; in, It is the current moment k For the future moment k+i The predicted value of the motor state, is the state transition matrix, is the input matrix, is the control input sequence, which includes k To the future k + N c The predicted value of the input voltage is -1.
[0028] S3. Based on the state prediction matrix, the comprehensive objective function of the high-voltage variable frequency motor predictive control model is constructed: The comprehensive objective function J Including tracking error J 1 and control input change J 2, which means as follows: ; in, l is the weight coefficient that weighs the tracking performance and control input stability. lWhen it is large, the penalty term of the control input change accounts for a large proportion in the comprehensive objective function, and more emphasis is placed on the stability of motor control. l When it is small, the penalty term of the control input change accounts for a smaller proportion in the comprehensive objective function, and more emphasis is placed on the tracking performance of the motor state to the reference state; The tracking error term J 1 means the following: ; in, oh ref ( i | k ) is the current moment k For the future moment k + i The target setting value of the angular velocity, It is the current moment k For the future moment k + i The predicted value of the angular velocity, i dref ( i | k ) is the current moment k For the future moment k + i of d The target setting value of the stator current in the axial direction, It is the current moment k For the future moment k + i of d The predicted value of the stator current in the axial direction, i qref ( i | k ) is the current moment k For the future moment k + i of q The target setting value of the stator current in the axial direction, It is the current moment k For the future moment k + i of q The predicted value of the stator current in the axial direction, α1, α2 and α3 are weight coefficients used to adjust the angular velocity, d The stator current in the axial direction and q The importance of the tracking error of the axial stator current in the comprehensive objective function; The control input variation term J 2 means the following: ; in, and The current moment k For the future moment k + i and k + i -1 d Predicted value of input voltage in the axial direction, and The current moment k For the future moment k + i and k + i -1 q Predicted value of input voltage in the axial direction, β 1 and β 2 is the weight coefficient, used to adjust d Axis and q The importance of axis direction control input variation in the comprehensive objective function.
[0029] S4. Determine the constraints of the high-voltage variable-frequency motor predictive control model: The constraint conditions include current constraint and voltage constraint; The current constraint is expressed as follows: ; in, i dmax yes d The maximum value of the stator current in the axial direction, i qmax yes q The maximum value of the stator current in the axial direction; The voltage constraint is expressed as follows: ; in, u dmax yes d The maximum value of the input voltage in the axial direction, u qmax yes q Maximum value of the input voltage in the axial direction.
[0030] S5. Solve the high-voltage variable-frequency motor predictive control model based on the quadratic programming algorithm: First, the comprehensive objective function is linearly transformed, and the control input sequence get: ; Expand J 1. J 2 About z The expression of the state variable oh ( i |k )and z The linear relationship is expressed as follows: ; in, P ij_1 are state variables in different time domains With input variables The correlation coefficient, Q ij_1 are state variables in different time domains With input variables The correlation coefficient of R i_1 is a state variable The independence coefficient of Then we have: ; For state variables Do the same processing as above to get: ; in, P ij_2 are state variables in different time domains With input variables The correlation coefficient, Q ij_2 are state variables in different time domains With input variables The correlation coefficient of R i_2 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ; in, P ij_3 are state variables in different time domains With input variables The correlation coefficient, Q ij_3 are state variables in different time domains With input variables The correlation coefficient of R i_3 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ; in, P ij_4 are state variables in different time domains With input variables The correlation coefficient, Qij_4 are state variables in different time domains With input variables The correlation coefficient of R i_4 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ; in, P ij_5 are state variables in different time domains With input variables The correlation coefficient, Q ij_5 are state variables in different time domains With input variables The correlation coefficient of R i_5 is a state variable The independence coefficient of Then the comprehensive objective function can be written as: ; in, H is a symmetric positive definite matrix whose elements are J In the expanded form z The quadratic coefficient of the element is determined, f is a vector whose elements are J In the expanded form z The coefficients of the first-order terms of the elements are determined; for example, assuming u d The elements in z is an even index, then for z In u d ( m | k )and u q ( n | k )( m , n =0, … , N c -1), H The corresponding element in H (2m,2n) Is J In the expanded u d ( m | k )and u q ( n | k) is determined by the coefficient of z In u d ( m | k ), f The corresponding element in f 2m Is J In the expanded u d ( m | k ) is determined by the coefficient of .
[0031] Then, the current constraint and voltage constraint are linearly transformed respectively, and the following is obtained: ; in, A dj_1 is the state variable in the constraint i d ( k ) and input variables u d ( j | k ), B dj_1 is the state variable in the constraint i d ( k ) and input variables u q ( j | k ), C d_1 is the state variable in the constraint i d ( k ) is the independent coefficient, A dj_2 is the state variable in the constraint i q ( k ) and input variables u d ( j | k ), B dj_2 is the state variable in the constraint i q ( k ) and input variables u q ( j | k ), C d_2 is the state variable in the constrainti q ( k ) is the independent coefficient, A dj_3 is the state variable in the constraint u d ( k ) and input variables u d ( j | k ), B dj_3 is the state variable in the constraint u d ( k ) and input variables u q ( j | k ), C d_3 is the state variable in the constraint u d ( k ) is the independent coefficient, A dj_4 is the state variable in the constraint u q ( k ) and input variables u d ( j | k ), B dj_4 is the state variable in the constraint u q ( k ) and input variables u q ( j | k ), C d_4 is the state variable in the constraint u q ( k )’s independence coefficient; Then the constraints can be written as: ; Writing the above formula in matrix form, we can get: ; in, A m is the constraint matrix, whose row vectors are transformed from constraints into z The coefficient composition of the linear inequality is: bis the constraint vector, whose elements are determined by the constraints and the constant terms in the predictive control model; Finally, the high-voltage variable-frequency motor predictive control model is solved: 1) Set the initial feasible solution z 0 : z 0 is the control input sequence z = U ( k ) and satisfy all constraints ; 2) Set the convergence accuracy , step size parameter α and barrier parameters m 0 ; 3) Update solution: ; Among them, 0< α <1, , z k and z k+1 are three adjacent solutions during the iteration process; 4) Update barrier parameters: ; in, r is the update factor, 0< r <1; 5) Check whether the convergence conditions are met: Calculate the current solution z k+1 With the previous solution z k The difference between J ( z k+1 ) and the comprehensive objective function value of the previous iteration J ( z k ), when: ; When , the algorithm is considered to have converged and the iteration is stopped; otherwise, the next iteration is continued; 6) Obtain the optimal solution: When the algorithm converges, the solution This is the desired optimal control input sequence.
[0032] In this embodiment, .
[0033] S6. Apply the optimal voltage to the high-voltage variable-frequency motor based on the solution result: At the current sampling time k, the optimal control input sequence The first element in Applied to the high voltage variable frequency motor as input voltage; At the next sampling time k+1, the actual state x(k+1) of the motor is obtained, and steps S1-S6 are repeated to continuously update the predictive control model and the comprehensive objective function, solve the optimization problem and apply control input to achieve dynamic control of the motor operation process.
[0034] Example 2 In order to verify the correctness and effectiveness of the method of the present invention, a simulation model of a high-voltage variable-frequency motor was built in MATLAB / Simulink; based on this simulation program, we obtained the following Figure 2 The motor speed changes with time as shown in Figure 3 The motor control input changes with time as shown. Through these two figures, we can observe the effect of the high-voltage variable-frequency motor starting process based on model predictive control.
[0035] In this embodiment, the motor control input is only responsible for the motor starting process and adjusting the current operating state of the motor. When the motor enters a stable operating state, the working voltage for maintaining the normal operation of the motor is provided by another structure, which has nothing to do with the content of the present invention. At this time, the motor control input voltage mentioned in the present invention is zero.
[0036] exist Figure 2 In the figure, the solid line represents the actual speed of the motor, and the dotted line represents the target speed. It is expected that the speed of the motor can smoothly follow the change of the target speed. The initial speed of the motor is 0, which means that the motor starts from a stationary state. In the first few moments of the simulation, the motor speed begins to increase rapidly, mainly because MPC calculates the future control input based on the current state (initially stationary) to gradually make the speed reach the target value.
[0037] The control input (voltage / frequency) usually refers to the output of the motor inverter, which determines the speed and acceleration of the motor. The control input will change over time, and since the comprehensive objective function of MPC includes a penalty for changes in the control input (i.e. J 2), so the system will try to avoid excessive control input fluctuations after the motor enters a steady state.
[0038] exist Figure 3In the process, the motor starts from rest. MPC calculates the initial value of the control input as zero according to the current state at the initial moment, and gradually increases the voltage / frequency through the control command signal to start the motor. Based on MPC, the control input is dynamically adjusted through the optimization process so that the motor speed gradually approaches the target speed; when the motor speed approaches the target value, the amplitude of the control input decreases. When the motor speed reaches the target value, the control input should drop to zero and subsequent changes tend to be stable. That is, ideally, the motor speed is stable near the target value, and the control input remains zero until the motor's operating state changes or the target value is adjusted. According to the settings in the program, the control input is limited to the theoretical range. This ensures that the voltage or frequency does not exceed the physical limitations of the motor.
[0039] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any ordinary technician in the industry can smoothly implement the present invention as shown in the drawings and described above. However, any equivalent changes, modifications and evolutions made by technicians familiar with the profession without departing from the scope of the technical solution of the present invention using the technical content disclosed above are all equivalent embodiments of the present invention. At the same time, any equivalent changes, modifications and evolutions made to the above embodiments based on the essential technology of the present invention are still within the protection scope of the technical solution of the present invention.
Claims
1. A method for starting and controlling a high-voltage variable-frequency motor based on model predictive control, characterized in that: The following steps are involved: S1, establish high voltage variable frequency motor The equation of state in the rotating coordinate system; S2, discretize the state equation to obtain the state prediction matrix of the high-voltage variable frequency motor; S3. Based on the state prediction matrix, a comprehensive objective function of the high-voltage variable frequency motor predictive control model is constructed; S4, determining the constraint conditions of the high-voltage variable-frequency motor predictive control model; S5. Solving the high-voltage variable-frequency motor predictive control model based on a quadratic programming algorithm; S6. Apply the optimal voltage to the high-voltage variable-frequency motor based on the solution result.
2. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 1 is characterized in that: The step S1 comprises the following steps: By establishing the state equation of the high-voltage variable-frequency motor, we can obtain: ； in, A is the system matrix, B is the input matrix, yes The derivative of is the state variable of the motor, yes d The stator current in the axial direction, yes q The stator current in the axial direction, ω is the motor angular velocity, yes d Input voltage in the axial direction, yes q Input voltage in the axial direction, T is the transposed matrix; Under continuous time conditions, the motor is k The state equation is expressed as follows: ； ； in, yes d The rotor current in the axial direction, yes q The rotor current in the axial direction, is the resistance of the stator winding, is the resistance of the rotor winding, is the stator inductance, is the rotor inductance, It is mutual induction. is the number of pole pairs of the motor, J is the moment of inertia of the motor, B is the coefficient of friction.
3. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 2 is characterized in that: The step S2 comprises the following steps: Set the prediction time domain N p and control time domain N c ,and N c ≤ N p , the state equation is discretized using the first-order Euler discretization method, and we can get: ； in, T s is the sampling time; In the prediction time domain, the discretized state equation is iterated to obtain: ； Writing the above formula in matrix form, we can get: ； in, It is the current moment k For the future moment k+i The predicted value of the motor state, is the state transition matrix, is the input matrix, is the control input sequence, which includes k To the future k + N c The predicted value of the input voltage is -1.
4. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 3 is characterized in that: The step S3 comprises the following steps: The comprehensive objective function J Including tracking error J 1 and control input change J 2, which means as follows: ； in, λ is the weight coefficient that balances tracking performance and control input stability; The tracking error term J 1 means the following: ； in, ω ref ( i | k ) is the current moment k For the future moment k + i The target setting value of the angular velocity, It is the current moment k For the future moment k + i The predicted value of the angular velocity, i dref ( i | k ) is the current moment k For the future moment k + i of d The target setting value of the stator current in the axial direction, It is the current moment k For the future moment k + i of d The predicted value of the stator current in the axial direction, i qref ( i | k ) is the current moment k For the future moment k + i of q The target setting value of the stator current in the axial direction, It is the current moment k For the future moment k + i of q The predicted value of the stator current in the axial direction, α1, α2 and α3 are weight coefficients; The control input variation term J 2 means the following: ； in, and The current moment k For the future moment k + i and k + i -1 d Predicted value of input voltage in the axial direction, and The current moment k For the future moment k + i and k + i -1 q Predicted value of input voltage in the axial direction, β 1 and β 2 is the weight coefficient.
5. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 4 is characterized in that: The step S4 comprises the following steps: The constraint conditions include current constraint and voltage constraint; The current constraint is expressed as follows: ； in, i dmax yes d The maximum value of the stator current in the axial direction, i qmax yes q The maximum value of the stator current in the axial direction; The voltage constraint is expressed as follows: ； in, u dmax yes d The maximum value of the input voltage in the axial direction, u qmax yes q Maximum value of the input voltage in the axial direction.
6. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 5 is characterized in that: The step S5 comprises the following steps: First, the comprehensive objective function is linearly transformed, and the control input sequence get: ； Expand J 1. J 2 About z The expression of the state variable ω ( i | k )and z The linear relationship is expressed as follows: ； in, P ij_1 are state variables in different time domains With input variables The correlation coefficient, Q ij_1 are state variables in different time domains With input variables The correlation coefficient of R i_1 is a state variable The independence coefficient of Then we have: ； For state variables Do the same processing as above to get: ； in, P ij_2 are state variables in different time domains With input variables The correlation coefficient, Q ij_2 are state variables in different time domains With input variables The correlation coefficient of R i_2 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ； in, P ij_3 are state variables in different time domains With input variables The correlation coefficient, Q ij_3 are state variables in different time domains With input variables The correlation coefficient of R i_3 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ； in, P ij_4 are state variables in different time domains With input variables The correlation coefficient, Q ij_4 are state variables in different time domains With input variables The correlation coefficient of R i_4 is a state variable The independence coefficient of For state variables Do the same processing as above to get: ； in, P ij_5 are state variables in different time domains With input variables The correlation coefficient, Q ij_5 are state variables in different time domains With input variables The correlation coefficient of R i_5 is a state variable The independence coefficient of Then the comprehensive objective function can be written as: ； in, H is a symmetric positive definite matrix whose elements are J In the expanded form z The quadratic coefficient of the element is determined, f is a vector whose elements are J In the expanded form z The coefficient of the first-order term of the element is determined; Then, the current constraint and voltage constraint are linearly transformed respectively, and the following is obtained: ； in, A dj_1 is the state variable in the constraint i d ( k ) and input variables u d ( j | k ), B dj_1 is the state variable in the constraint i d ( k ) and input variables u q ( j | k ), C d_1 is the state variable in the constraint i d ( k ) is the independent coefficient, A dj_2 is the state variable in the constraint i q ( k ) and input variables u d ( j | k ), B dj_2 is the state variable in the constraint i q ( k ) and input variables u q ( j | k ), C d_2 is the state variable in the constraint i q ( k ) is the independent coefficient, A dj_3 is the state variable in the constraint u d ( k ) and input variables u d ( j | k ), B dj_3 is the state variable in the constraint u d ( k ) and input variables u q ( j | k ), C d_3 is the state variable in the constraint u d ( k ) is the independent coefficient, A dj_4 is the state variable in the constraint u q ( k ) and input variables u d ( j | k ), B dj_4 is the state variable in the constraint u q ( k ) and input variables u q ( j | k ), C d_4 is the state variable in the constraint u q ( k )’s independence coefficient; Then the constraints can be written as: ； Writing the above formula in matrix form, we can get: ； in, A m is the constraint matrix, whose row vectors are transformed from constraints into z The coefficient composition of the linear inequality is: b is a constraint vector whose elements are determined by the constraints and constant terms in the predictive control model.
7. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 6 is characterized in that: The step S5 further comprises the following steps: 1) Set the initial feasible solution z 0 : z 0 is the control input sequence z = U ( k ) and satisfy all constraints ; 2) Set the convergence accuracy , step size parameter α and barrier parameters μ 0 ; 3) Update solution: ； Among them, 0< α <1, , z k and z k+1 are three adjacent solutions during the iteration process; 4) Update barrier parameters: ； in, ρ is the update factor, 0< ρ <1; 5) Check whether the convergence conditions are met: Calculate the current solution z k+1 With the previous solution z k The difference between J ( z k+1 ) and the comprehensive objective function value of the previous iteration J ( z k ), when: ； When , the algorithm is considered to have converged and the iteration is stopped; otherwise, the next iteration is continued; 6) Get the optimal solution: When the algorithm converges, the solution This is the desired optimal control input sequence.
8. The high-voltage variable-frequency motor starting and control method based on model predictive control according to claim 7 is characterized in that: The step S6 comprises the following steps: At the current sampling time k , the optimal control input sequence The first element in Applied to the high voltage variable frequency motor as input voltage; At the next sampling moment k +1, get the actual state of the motor x ( k +1), repeat steps S1-S6, continuously update the predictive control model and the comprehensive objective function, solve the optimization problem and apply control input to achieve dynamic control of the motor operation process.