Sensorless control method of staggered magnetic pole series-parallel magnetic circuit doubly salient bearingless motor

Abstract

The invention discloses a sensorless control method for a staggered magnetic pole series-parallel magnetic circuit doubly salient bearingless motor, and the method mainly comprises the steps: introducing a catfish effect, and based on a conventional particle swarm algorithm, providing an improved catfish effect quantum particle swarm algorithm. Therefore, the problem of simultaneous detection of two rotor space positions of the tangential position and the radial displacement signal of the rotor of the staggered magnetic pole series-parallel magnetic circuit doubly salient bearingless motor is solved. The method overcomes the limitation of a mechanical sensor on the increase of the critical rotating speed, reduces the manufacturing cost of the system, has better capability of quickly locking a signal target and quickly converging a function compared with the traditional particle swarm algorithm, and improves the signal detection efficiency.

Description

Technical Field

[0001] The invention belongs to the field of electric drive control equipment, and particularly relates to a sensorless control method for an interleaved pole hybrid magnetic circuit doubly salient bearingless motor. Background Technique

[0002] The interleaved pole hybrid magnetic circuit doubly salient bearingless motor is a new type of motor integrating magnetic bearing technology and doubly salient motor technology. It not only has excellent characteristics such as no mechanical friction, no need for lubrication, low noise, and long service life of magnetic bearings, but also has characteristics such as high torque (power) density, wide speed regulation ability, and high operation reliability of doubly salient motors, and has broad application prospects in fields such as aerospace, life science, and semiconductor manufacturing. When the interleaved pole hybrid magnetic circuit doubly salient bearingless motor is applied to industrial control, the accurate detection of the rotor spatial position (tangential position and radial displacement) is a key link for the stable operation of the motor. In the past, mechanical sensors were installed at the end of the motor to obtain the rotor tangential position and radial displacement signals, and then the two signals were used to control the rotation and suspension of the motor. However, installing mechanical sensors in the motor hinders the reliable operation and integrated development, limits the improvement of the critical speed, and increases the system manufacturing cost.

[0003] As a new type of bearingless motor, there are few relevant studies on the sensorless accurate estimation methods for the two rotor spatial positions of the rotor tangential position and radial displacement signals of the interleaved pole hybrid magnetic circuit doubly salient bearingless motor. Some literature has proposed sensorless control methods for bearingless motors using neural network methods and Kalman filter methods, but there are almost no sensorless control methods for bearingless motors using particle swarm algorithms. Therefore, there is an urgent need for relevant research on the accurate estimation method for the rotor spatial position of the interleaved pole hybrid magnetic circuit doubly salient bearingless motor and its sensorless control. Summary of the Invention

[0004] Object of the Invention: The object of the invention is to provide a sensorless control method for an interleaved pole hybrid magnetic circuit doubly salient bearingless motor. Based on the traditional particle swarm algorithm, by introducing the "catfish effect", an improved catfish effect quantum particle swarm algorithm is invented to solve the problem of simultaneous detection of the rotor tangential position and radial displacement signals of the interleaved pole hybrid magnetic circuit doubly salient bearingless motor.

[0005] Technical Solution: A sensorless control method for an interleaved pole hybrid magnetic circuit doubly salient bearingless motor of the invention comprises the following steps:

[0006] Step 1, based on the rotor tangential position, radial displacement signals and double power sliding mode reaching control input signals in the interleaved pole hybrid magnetic circuit doubly salient bearingless motor, construct a system state variable equation;

[0007] Step 2: Construct a double-power sliding mode function to obtain the double-power sliding mode reaching control input signal;

[0008] Step 3: Represent the double-power sliding mode reaching control input signal through a quantum particle swarm optimization algorithm;

[0009] Step 4: Introduce a catfish operator into the quantum particle swarm optimization algorithm, establish a rotor spatial position prediction model for the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit, and predict the spatial position of the rotor in the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit through the rotor spatial position prediction model.

[0010] Furthermore, in Step 1, the system state variable equation is:

[0011]

[0012] where \(w(t)\) is the electromagnetic energy storage in the air gap of the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit, \(f(w(t))\) and \(g(w(t))\) are respectively the extraction of the rotor tangential position and radial displacement signals, is the derivative of \(w(t)\), and \(u\) is the double-power sliding mode reaching control input signal.

[0013] Furthermore, Step 2 is specifically: in the bearingless doubly salient motor system with interleaved pole hybrid magnetic circuit, the goal is to design the double-power sliding mode reaching control input signal \(u\) so that the system state variable equation is stable and the stable error \(e\) converges quickly to the equilibrium point, and the rotor spatial position of the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit will be the same as the required rotor spatial position, where is the estimated value of the electromagnetic energy storage;

[0014] Construct the double-power sliding mode function to eliminate the uncertain error value and external interference of the control system. The double-power sliding mode function is as follows:

[0015]

[0016] where \(l_1\), \(l_2\), \(\alpha_1\), \(\alpha_2\) are all constants, \(l_1\gt0\), \(l_2\gt0\), \(\alpha_1\gt1\), \(0\lt\alpha_2\lt1\), is the derivative of \(\sigma\);

[0017] Control the rotor spatial position to make the control system state reach the sliding mode surface within a finite time and complete signal detection. Therefore, the double-power sliding mode reaching control input signal needs to be obtained:

[0018]

[0019] In Equation (3), sign() represents the sign function, where is a constant, and arctan() represents the arctangent function; therefore, Equation (3) will make the state sum converge in finite time respectively.

[0020] Moreover, by designing a sliding mode reaching law with double power orders for Equations (2) and (3), in addition to obtaining the finite system state convergence time, it can further reduce the chattering and steady-state error problems, and then make the control more accurate and ensure the system stability.

[0021] Furthermore, Step 3 is specifically as follows:

[0022] Express the input signal of the double power order sliding mode approaching control using the quantum particle swarm optimization algorithm as:

[0023]

[0024]

[0025] where β k is the compression / expansion factor, Q g and τ k are random numbers uniformly distributed on (0, 1), Q k is the average value of the regional best positions in the search processes of all particles in the swarm. Updating all the particles in the particle swarm corresponding to the rotor spatial position once is called one round of iteration Q g , and the current velocity of the particle is determined by three factors: the velocity at the previous moment the individual extreme value and the global extreme value When the algorithm shows the premature phenomenon, the global extreme value is the local optimal solution. Therefore, by changing the global extreme value or changing the individual extreme value to make the particle escape from the local optimal solution region and enter other regions for search, and finally find the global optimal solution of the rotor spatial position.

[0026] Furthermore, Step 4 is specifically as follows: Introduce the catfish operator c1·rand() to perturb the global extreme value or the individual extreme value , establish a prediction model for the rotor spatial position of the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit, and update the particle formulas (6) and (7) corresponding to the rotor tangential position and radial displacement signals respectively

[0027]

[0028]

[0029] where ω is the inertia weight; c1 represents the perturbation intensity of the catfish on the individual extreme value and the global optimal extreme value; rand() is a random number between (0, 1); and are respectively the tangential position and the radial displacement signal of the rotor in the d-th dimension during the k-th iteration of the particle, and are respectively the tangential position and the radial displacement signal of the rotor in the d-th dimension during the (k + 1)-th iteration of the particle, is the position of the individual extreme value of the particle in the d-th dimension, is the position of the global extreme value of the population in the d-th dimension;

[0030] When the search particle swarm of the catfish effect quantum particle swarm algorithm still does not evolve after multiple iterations, it is determined that it has fallen into the regional optimal solution. At this time, the worst 10% of all particles will be removed, and then catfish particles will be introduced. According to formulas (6) and (7), the tangential position and the radial displacement signal of the rotor of the particles are updated. These catfish particles will find better solutions, and these catfish particles will guide all the particles to a new area near the optimal solution. Finally, it is judged whether the termination condition is satisfied. When the termination condition is reached, the particles will converge, and the final solution will be output, and the algorithm ends.

[0031] Beneficial effects: Compared with the prior art, the present invention has the following remarkable advantages:

[0032] 1. The sensorless control method for the interleaved-pole hybrid magnetic circuit doubly salient bearingless motor based on the improved catfish effect quantum particle swarm algorithm of the present invention uses the improved catfish effect quantum particle swarm algorithm to replace the traditional particle swarm algorithm. It has higher precision, lower cost and better computing performance. For a control object with relatively prominent non-linear characteristics such as the interleaved-pole hybrid magnetic circuit doubly salient bearingless motor, the advantages of the improved catfish effect quantum particle swarm algorithm can be significantly seen. This method can quickly realize the effective identification of the regional maximum value and the global maximum value of the interleaved-pole hybrid magnetic circuit doubly salient bearingless motor, and effectively obtain the tangential position and the radial displacement signal of the rotor. This method also reduces the accuracy requirements for the multi-physical field model of the interleaved-pole hybrid magnetic circuit doubly salient bearingless motor, saves the time required for information discrimination, and improves the efficiency of signal detection.

[0033] 2. When the number of iterations approaches infinity, the traditional particle swarm clustering algorithm cannot accurately estimate the rotor spatial position at a relatively fast speed. In addition, there are limitations in the speed and search range of a single particle of the traditional particle swarm clustering algorithm, and it cannot search the entire spatial position of the tangential position and the radial displacement of the rotor, which limits the global search ability of the algorithm. The improved catfish effect quantum particle swarm algorithm can improve the above disadvantages, quickly and accurately track the rotor spatial position of the interleaved-pole hybrid magnetic circuit doubly salient bearingless motor, and has the ability of fast online learning and convergence.

[0034] 3. Compared with the traditional particle swarm clustering algorithm, the improved catfish effect quantum particle swarm algorithm has a completely random iterative equation. When particles move in the quantum space, they can change their trajectories according to the following situation. The entire iterative equation has fewer parameters and is easier to control, enabling the particles to track the optimal solution at the rotor space position of the global feasible solution. Description of the Drawings

[0035] Figure 1 is the structural diagram of a bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits;

[0036] Figure 2 is the operation diagram of a bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits at 1500 revolutions per minute;

[0037] Figure 3 is the back electromotive force observation diagram of a bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits. Detailed Implementation Manner

[0038] The technical solution of the present invention will be further described below in conjunction with the drawings.

[0039] The present invention is realized through the following technical solutions: A sensorless control method for a bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits is to establish a model of the bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits, and improve the quantum particle swarm algorithm in combination with the catfish effect, including the following steps:

[0040] The system state variable equation involved in a bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits is as follows:

[0041]

[0042] Among them, w(t) is the electromagnetic energy storage in the air gap of the bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits, f(w(t)) and g(w(t)) are respectively the extraction of the rotor tangential position and radial displacement signals, is the derivative of w(t), and u is the double-power sliding mode reaching control input signal.

[0043] In the bearingless doubly salient motor system with interleaved magnetic poles and mixed magnetic circuits, the goal is to design the double-power sliding mode reaching control input signal u to make Equation (1) stable and the stable error e quickly converge to the equilibrium point, and the rotor space position of the bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits will be the same as the required rotor space position. Among them is the estimated value of the electromagnetic energy storage,

[0044] Due to the uncertain error values of the control system and external interference, the rotor spatial position of the interleaved pole hybrid magnetic circuit doubly salient bearingless motor is affected, making it difficult to accurately estimate. Therefore, a double-power sliding mode function is constructed as follows

[0045]

[0046] where l1, l2, α1, and α2 are all constants, with l1 > 0, l2 > 0, α1 > 1, and 0 < α2 < 1 is the derivative of σ. By controlling the rotor spatial position, the state of the control system can reach the sliding mode surface within a finite time to complete signal detection. Therefore, it is necessary to obtain the double-power sliding mode reaching control input signal

[0047]

[0048] In Equation (3), sign() represents the sign function, where is a constant, and arctan() represents the arctangent function. Therefore, Equation (3) will cause the state and to converge in a finite time respectively. Moreover, by designing a double-power sliding mode reaching law for Equations (2) and (3), in addition to obtaining the finite system state convergence time, the chattering and steady-state error problems can be reduced, making the control more accurate and ensuring system stability. The double-power sliding mode reaching control input signal is expressed using the quantum particle swarm optimization algorithm as

[0049]

[0050]

[0051] where β k is the compression / expansion factor, Q g and τ k are random numbers uniformly distributed on (0, 1), and Q k is the average value of the regional best positions in the search history of all particles in the swarm. Updating all particles in the particle swarm corresponding to the rotor spatial position once is called one round of iteration Q g . The current velocity of the particle is determined by three factors: the velocity at the previous moment the individual extreme value and the global extreme value Once the algorithm shows the phenomenon of premature convergence, the global extreme value must be a local optimal solution. Therefore, by changing the global extreme value or indirectly changing the individual extreme value the particle can escape from the local optimal solution region and enter other regions for search, finally finding the global optimal solution of the rotor spatial position

[0052] Introduce the catfish operator c1·rand() to the global extreme value or individual extreme value Perturbations are carried out to establish a prediction model for the rotor spatial position of a bearingless doubly salient motor with interleaved pole hybrid magnetic circuit, and update the particle formulas (6) and (7) corresponding to the rotor tangential position and radial displacement signals respectively

[0053]

[0054]

[0055] where ω is the inertia weight; c1 represents the perturbation intensity of the catfish to the individual extreme value and the global optimal extreme value; rand() is a random number between (0,1); and are the rotor tangential position and radial displacement signals of the d-th dimension of the particle in the k-th iteration respectively, and are the rotor tangential position and radial displacement signals of the d-th dimension of the particle in the (k + 1)-th iteration respectively, is the position of the individual extreme value of the particle in the d-th dimension, is the position of the global extreme value of the population in the d-th dimension.

[0056] When the catfish effect quantum particle swarm algorithm determines that the search particle swarm has not evolved after multiple iterations, that is, it is judged to fall into the regional optimal solution. At this time, the worst 10% of all particles will be removed, and then catfish particles will be introduced. The rotor tangential position and radial displacement signals of the particles are updated according to formulas (6) and (7). These catfish particles will find better solutions, and these catfish particles will guide all particles to a new area near the optimal solution. Finally, it is judged whether the termination condition is satisfied. When the termination condition is reached, the particles will converge, and the final solution will be output, and the algorithm ends.

[0057] Figure 1 is the cross-sectional view of the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit; Figure 2 is the operation diagram of the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit at a speed of 1500 revolutions per minute; Figure 3 is the back electromotive force observation diagram of the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit.

Claims

1. A sensorless control method for an interleaved pole hybrid magnetic circuit doubly salient bearingless motor, characterized in that, It includes the following steps: Step 1: Based on the tangential position of the rotor, the radial displacement signal and the double-power sliding mode reaching control input signal in the switched reluctance bearingless motor with interleaved poles and mixed magnetic circuits, construct the system state variable equation; Step 2: Construct the double-power sliding mode function to obtain the double-power sliding mode reaching control input signal; Step 3: Represent the double-power sliding mode reaching control input signal by the quantum particle swarm optimization algorithm; Step 4: Introduce the catfish operator into the quantum particle swarm optimization algorithm, establish the rotor spatial position prediction model for the switched reluctance bearingless motor with interleaved poles and mixed magnetic circuits, and predict the spatial position of the rotor in the switched reluctance bearingless motor with interleaved poles and mixed magnetic circuits through the rotor spatial position prediction model.

2. The sensorless control method of a switched reluctance bearingless motor with interleaved magnetic poles and hybrid magnetic circuit according to claim 1, characterized in that, In Step 1, the system state variable equation is: Among them, \(w(t)\) is the electromagnetic energy storage in the air gap of the bearingless doubly salient motor with interleaved magnetic poles and mixed magnetic circuits, \(f(w(t))\) and \(g(w(t))\) are respectively the extraction of the rotor tangential position and radial displacement signals, is the derivative of \(w(t)\), and \(u\) is the input signal of the bi-power sliding mode reaching control.

3. The sensorless control method of a bearingless doubly salient motor with interleaved magnetic poles and hybrid magnetic circuit according to claim 1, characterized in that, Step 2 is specifically as follows: In the interleaved pole hybrid magnetic circuit doubly salient bearingless motor system, the goal is to design the double power sliding mode reaching control input signal u to make the system state variable equation stable and the stable error e quickly converge to the equilibrium point. The rotor spatial position of the interleaved pole hybrid magnetic circuit doubly salient bearingless motor will be the same as the required rotor spatial position, where is the estimated value of the electromagnetic energy storage; Construct a double-power sliding mode function to eliminate the uncertain error value and external interference of the control system. The double-power sliding mode function is as follows: where \(l_1\), \(l_2\), \(\alpha_1\), \(\alpha_2\) are all constants, \(l_1\gt0\), \(l_2\gt0\), \(\alpha_1\gt1\), \(0\lt\alpha_2\lt1\). is the derivative of \(\sigma\); Control the rotor spatial position to make the control system state reach the sliding mode surface within a finite time and complete signal detection. Therefore, it is necessary to obtain the double-power sliding mode reaching control input signal: In Equation (3), sign() represents the sign function, where is a constant, and arctan() represents the arctangent function; therefore, Equation (3) will make the state and converge finitely in finite time, respectively.

4. A sensorless control method for a bearingless doubly salient motor with interleaved magnetic poles and hybrid magnetic circuit according to claim 1, characterized in that Step 3 is specifically: Represent the double-power sliding mode reaching control input signal by the quantum particle swarm optimization algorithm as: Among them, β k is the compression / expansion factor, Q g and τ k are random numbers uniformly distributed on (0, 1), Q k is the average value of the regional best positions in the search processes of all particles in the population. Updating all particles in the particle swarm corresponding to the rotor spatial position once is called one round of iteration. g , and the current velocity of the particle is determined by three factors: the velocity at the previous moment personal extreme value and the global extreme value When the premature phenomenon appears in the algorithm, the global extreme value is the local optimal solution. Therefore, by changing the global extreme value or changing the personal extreme value to make the particle escape from the local optimal solution region and enter other regions for search, and finally find the global optimal solution of the rotor spatial position.

5. A sensorless control method for a bearingless doubly salient motor with interleaved magnetic poles and hybrid magnetic circuits according to claim 1, characterized in that, Step 4 specifically is: introducing the catfish operator c1·rand() to perturb the global extreme value or the individual extreme value to establish a rotor spatial position prediction model for the bearingless doubly salient motor with interleaved pole hybrid magnetic circuit, and update the particle formulas (6) and (7) corresponding to the rotor tangential position and radial displacement signals respectively where ω is the inertia weight; c1 represents the perturbation intensity of the catfish on the individual extreme value and the global optimal extreme value; rand() is a random number between (0, 1); and are the tangential position and the radial displacement signal of the rotor in the d-th dimension at the k-th iteration of the particle, and are the tangential position and the radial displacement signal of the rotor in the d-th dimension at the (k + 1)-th iteration of the particle, is the position of the individual extreme value of the particle in the d-th dimension, is the position of the global extreme value of the population in the d-th dimension; When the catfish effect quantum particle swarm algorithm determines that the search particle swarm has not evolved after multiple iterations, that is, it is judged to fall into the local optimal solution. At this time, the worst 10% of all particles will be removed, and then catfish particles are introduced. According to formulas (6) and (7), update the tangential position and radial displacement signals of the rotor of the particles. These catfish particles will find better solutions, and these catfish particles will guide all the particles to a new area near the optimal solution. Finally, judge whether the termination condition is satisfied. When the termination condition is reached, the particles will converge, and the final solution will be output, and the algorithm ends.

Images