Predictive phase-locked loop with low computational burden, permanent magnet synchronous motor control device and method

Abstract

The invention discloses a predictive phase-locked loop with low calculation burden and a permanent magnet synchronous motor control device and method, and the method comprises the steps: obtaining a target integral signal containing position information through employing an integral processing equivalent counter electromotive force containing an integral correction signal, and further calculating an equivalent position error and a cost function according to the target integral signal. And setting an initial position compensation angle according to the calculated equivalent position error and a cost function, and superposing the initial position compensation angle to the estimated position of the last period to obtain an initial position sector boundary. And designing a Newton iteration function based on the target integral signal and an initial position sector boundary to realize rotor position search, and obtaining an estimated rotor position through two Newton iteration operations. The problem that the dynamic performance of a traditional phase-locked loop scheme is limited and the problems that an existing prediction phase-locked loop scheme is too high in calculation burden and poor in steady-state observation precision are solved, and therefore it is guaranteed that a permanent magnet synchronous motor sensorless driving system can have good steady-state and dynamic performance.

Description

Technical Field

[0001] The present invention belongs to the field of permanent magnet synchronous motor control, and in particular relates to a predictive phase-locked loop with low computational burden, a permanent magnet synchronous motor control device and a method. Background Art

[0002] Permanent magnet synchronous motor position sensorless drive has the advantages of high efficiency, high power density, low cost and strong robustness, and has been widely used in fields such as robots, industrial drives and household appliances. Permanent magnet synchronous motor position sensorless control methods are mainly divided into two categories: high-frequency signal injection method and fundamental frequency model method. The fundamental frequency model method is to design the back-EMF / flux observer based on the motor fundamental frequency model, and extract the rotor position information from the observed back-EMF / flux signal. It does not rely on the motor salient pole characteristics and the injection of auxiliary signals, and is widely used in the medium and high-speed operation of permanent magnet synchronous motor position sensorless drives.

[0003] At present, many scholars have realized the extraction of motor fundamental frequency signals based on extended Kalman filters, sliding mode observers, model reference adaptive observers, Romberg observers and extended state observers, and further adopted phase-locked loops to extract the phase of fundamental frequency signals as rotor position signals. However, the traditional phase-locked loop uses PI controllers as loop filters. Its fixed parameters and inherent linear characteristics cannot meet the frequently changing working conditions of the motor. For example, when the input is a frequency ramp signal, the accurate tracking of the motor rotor position cannot be achieved, which limits the use of traditional phase-locked loops in applications where the motor frequency changes rapidly.

[0004] In the prior art, the predictive phase-locked loop is usually based on an equal sector mechanism to set the position sector boundary, and then calculate and compare the cost function of each sector boundary to obtain the optimal rotor position, but the equal sector mechanism is undoubtedly the main source of the computational burden. The predictive phase-locked loop is designed based on the Newton iteration principle. Although the initial sector boundary can converge to the optimal rotor position through 3 Newton iterations, and the phase-locking accuracy reaches 1.26×10-8rad, this scheme still requires 4 iterative operations to obtain the initial position sector boundary that satisfies local convergence, so the total number of iterations of this scheme reaches 7 times. In addition, there is also a predictive phase-locked loop based on adaptive step size, which adjusts the iteration step size and direction in real time through the designed cost function to accelerate the iteration process, which can further reduce the number of iterations to 6 times and improve the theoretical accuracy to 5.6×10-8rad.

[0005] Although the above-mentioned predictive phase-locked loop can achieve the number of iterations of the predictive phase-locked loop while ensuring the phase-locked accuracy, its computational complexity still needs further optimization; in addition, the speed and rotor position ripples of the predictive phase-locked loop under steady-state conditions are slightly larger than those of the traditional phase-locked loop. How to improve the filtering capability of the predictive phase-locked loop needs further research. Summary of the invention

[0006] The purpose of the present invention is to provide a predictive phase-locked loop with low computational burden, a permanent magnet synchronous motor control device and method, which are used to solve the problem of limited dynamic performance of traditional phase-locked loop schemes and the problems of excessive computational burden and poor steady-state observation accuracy in existing predictive phase-locked loop schemes, thereby ensuring that the permanent magnet synchronous motor position sensorless drive system can have good steady-state and dynamic performance.

[0007] The technical solution provided by the present invention is as follows:

[0008] In a first aspect, the present invention provides a predictive phase-locked loop with low computational burden, the predictive phase-locked loop comprising: a cost function module, an initial position sector setting module and a Newton iterative rotor position search module; wherein,

[0009] The cost function module estimates the equivalent back electromotive force of the permanent magnet synchronous motor according to the αβ axis voltage and current signals, obtains the target integral signal containing the position information by using the estimated equivalent back electromotive force through integral processing, and calculates the equivalent position error and the cost function;

[0010] The initial position sector setting module sets the initial position compensation angle according to the calculated equivalent position error and the cost function, and superimposes it on the estimated position of the previous cycle to obtain the initial position sector boundary;

[0011] The Newton iterative rotor position search module designs a Newton iterative function based on the target integral signal and the initial position sector boundary to realize rotor position search and obtain an estimated rotor position signal.

[0012] In one implementation, considering the harmonic components caused by sampling error, inverter nonlinearity and motor parameter mismatch non-ideal factors, the estimated equivalent back electromotive force of the permanent magnet synchronous motor is expressed as:

[0013]

[0014] In the formula, u α 、u β with i α 、i β are the stator voltage and stator current of the αβ axis system, R s is the stator resistance, p is the differential operator, L q is the q-axis inductance, θ e is the motor rotor position, To estimate the equivalent back EMF, is the estimated equivalent back electromotive force of the αβ axis system, To estimate the equivalent back EMF vector, e e1 To estimate the equivalent back EMF fundamental frequency vector, E α0 、Eβ0 are the estimated equivalent back electromotive force of the α-axis and the β-axis, e e0 =[E α0 E β0 ] T To estimate the equivalent back EMF DC vector, E h are respectively used to estimate the amplitude of the hth equivalent back-EMF harmonic, To estimate the hth equivalent back-EMF harmonic vector.

[0015] In one embodiment, in order to effectively suppress the negative impact of DC offset and hth harmonic components on position observation, a synchronous rotating low-pass filter is introduced to extract the baseband signal of the integral signal, and the integral signal output by the synchronous rotating low-pass filter is used as the integral correction signal. and It can be obtained through simple coordinate transformation and low-pass filtering:

[0016]

[0017] In the formula, and is the integrated signal of the αβ axis system, and is the integrated signal of the γδ axis system, and is the integral signal of the γδ axis system output by the low-pass filter, and is the integral signal output by the synchronous rotating low-pass filter, ω c is the cutoff frequency of the low-pass filter, is the estimated position of the rotor, and s represents a complex variable in the complex plane.

[0018] In one embodiment, the integrated signal output by the synchronously rotating low-pass filter is and As the integral correction signal, the integral signal of the αβ axis system and Calculated by the following formula:

[0019]

[0020] in, and are the proportional and integral gains of the PI controller and satisfy:

[0021]

[0022] In the formula, and is the integral correction signal, ω eis the electrical angular velocity of the motor, D α , D β are the integral compensation amounts of the α-axis and β-axis respectively, and n is an adjustable parameter greater than 0.

[0023] In one implementation, the equivalent position error is:

[0024]

[0025] In the formula, is the position estimation error, ψ af is the amplitude of the effective flux, which can be obtained by calculate, is the estimated rotor position.

[0026] In one embodiment, the cost function is:

[0027] J F =|E err |.

[0028] In one embodiment, the transfer function G is defined according to the harmonic suppression capability of the cost function and the parameter tuning criteria. αβ (s) is:

[0029]

[0030] In the formula, H αβ (s) and The transfer function between .

[0031] In one implementation, the initial position sector boundary is obtained by the following formula:

[0032]

[0033] Where k represents the kth current sampling period, is the initial position sector boundary, θ c is the position compensation angle.

[0034] In one implementation, the Newton iterative function is:

[0035]

[0036] In the formula, j represents the number of iterations, and the subscript (j-1) represents the j-1th iteration parameter. Representatives and The associated equivalent position error, Representatives and The relevant Newton iteration formula;

[0037] The sufficient condition for determining the convergence interval of Newton iteration operation is: (1) about The first-order derivative of is not 0; (2) about The second-order derivative of is not 0; (3) satisfies The expression related to the sufficient condition is expressed as:

[0038]

[0039] In a second aspect, the present invention provides a permanent magnet synchronous motor control device based on a predictive phase-locked loop with low computational burden, the device comprising: the above-mentioned predictive phase-locked loop with low computational burden.

[0040] In a third aspect, the present invention provides a permanent magnet synchronous motor control method based on the above-mentioned predictive phase-locked loop with low computational burden, the method comprising:

[0041] Obtaining an estimated rotor position signal via a predictive phase-locked loop with low computational burden;

[0042] The estimated rotor position signal is processed by differentiation and low-pass filter to obtain an estimated speed signal;

[0043] The estimated rotor position signal and speed signal are used for closed-loop control to achieve position sensorless control of the permanent magnet synchronous motor.

[0044] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0045] (1) The present invention proposes a new cost function based on an integral signal, which has good DC and harmonic suppression capabilities, thereby solving the problem that the traditional iterative phase-locked loop is susceptible to harmonic interference;

[0046] (2) The present invention proposes a simple initial position sector boundary setting scheme, which can search for an initial value of the iteration that satisfies the local convergence condition of the Newton iteration and is close to the actual position only through a new cost function, thereby reducing the computational burden brought by the initial value of the iteration setting;

[0047] (3) The present invention proposes a Newton iterative rotor position search method based on the integral signal, which can achieve 2.19×10 -6 The phase-locking accuracy of rad can effectively reduce the computational burden brought by the rotor position search and significantly improve the dynamic response capability of the system.

[0048] (4) The predictive phase-locked loop provided by the present invention can also be used for grid-connected synchronization technology, asynchronous motor position sensorless control technology and synchronous reluctance motor position sensorless control technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are part of the present invention and are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but do not constitute an improper limitation of the present invention. Obviously, the accompanying drawings described below are only some embodiments. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0050] Figure 1 A block diagram of a predictive phase-locked loop structure with low computational burden provided by an embodiment of the present invention;

[0051] Figure 2 G under different parameter n conditions provided in an embodiment of the present invention αβ Bode plot of (s);

[0052] Figure 3 is the calculation domain θ provided in one embodiment of the present invention err E within ∈(-π,π)rad err , J F With Δθ c ;

[0053] Figure 4 The value of the sufficient condition expression for determining the convergence interval of the Newton iteration operation provided in one embodiment of the present invention in the entire calculation domain;

[0054] Figure 5 A flow chart of a predictive phase-locked loop with low computational burden provided in one embodiment of the present invention;

[0055] Figure 6 The steady-state experimental comparison results of the conventional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention under the conditions of 50% rated load and 500 rpm speed provided in one embodiment of the present invention are shown; wherein, Figure 6 (a) is the steady-state experimental result of the traditional prediction phase-locked loop. Figure 6 (b) The steady-state experimental results of the predicted phase-locked loop provided by the present invention;

[0056] Figure 7 The experimental comparison results of the conventional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention under the α-axis 5V DC voltage disturbance provided in one embodiment of the present invention are shown; wherein, Figure 7 (a) is the steady-state experimental result of the traditional prediction phase-locked loop. Figure 7 (b) The steady-state experimental results of the predicted phase-locked loop provided by the present invention;

[0057] Figure 8Experimental comparison results of a conventional phase-locked loop, a conventional predictive phase-locked loop and a predictive phase-locked loop provided by the present invention under rapid acceleration and deceleration conditions; wherein, Figure 8 (a) is the experimental result of the traditional phase-locked loop. Figure 8 (b) is the experimental result of the traditional prediction phase-locked loop. Figure 8 (c) Experimental results of the predicted phase-locked loop provided by the present invention;

[0058] Fig. 9 Experimental comparison results of a conventional phase-locked loop, a conventional predictive phase-locked loop and a predictive phase-locked loop provided in an embodiment of the present invention under sudden load addition and sudden load removal conditions; wherein, Fig. 9 (a) is the experimental result of the traditional phase-locked loop. Fig. 9 (b) is the experimental result of the traditional prediction phase-locked loop. Fig. 9 (c) Experimental results of the predicted phase-locked loop provided by the present invention;

[0059] It should be noted that these drawings and textual descriptions are not intended to limit the conceptual scope of the present invention in any way, but are intended to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0060] In order to deepen the recognition and understanding of the present invention, the technical solution of the present invention is further introduced below in conjunction with the accompanying drawings and specific implementation methods. The following embodiments are explanations of the present invention and the present invention is not limited to the following embodiments.

[0061] The embodiment of the present invention provides a predictive phase-locked loop with low computational burden, and its structural block diagram is as follows: Figure 1 The low computational burden predictive phase-locked loop includes: a cost function module, an initial position sector setting module and a Newton iterative rotor position search module, wherein:

[0062] The cost function module is used to calculate the equivalent back electromotive force of the permanent magnet synchronous motor containing fundamental frequency components, DC components and harmonic components, and use integral processing of the equivalent back electromotive force to obtain the target integral signal containing position information, and calculate the equivalent position error and cost function;

[0063] An initial position sector setting module is used to set the initial position compensation angle according to the calculated equivalent position error and cost function, and superimpose it on the estimated position of the previous cycle to obtain the initial position sector boundary;

[0064] The Newton iterative rotor position search module is used to design a Newton iterative function based on the target integral signal and the initial position sector boundary to realize the rotor position search and obtain the estimated rotor position information.

[0065] The low computational burden predictive phase-locked loop of this embodiment solves the problem of limited dynamic performance of the traditional phase-locked loop solution and the problem of excessive computational burden and poor steady-state observation accuracy of the existing predictive phase-locked loop solution by improving the cost function module, the initial position sector setting module and the Newton iteration rotor position search module, thereby ensuring that the permanent magnet synchronous motor position sensorless drive system can have good steady-state and dynamic performance. The improvement implementation process of each part is described in detail below.

[0066] The first part of the cost function module (cost function with harmonic suppression capability):

[0067] This part takes into account the harmonic components caused by non-ideal factors such as sampling error, inverter nonlinearity and motor parameter mismatch. The cost function in the embodiment of the present application has the ability to suppress harmonics.

[0068] Furthermore, the equivalent back electromotive force of the permanent magnet synchronous motor estimated by considering the harmonic components caused by sampling error, inverter nonlinearity and motor parameter mismatch non-ideal factors is expressed as:

[0069]

[0070] In the formula, u α 、u β with i α 、i β are the stator voltage and stator current of the αβ axis system, R s is the stator resistance, p is the differential operator, L q is the q-axis inductance, θ e is the motor rotor position, Estimate the equivalent back EMF for the α-axis and β-axis respectively, To estimate the equivalent back EMF vector, e e1 To estimate the equivalent back EMF fundamental frequency vector, E α0 、E β0 are the estimated equivalent back electromotive force of the α-axis and the β-axis, e e0 =[E α0 E β0 ] T To estimate the equivalent back EMF DC vector, E h are respectively used to estimate the amplitude of the hth equivalent back-EMF harmonic, To estimate the hth equivalent back-EMF harmonic vector.

[0071] Integrate the estimated equivalent back electromotive force:

[0072]

[0073] In the formula, ψ af is the effective flux linkage of the motor, ω eis the electrical angular velocity of the motor.

[0074] From formula (2), we can see that under the action of integration, the DC component in the equivalent back electromotive force will lead to the generation of an integral signal that increases linearly with time, thus leading to the integral saturation of the integral signal. On the other hand, when the system is in a steady state, the amplitude ratio of the hth harmonic component of the integral signal to the fundamental component is λ Ih It can be expressed as:

[0075]

[0076] Obviously, the amplitude ratio of the hth harmonic component to the fundamental component of the integrated signal is h times the amplitude ratio of the hth harmonic component to the fundamental component of the back-electromotive force signal, that is, the introduction of the integrated signal can effectively suppress the negative impact of the hth harmonic component on position observation, but the negative impact caused by the DC offset cannot be ignored.

[0077] To solve this problem, in order to effectively suppress the negative impact of the hth harmonic component on position observation and DC offset, a synchronous rotating low-pass filter is used to extract the baseband signal of the integral signal, and the integral signal output by the synchronous rotating low-pass filter is used as the integral correction signal.

[0078] The base frequency signal of the integrated signal extracted by introducing a synchronous rotating low-pass filter is:

[0079]

[0080] In the formula, the superscript “^” indicates the estimated value of the variable. and is the integrated signal of the αβ axis system, and is the integrated signal of the γδ axis system, and is the integral signal of the γδ axis system output by the low-pass filter, and is the integral signal output by the synchronous rotating low-pass filter, ω c is the cutoff frequency of the low-pass filter, is the estimated position of the rotor, and s represents a complex variable in the complex plane.

[0081] After Laplace transform, the transfer function of the synchronous rotating low-pass filter can be expressed as:

[0082]

[0083] In the formula, ω c It is the cutoff frequency of the low-pass filter, which is usually set to a small value to completely filter out high-frequency disturbance components.

[0084] Furthermore, the integral signal output by the synchronously rotating low-pass filter is and As the integral correction signal, the integral signal of the αβ axis system can be calculated by the following formula:

[0085]

[0086] in, and are the proportional and integral gains of the PI controller and satisfy:

[0087]

[0088] In the formula, and is the integral correction signal, ω e is the electrical angular velocity of the motor, D α , D β are the integral compensation amounts of the α-axis and β-axis respectively, and n is an adjustable parameter greater than 0.

[0089] Furthermore, according to the integrated signal of the αβ axis system and Design of a new equivalent position error E err :

[0090]

[0091] In the formula, is the position estimation error, ψ af is the magnitude of the effective flux, which can be obtained by calculate, is the estimated rotor position.

[0092] Furthermore, according to the new equivalent position error E err Design a cost function J with harmonic suppression capability F :

[0093] J F =|E err | (11)

[0094] Substituting formula (8) into formula (7), and The transfer function H αβ It can be expressed as:

[0095]

[0096] Taking into account the DC disturbance component, using the Laplace transform final value theorem, the integral signal component introduced by e0 can be expressed as:

[0097]

[0098] It can be seen from formula (13) that the negative impact of e0 can be eliminated by introducing an integral correction signal.

[0099] For the baseband signal, and s=jω e Substituting into formula (12) we can get:

[0100]

[0101] It can be seen from formula (14) that the introduction of the integral correction signal will not affect the effect of the integral on the baseband signal.

[0102] In order to further clarify the harmonic suppression capability and parameter tuning criteria of the new cost function, the transfer function G is defined αβ (s) is:

[0103]

[0104] G under different parameter n conditions αβ The Bode plot of (s) is as follows Figure 2 As shown, like Figure 2 As shown, the novel integration method provided by the present invention has the ability to suppress DC and high-order harmonic components, and the change of parameter n will not affect the ability to suppress high-order harmonics. Increasing parameter n will increase the bandwidth of the system, while a larger parameter n will reduce the ability to suppress DC bias components and may amplify low-order harmonic disturbances. Considering the above factors, n is set to 2 in the present invention.

[0105] The second part of the initial position sector setting module:

[0106] Most of the reported prediction phase-locked loops use an equal-sector mechanism to construct the position sector boundaries, and the number of iterations is consistent with the number of sector boundaries. The equal-sector mechanism is undoubtedly the main source of the computational burden. The present invention uses the Newton iteration operation mechanism to search for the rotor position to reduce the number of iterations of the prediction phase-locked loop, but the Newton iteration operation requires the design of a reasonable initial value of the iteration to ensure local convergence; in addition, the closer the initial position sector boundary is to the actual motor rotor position, the higher the theoretical accuracy of the prediction phase-locked loop. Figure 3 The setting principle of the initial position sector boundary provided by the present invention is provided. Figure 3 The cost function J shown F The symmetry of the position angle compensation step Δθ can be designed c :

[0107]

[0108] Furthermore, consider Figure 3 The new equivalent position error E shownerr The monotonicity of the final position compensation angle θ c Can be set to:

[0109] θ c =sign(E err )·Δθ c (17)

[0110] Then the final initial position sector boundary can be obtained by the following formula:

[0111]

[0112] Where, k represents the kth current sampling period, is the initial position sector boundary. Considering the position compensation angle θ c After that, the selection range of the initial position sector boundary can be converted from the entire computational domain to θ err ∈(θ e -π / 6,θ e +π / 6), which helps the convergence and convergence speed.

[0113] The third part is Newton iteration rotor position search module:

[0114] Furthermore, after obtaining the initial position sector boundary, the rotor position search can be equivalent to solving:

[0115]

[0116] about The derivative of can be expressed as:

[0117]

[0118] definition is the initial value of the jth iteration, then The first-order Taylor series expansion at the initial value of the iteration can be expressed as:

[0119]

[0120] Where j is the number of iterations. Substitute into the above formula and ignore the higher-order infinitesimal terms We can get:

[0121]

[0122] According to formulas (19), (20) and (22), the Newton iteration formula can be expressed as:

[0123]

[0124] In the formula, j represents the number of iterations, and the subscript (j-1) represents the j-1th iteration parameter. Representatives and The associated equivalent position error, Representatives and The relevant Newton iteration formula;

[0125] The sufficient condition for determining the convergence interval of Newton iteration operation is: (1) about The first-order derivative of is not 0; (2) about The second-order derivative of is not 0; (3) satisfies

[0126] The expression related to the above sufficient condition can be expressed as:

[0127]

[0128] The values ​​of formulas (24)-(26) in the entire calculation domain are as follows: Figure 4 As shown, when θ err When it belongs to the interval (-π / 2,π / 2), the sufficient conditions for the convergence of the above Newton iteration method can be satisfied, so θ err The computational domain of ∈(-π / 6,π / 6) can achieve local convergence of Newton iteration.

[0129] Furthermore, for the predictive phase-locked loop, the number of iterations is proportional to the computational burden of the digital controller. To determine a reasonable number of iterations and the corresponding iteration accuracy, the phase-locked accuracy after the jth iteration can be calculated by the following formula:

[0130]

[0131] Among them, θ err(j-1) With θ err(j) is the calculation error between the j-1th and jth iterations, and is the rotor position information calculated at the j-1th and jth iterations.

[0132] According to the calculation accuracy formula, the estimation error of the Newton iteration method after different iteration times under different initial value conditions is shown in Table 1. It can be seen from the table that even if there is an error of ±π / 6rad in the initial value of the iteration, the estimation error can be limited to 2.19×10 -6 rad, which can meet the accuracy requirements of most fixed-point or floating-point operations.

[0133] Table 1 Estimation error of Newton iteration method after different iteration times under different initial value conditions

[0134]

[0135] Furthermore, the speed signal is obtained by differentiating the position signal, and a low-pass filter (LPF) is used to smooth the estimated speed signal:

[0136]

[0137] Among them, T c is the sampling period.

[0138] In one embodiment, a permanent magnet synchronous motor control device based on a predictive phase-locked loop with low computational burden is proposed. The device includes: the predictive phase-locked loop with low computational burden described above, which is used to obtain an estimated rotor position.

[0139] For the description of the predictive phase-locked loop based on low computational burden, please refer to the description of the same or similar parts mentioned above, which will not be repeated here.

[0140] Continue to refer to Figure 1 As shown, in one embodiment, a permanent magnet synchronous motor control method based on a predictive phase-locked loop with low computational burden is proposed, the method comprising the following steps:

[0141] Obtaining an estimated rotor position signal via a predictive phase-locked loop with low computational burden;

[0142] The estimated rotor position signal is processed by differentiation and low-pass filter to obtain an estimated speed signal;

[0143] The estimated rotor position signal and speed signal are used for closed-loop control to achieve position sensorless control of the permanent magnet synchronous motor.

[0144] Predict the process of the phase-locked loop to obtain the estimated rotor position information, refer to Figure 5 The cost function module has the ability to suppress harmonics. It first estimates the equivalent back electromotive force of the permanent magnet synchronous motor. Then, the equivalent back electromotive force is processed by integrating the integral correction signal to obtain the target integral signal containing the position information, and the equivalent position error E is further calculated based on the target integral signal. err With the cost function J F Initial position sector setting, design position angle compensation step length Δθ based on the calculated equivalent position error and cost function c , and superimpose it to the estimated position of the previous cycle to obtain the initial position sector boundary. Newton iterative rotor position search, based on the target integral signal and the initial position sector boundary, the Newton iterative function is designed to realize the rotor position search, and the estimated rotor position is obtained after two Newton iterative operations.

[0145] For a detailed description of the predictive phase-locked loop with low computational burden, please refer to the description of the same or similar parts mentioned above, which will not be repeated here.

[0146] In a specific embodiment, the beneficial effects of the present invention are verified by a 750W permanent magnet synchronous motor experimental platform. The parameters of the 750W permanent magnet synchronous motor are: rated frequency 250Hz, rated speed 3000rpm, pole pair number 5, d-axis inductance 4.7mH, q-axis inductance 6.7mH, stator resistance 0.9Ω, permanent magnet flux linkage 0.054Wb, and rated current 4.2A. A hysteresis brake is used as the motor load, a two-level three-phase inverter is used to drive the permanent magnet synchronous motor, and a DSP TMS320F28335 is used to implement the control algorithm, and the sampling frequency and switching frequency are both 10kHz.

[0147] Figure 6 The steady-state experimental comparison results of the conventional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention under the conditions of 50% rated load and 500 rpm speed are shown in FIG. Figure 6 As shown in (a), for the traditional predictive phase-locked loop, the estimated back-EMF signal has significant distortion. The fast Fourier transform (FFT) analysis results show that the amplitudes of the 5th, 7th, 11th and 13th harmonic components are 7.24%, 4.45%, 2.67% and 0.97% of the amplitude of the fundamental component, which leads to the existence of significant 6th rotor position ripples in the rotor position signal estimated by the traditional predictive phase-locked loop, and the speed estimation error can reach ±23.5rpm; Figure 6 As shown in (b), by using the predictive phase-locked loop provided by the present invention, the obtained target integral signal has good sinusoidality, the amplitudes of the 5th, 7th, 11th and 13th harmonic components can be reduced to 1.90%, 0.97%, 0.36% and 0.59% of the amplitude of the fundamental component, the 6th rotor position ripple can be eliminated, and the speed estimation error can be limited to within ±6.8rpm. The predictive phase-locked loop provided by the present invention has good harmonic suppression capability and observation accuracy.

[0148] Figure 7 The experimental comparison results of the conventional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention under the α-axis 5V DC voltage disturbance are shown in FIG. Figure 7 As shown in (a), when E is applied α0 = 0 DC voltage disturbance to stator voltage u α After that, a 5V bias error appeared in the estimated α-axis back EMF, which in turn caused a 10.0° fundamental frequency position ripple in the traditional prediction phase-locked loop observation results, indicating that the traditional prediction phase-locked loop does not have the ability to suppress DC bias errors; Figure 7 As shown in (b), when E is applied α0 = 0 DC voltage disturbance to stator voltage u αAfter that, the target integral signal can effectively compensate for the DC bias of the α-axis integral signal, ensuring that the amplitude of the α-axis integral signal is consistent with that of the β-axis integral signal, and the position estimation error can converge to about 0° within one electrical cycle. The prediction phase-locked loop provided by the present invention has good DC bias resistance.

[0149] Figure 8 The experimental comparison results of the conventional phase-locked loop, the conventional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention under the conditions of rapid acceleration and deceleration are shown in FIG. 1 , wherein the speed variation trend is 1500-3000-1500rpm, and the accelerations in acceleration and deceleration are 1800rpm / s and 900rpm / s respectively. Figure 8 (a), limited by the limited bandwidth of the traditional phase-locked loop, the maximum position estimation error can reach 22.9° and 17.7° during rapid acceleration and deceleration; refer to Figure 8 (b) and (c), for the traditional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention, the position estimation error fluctuates around 0 degrees during the dynamic process. The position estimation error of the traditional predictive phase-locked loop is maintained within ±11.1°, while the predictive phase-locked loop provided by the present invention can further reduce the position estimation error to within ±9.3°.

[0150] Fig. 9 This is the experimental comparison result of the conventional phase-locked loop, the conventional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention under the conditions of sudden load addition and sudden load removal, wherein the speed reference is 2400rpm and the load step signal is 0~80~0%. Fig. 9 (a) Under step load disturbance, the traditional phase-locked loop has significant position estimation error, and the maximum position estimation error can reach 15.9° during the dynamic process; refer to Fig. 9 (b) and (c), the traditional predictive phase-locked loop and the predictive phase-locked loop provided by the present invention do not produce obvious position estimation errors during load changes, indicating that the predictive phase-locked loop provided by the present invention has good dynamic and static performance.

[0151] The above experimental effect diagrams show that compared with the traditional phase-locked loop, the predictive phase-locked loop provided by the present invention has a faster dynamic response capability; and compared with the traditional predictive phase-locked loop, the predictive phase-locked loop provided by the present invention has a stronger DC and harmonic disturbance suppression capability, a lower computational burden and a higher estimation accuracy.

[0152] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0153] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A predictive phase-locked loop with low computational burden, characterized in that: The predictive phase-locked loop includes: a cost function module, an initial position sector setting module and a Newton iterative rotor position search module; wherein, The cost function module estimates the equivalent back electromotive force of the permanent magnet synchronous motor according to the αβ axis voltage and current signals, obtains the target integral signal containing the position information by using the estimated equivalent back electromotive force through integral processing, and calculates the equivalent position error and the cost function; The initial position sector setting module sets the initial position compensation angle according to the calculated equivalent position error and the cost function, and superimposes it on the estimated position of the previous cycle to obtain the initial position sector boundary; The Newton iterative rotor position search module designs a Newton iterative function based on the target integral signal and the initial position sector boundary to realize rotor position search and obtain an estimated rotor position signal.

2. The predictive phase-locked loop with low computational burden according to claim 1, characterized in that: Considering the harmonic components caused by sampling error, inverter nonlinearity and motor parameter mismatch non-ideal factors, the estimated equivalent back electromotive force of the permanent magnet synchronous motor is expressed as: In the formula, u α 、u β with i α 、i β are the stator voltage and stator current of the αβ axis system, R s is the stator resistance, p is the differential operator, L q is the q-axis inductance, θ e is the motor rotor position, Estimate the equivalent back EMF for the α-axis and β-axis respectively, To estimate the equivalent back EMF vector, e e1 To estimate the equivalent back EMF fundamental frequency vector, E α0 、E β0 are the estimated equivalent back electromotive force of the α-axis and the β-axis, e e0 =[E α0 E β0 ] T To estimate the equivalent back EMF DC vector, E h are respectively used to estimate the amplitude of the hth equivalent back-EMF harmonic, To estimate the hth equivalent back-EMF harmonic vector.

3. The predictive phase-locked loop with low computational burden according to claim 1 or 2, characterized in that: In order to effectively suppress the negative impact of DC offset and hth harmonic components on position observation, a synchronous rotating low-pass filter is introduced to extract the base frequency signal of the integral signal, and the integral signal output by the synchronous rotating low-pass filter is used as the integral correction signal; the synchronous rotating low-pass filter extracts the integral signal and It can be obtained through simple coordinate transformation and low-pass filtering: In the formula, and is the integrated signal of the αβ axis system, and is the integrated signal of the γδ axis system, and is the integral signal of the γδ axis system output by the low-pass filter, and is the integral signal output by the synchronous rotating low-pass filter, ω c is the cutoff frequency of the low-pass filter, is the estimated position of the rotor, and s represents a complex variable in the complex plane.

4. The predictive phase-locked loop with low computational burden according to claim 3, characterized in that: The integrated signal output by the synchronously rotating low-pass filter and As the integral correction signal, the integral signal of the αβ axis system and Calculated by the following formula: in, With k i D are the proportional and integral gains of the PI controller and satisfy: In the formula, and is the integral correction signal, ω e is the electrical angular velocity of the motor, D α , D β are the integral compensation amounts of the α-axis and β-axis respectively, and n is an adjustable parameter greater than 0.

5. The low computational burden predictive phase-locked loop according to claim 4, characterized in that: The equivalent position error E err for: In the formula, is the position estimation error, ψ af is the magnitude of the effective flux, which can be obtained by calculate, is the estimated rotor position.

6. The predictive phase-locked loop with low computational burden according to claim 5, characterized in that: The cost function is: J F =|E err |。 7. The predictive phase-locked loop with low computational burden according to claim 5, characterized in that: The initial position sector boundary is obtained by the following formula: Where k represents the kth current sampling period, is the initial position sector boundary, θ c is the position compensation angle.

8. The predictive phase-locked loop with low computational burden according to claim 7, characterized in that: The Newton iterative function is: In the formula, j represents the number of iterations, and the subscript (j-1) represents the j-1th iteration parameter. Representatives and The associated equivalent position error, Representatives and The relevant Newton iteration formula; The sufficient condition for determining the convergence interval of Newton iteration operation is: (1) about The first-order derivative of is not 0; (2) about The second-order derivative of is not 0; (3) The expression related to the sufficient condition is expressed as:

9. A permanent magnet synchronous motor control device based on a predictive phase-locked loop with low computational burden, characterized in that: The device comprises: a predictive phase-locked loop with low computational burden as described in any one of claims 1-8.

10. A permanent magnet synchronous motor control method based on the predictive phase-locked loop with low computational burden according to any one of claims 1 to 8, characterized in that: The method comprises: Obtaining an estimated rotor position signal via a predictive phase-locked loop with low computational burden; The estimated rotor position signal is processed by differentiation and low-pass filter to obtain an estimated speed signal; The estimated rotor position signal and speed signal are used for closed-loop control to achieve position sensorless control of the permanent magnet synchronous motor.

Images