A speed control method applicable to a magnetic levitation pump at low speeds

Abstract

The present invention discloses a rotational speed control method applicable to a magnetic levitation pump at low rotational speeds. Based on a resonant controller, the present invention adaptively controls the resonant frequency point and the resonant gain respectively, and introduces a compensation phase angle to achieve adaptive resonant control, compensating for system delay and thus improving system stability. By continuously controlling the actual rotational speed of the rotor, the actual rotational speed of the rotor is maintained at the target rotational speed. The present invention correspondingly improves the rotational performance of the magnetic levitation pump at low speeds from two aspects of measurement accuracy and control mode. The accuracy of the sensor measurement value is improved by EKF filtering, and the periodic fluctuation of the rotational speed is effectively reduced by an adaptive PIR controller. The combination of the two comprehensively optimizes the rotational characteristics of the magnetic levitation pump at low speeds.

Description

Technical Field

[0001] The present invention relates to the field of magnetic levitation pump control, and particularly to a speed control method applicable to low speeds of a magnetic levitation pump. Background Art

[0002] With the vigorous development of industries such as nanotechnology, semiconductor manufacturing processes, and medical devices, the demand for ultra-clean, low particle numbers, continuous flow, or smooth fluid medium transfer in the high-purity market is becoming increasingly urgent. The magnetic levitation motor has a compact structure, and its design without mechanical contact ensures that only very few particulate matters are generated during its movement. In addition, due to its low shear stress, it can ensure gentle transportation of sensitive fluids, and the transportation process is smooth and pulsation-free, so it has been widely used.

[0003] However, due to the absence of bearings, the control of the rotor is more complex. Especially at low speeds, it is easily affected by cogging torque, current harmonics, manufacturing and installation errors. In addition, there are measurement errors and feedback errors caused by insufficient sensor accuracy, resulting in uneven rotation speed or rotor jitter, which affects the transportation efficiency and motor performance. At high speeds, due to the effect of inertia, the rotation speed can still remain relatively stable. However, at low speeds, the fluctuations of current and measurement errors will have a more obvious impact on the rotation speed fluctuations.

[0004] Currently, the traditional PID control is used for the speed control of magnetic levitation pumps. Although this control method is simple and easy to adjust, it is difficult to handle the situation affected by periodic disturbances, and it is easily affected by various factors during actual control, with uncertainty and time-variation. According to the internal model principle, to accurately track a periodic reference or completely suppress a periodic disturbance, an internal model of the periodic reference or periodic disturbance must be established in the controller. Therefore, a new control method needs to be considered to reduce the periodic disturbance. Summary of the Invention

[0005] To solve the problems in the background art, the present invention provides a speed control method applicable to low speeds of a magnetic levitation pump. More accurate current, angle, and calculated speed values in the present invention are used as inputs to the control system, reducing the errors caused by insufficient sensor accuracy or time delay. Through an adaptive PIR controller, the parameters are automatically adjusted according to different speeds, effectively suppressing the harmonic components generated in the current and the periodic disturbance torques received by the rotor and other disturbances.

[0006] The technical solution adopted by the present invention is:

[0007] 1) After the magnetic levitation pump starts, after measuring the angle and rotational speed of the rotor in the magnetic levitation pump and the current values flowing through the a and b phases of the driving windings of the magnetic levitation pump by sensors, the angle and rotational speed of the rotor are respectively converted into electrical angles and electrical rotational speeds, and the electrical angle, electrical rotational speed, and the current values of the a and b phases are corrected according to the extended Kalman filtering algorithm to obtain the corrected electrical angle, electrical rotational speed, and the current values of the a and b phases;

[0008] 2) Using the corrected electrical angle, the corrected current values of the a and b phases are converted into the actual current values of the driving d and q axes of the rotor through Park transformation, and the corrected electrical rotational speed is converted into the mechanical rotational speed and used as the actual rotational speed of the rotor;

[0009] 3) Using the target rotational speed, actual rotational speed, actual current values of the driving d and q axes of the rotor, and the target current value, the target voltage of the driving d axis of the rotor and the target voltage of the driving q axis are respectively obtained through multiple adaptive PIR controllers;

[0010] 4) The target voltage of the driving d axis of the rotor, the target voltage of the driving q axis, and the corrected electrical angle are converted into the voltage values of the a and b phases through inverse Park transformation, and then the rotor of the magnetic levitation pump is driven and controlled through a digital-to-analog converter and a DC power amplifier to rotate;

[0011] 5) Repeat steps 1) to 4), and maintain the actual rotational speed of the rotor at the target rotational speed by continuously controlling the actual rotational speed of the rotor.

[0012] The specific content of step 3) is as follows: Set the target rotational speed of the rotor. After subtracting the actual rotational speed of the rotor from the target rotational speed, input the difference and the actual rotational speed of the rotor into the first PIR controller to obtain the target current value of the driving q axis of the rotor; Set the target current value of the driving d axis of the rotor to 0. After subtracting the target current value from the actual current value of the driving d axis of the rotor, input the difference and the actual rotational speed of the rotor into the second PIR controller to obtain the target voltage of the driving d axis of the rotor. At the same time, subtract the target current value from the actual current value of the driving q axis of the rotor, and input the difference and the actual rotational speed of the rotor into the third PIR controller to obtain the target voltage of the driving q axis of the rotor;

[0013] The first, second, and third PIR controllers mentioned above are all adaptive PIR controllers.

[0014] The transfer function G(s) of the adaptive PIR controller is expressed as follows:

[0015]

[0016] Among them, K r0 is the initial gain coefficient, c is the control coefficient, ω C is the bandwidth of the adaptive PIR controller, and s is the Laplace operator; Kp is the proportionality coefficient, K i is the integral coefficient, ω m is the actual rotational speed of the rotor is the compensation phase angle

[0017] After analyzing the system stability test and the Nyquist curves of the transfer function at different compensation angles and rotational speeds, the compensation phase angle is determined according to the actual rotational speed of the rotor. When the mechanical rotational speed of the rotor is less than or equal to 300 rpm, the main factor affecting the resonance control effect is the gain coefficient, and the phase compensation effect is very small at this time. The compensation phase angle is equal to 0 degrees. After the rotational speed increases, since the system stability is greatly affected by the motor speed, delay compensation must be carried out. That is, when the mechanical rotational speed of the rotor is greater than 300 rpm and less than or equal to 500 rpm, the compensation phase angle is equal to 20 degrees. When the mechanical rotational speed of the rotor is greater than 500 rpm, the compensation phase angle is equal to 30 degrees

[0018] The specific operation of inputting the difference value and the actual rotational speed of the rotor into the first PIR controller is as follows: The difference between the target rotational speed and the actual rotational speed of the rotor and the actual rotational speed of the rotor are simultaneously input through the proportional, integral, and resonance control links in the first PIR controller, and the target current value of the driving q-axis of the rotor is output

[0019] The adaptive PIR controller adaptively controls the resonance frequency point and the resonance gain respectively on the basis of the resonance controller and introduces the compensation phase angle It is obtained by paralleling the adaptive resonance controller and the PI controller, and then controls the rotational speed of the magnetic levitation pump, realizing adaptive resonance control, compensating for system delay, and improving system stability

[0020] The beneficial effects of the present invention are as follows

[0021] 1) Introducing the extended Kalman filter algorithm to obtain more accurate current, angle, and calculated rotational speed values as the input of the control system, reducing the errors caused by insufficient sensor accuracy or time delay

[0022] 2) Designing an adaptive PIR controller on the basis of the traditional resonance controller, automatically adjusting parameters according to different rotational speeds, effectively suppressing disturbances such as harmonic components generated in the current and periodic disturbance torques received by the rotor, and ensuring the uniform and stable rotation of the rotor

[0023] 3) Through the FOC strategy, the present invention makes the Maxwell force and the Lorentz force generated by the rotor in the same direction by applying current at the same moment, which can generate the maximum electromagnetic torque. In addition, the FOC strategy controls the torque and magnetic flux separately, which is convenient to control and provides a faster dynamic response. Description of the Drawings

[0024] Figure 1 Flow chart of estimating the rotational speed of the magnetic levitation pump by the extended Kalman filter algorithm.

[0025] Figure 2 PIR control block diagram of the current control loop and the speed control loop.

[0026] Figure 3 Schematic diagram of the rotational speed control system of the magnetic levitation pump. Detailed Implementation Manner

[0027] The present invention will be further described in detail below in conjunction with the drawings and specific embodiments.

[0028] As Figure 3 shown, the method includes the following steps:

[0029] 1) After the magnetic levitation pump starts, after measuring the angle and rotational speed of the rotor in the magnetic levitation pump and the current values flowing through the a and b phases of the drive windings of the magnetic levitation pump by sensors, the angle and rotational speed of the rotor are respectively converted into electrical angles and electrical rotational speeds. As Figure 1 shown, the electrical angle, electrical rotational speed, and current values of the a and b phases are corrected according to the extended Kalman filter algorithm to obtain the corrected electrical angle, electrical rotational speed, and current values of the a and b phases;

[0030] 2) Using the corrected electrical angle, the corrected current values of the a and b phases are converted into the actual current values of the drive d and q axes of the rotor through Park transformation, and the corrected electrical rotational speed is converted into the mechanical rotational speed and used as the actual rotational speed of the rotor;

[0031] 3) Set the target rotational speed of the rotor. After subtracting the target rotational speed of the rotor from the actual rotational speed, the difference and the actual rotational speed of the rotor are input into the first PIR controller to obtain the target current value of the drive q axis of the rotor;

[0032] 4) Set the target current value of the drive d axis of the rotor to 0. Subtract the actual current value of the drive d axis of the rotor from the target current value, and input the difference and the actual rotational speed of the rotor into the second PIR controller to obtain the target voltage of the drive d axis of the rotor. At the same time, subtract the actual current value of the drive q axis of the rotor from the target current value, and input the difference and the actual rotational speed of the rotor into the third PIR controller to obtain the target voltage of the drive q axis of the rotor;

[0033] 5) Convert the target voltages of the driving d-axis and q-axis of the rotor, as well as the corrected electrical angle, into the voltage values of phases a and b through the inverse Park transformation, and then drive and control the rotation of the rotor of the magnetic levitation pump through a digital-to-analog converter and a DC power amplifier;

[0034] 6) Repeat steps 1) to 5), and maintain the actual rotor speed at the target speed by continuously controlling the actual rotor speed.

[0035] Among them, the first, second, and third PIR controllers are all adaptive PIR controllers, and the transfer function G(s) of the adaptive PIR controller is expressed as follows:

[0036]

[0037] Among them, K r0 is the initial gain coefficient, c is the control coefficient, w C is the bandwidth of the adaptive PIR controller, s is the Laplace operator; K p is the proportional coefficient, K i is the integral coefficient, ω m is the actual speed of the rotor, is the compensation phase angle.

[0038] After analyzing the system stability test and the Nyquist curves of the transfer function at different compensation angles and speeds, the compensation phase angle is determined according to the actual speed of the rotor. When the mechanical speed of the rotor is less than or equal to 300 rpm, the main factor affecting the resonance control effect is the gain coefficient, and the effect of phase compensation is very small at this time, and the compensation phase angle is equal to 0 degrees; after the speed increases, since the system stability is greatly affected by the motor speed, delay compensation must be carried out, that is, when the mechanical speed of the rotor is greater than 300 rpm and less than or equal to 500 rpm, the compensation phase angle is equal to 30 degrees; when the mechanical speed of the rotor is greater than 500 rpm, the compensation phase angle is equal to 50 degrees.

[0039] In step 3), inputting the difference value and the actual speed of the rotor into the first PIR controller specifically means: inputting the difference between the target speed and the actual speed of the rotor and the actual speed of the rotor simultaneously through the proportional, integral, and resonance control links in the first PIR controller, and outputting the target current value of the driving q-axis of the rotor.

[0040] The adaptive PIR controller adaptively controls the resonance frequency point and the resonance gain respectively on the basis of the resonance controller and introduces the compensation phase angle It is obtained by paralleling an adaptive resonance controller and a PI controller, and then controls the speed of the magnetic levitation pump, achieving adaptive resonance control, compensating for system delay and thus improving system stability.

[0041] As Figure 2 shown, the difference between the rotational speed and the d / q-axis current is input to the controller. On the one hand, the input quantity passes through the PI controller, and on the other hand, the input quantity enters the resonance controller, whose resonance gain is The resonance bandwidth is ω c , and the resonance point is ω m , that is, it suppresses interference with a bandwidth of ω m above and below the frequency of ω c . After resonance control, it undergoes an angular compensation of a phase angle . The output result is added to the output result of the PI controller to obtain the target q-axis current / d and q-axis voltage.

[0042] The basic resonance controller is a resonance controller, and its transfer function is: This resonance controller only acts on a single frequency and has extremely strong frequency selectivity. However, in fact, due to the uncertainty of measurement sampling, the reference waveform may vary within a certain frequency range. Therefore, the deformation: is used to replace the resonance controller; where K r can adjust the resonance peak gain and at the same time adjust the resonance frequency band width; ω C can adjust the resonance frequency band width and at the same time adjust the resonance peak gain; ω0 is the resonance frequency point. By adding 2ω c after that, the gain at the resonance frequency is reduced, but a frequency band with a relatively large gain is formed near the resonance frequency, thereby reducing the influence of frequency offset;

[0043] In order to compensate for system delay and improve system stability, the present invention performs a delay compensation operation to obtain the transfer function of the resonance controller as where is the phase compensation angle, which represents the angle by which the phase is advanced compared to the uncompensated state when approaching the resonance point ω0 infinitely closely.

[0044] Since the resonance frequency point ω0 changes with the rotational speed, considering the overall system stability, as the resonance frequency point ω0 increases, the gain coefficient K r becomes smaller. Let where c is a control coefficient related to the parameters of the entire motor system, and K r0 is the initial gain coefficient, and the optimal values are obtained through tests in the current loop and the speed loop respectively; in view of the fact that the interference suffered by the bearingless permanent magnet synchronous motor is often a frequency related to the rotational speed, therefore, in the present invention, the mechanical rotational speed ω of the rotor is directly used mThe resonant frequency point ω0 of the adaptive resonant controller realizes adaptive resonant control; therefore, the adaptive resonant controller designed in the present invention is After the adaptive resonant controller and the PI controller are connected in parallel, an adaptive PIR controller is obtained, and its transfer function is

Claims

1. A speed control method applicable to a magnetic levitation pump at low speeds, characterized in that: The method comprises the following steps: 1) After the magnetic levitation pump is started, the angle and rotational speed of the rotor in the magnetic levitation pump and the current values flowing through the driving windings a and b of the magnetic levitation pump are measured by sensors. Then, the angle and rotational speed of the rotor are respectively converted into electrical angle and electrical rotational speed. According to the extended Kalman filtering algorithm, the electrical angle, electrical rotational speed and the current values of phases a and b are corrected to obtain the corrected electrical angle, electrical rotational speed and the current values of phases a and b; 2) The corrected current values of phases a and b are converted into the actual current values of the driving d and q axes of the rotor by Park transformation using the corrected electrical angle, and the corrected electrical rotational speed is converted into mechanical rotational speed as the actual rotational speed of the rotor; 3) Using the target rotational speed, actual rotational speed, actual current values of the driving d and q axes of the rotor and the target current value, the target voltage of the driving d axis and the target voltage of the driving q axis of the rotor are respectively obtained through a plurality of adaptive PIR controllers; The specific content of step 3) is: Set the target rotational speed of the rotor. After subtracting the actual rotational speed of the rotor from the target rotational speed of the rotor, the difference value and the actual rotational speed of the rotor are input into the first PIR controller to obtain the target current value of the driving q axis of the rotor; Set the target current value of the driving d axis of the rotor to 0. After subtracting the target current value from the actual current value of the driving d axis of the rotor, the difference value and the actual rotational speed of the rotor are input into the second PIR controller to obtain the target voltage of the driving d axis of the rotor. At the same time, after subtracting the target current value from the actual current value of the driving q axis of the rotor, the difference value and the actual rotational speed of the rotor are input into the third PIR controller to obtain the target voltage of the driving q axis of the rotor; The first, second and third PIR controllers are all adaptive PIR controllers; The specific content of "inputting the difference value and the actual rotational speed of the rotor into the first PIR controller" is: The difference value between the target rotational speed and the actual rotational speed of the rotor and the actual rotational speed of the rotor pass through the proportional, integral and resonant control links in the first PIR controller at the same time, and the target current value of the driving q axis of the rotor is output; The adaptive PIR controller adaptively controls the resonant frequency point and the resonant gain respectively on the basis of a resonant controller and introduces a compensation phase angle. It is obtained by paralleling an adaptive resonant controller and a PI controller. 4) The target voltage of the driving d axis, the target voltage of the driving q axis of the rotor and the corrected electrical angle are converted into the voltage values of phases a and b by inverse Park transformation, and then the rotor of the magnetic levitation pump is driven and controlled to rotate through a digital-to-analog converter and a DC power amplifier; 5) Repeat steps 1) to 4), and maintain the actual rotational speed of the rotor at the target rotational speed by continuously controlling the actual rotational speed of the rotor; The transfer function G(s) of the adaptive PIR controller is set as follows: Among them, K r0 is the initial gain coefficient, c is the control coefficient, ω C is the bandwidth of the adaptive PIR controller, s is the Laplace operator; K p is the proportional coefficient, K i is the integral coefficient, ω m is the actual rotational speed of the rotor, is the compensation phase angle; The compensation phase angle is determined according to the actual rotational speed of the rotor. When the actual rotational speed of the rotor is less than or equal to 300 rpm, the compensation phase angle is equal to 0 degrees; when the actual rotational speed of the rotor is greater than 300 rpm and less than or equal to 500 rpm, the compensation phase angle is equal to 20 degrees; when the actual rotational speed of the rotor is greater than 500 rpm, the compensation phase angle is equal to 30 degrees.

Images