Permanent magnet synchronous motor sensorless noise reduction control method for random frequency phase high-frequency square wave voltage injection

Abstract

The invention discloses a permanent magnet synchronous motor sensorless noise reduction control method for random frequency phase high-frequency square wave voltage injection. Four kinds of high-frequency square wave voltage signals with different frequencies and different phases are randomly combined and injected into an estimated rotor reference system, the injected four kinds of voltage signals follow the principle that the volt-second areas are equal, and it is guaranteed that the amplitude values of induced currents are equal. And meanwhile, the composition probability of each square wave is distributed by utilizing the combination module, so that the probability of each frequency square wave is equal, the distribution of the current power spectral density (PSD) is more uniform, and the amplitude of the PSD is further reduced. A high-frequency signal is utilized to excite the saturation salient polarity of the motor, an estimated position angle and an estimated rotating speed are extracted from high-frequency current response, demodulation is performed through two times of rotation transformation and absolute value processing, then normalization processing is performed, the influence of the amplitude of an injected signal and related parameters on the estimated position angle is avoided, and finally observation is completed through an orthogonal phase-locked loop. And position-sensorless control of the permanent magnet synchronous motor at a low speed is realized by estimating the position angle and the rotating speed.

Description

Technical Field

[0001] The present invention belongs to the field of permanent magnet synchronous motor drive control applications, and in particular relates to a position sensorless noise reduction control method for a permanent magnet synchronous motor with high-frequency square wave injection. Background Art

[0002] In motor drive systems, high-performance motor control strategies rely on position sensors to obtain real-time rotor position signals. Sensorless control technology eliminates mechanical position sensors, reducing system size and cost while enhancing system robustness, and has garnered widespread attention. Sensorless control technologies can be broadly categorized into two types: model-based and salient-pole methods. The model-based method uses a mathematical model of the motor to calculate back-EMF to estimate rotor position when the motor is operating at medium or high speeds. Since back-EMF is approximately proportional to speed, the signal-to-noise ratio decreases at low speeds, making the back-EMF method inapplicable. In contrast, the salient-pole method uses the induced current to track rotor salient poles to determine rotor position. Based on the type of injected signal, high-frequency signal injection methods can be categorized as sinusoidal or square wave. Compared to sine waves, square wave injection allows for higher-frequency signal injection, facilitating high-frequency signal extraction and increasing system bandwidth.

[0003] Although the high-frequency injection method can achieve good operation of the motor in the low-speed range, the injected high-frequency signal will cause a relatively serious noise problem, which seriously limits the promotion of this technology in industrial applications. To solve this problem, two methods are commonly used. One is the random frequency high-frequency square wave voltage injection method. This method randomly injects two square wave voltages of different frequencies into the rotor reference frame, expanding the power spectrum density of the current, reducing discrete spikes, and reducing noise. The second is the random phase high-frequency square wave injection method. This method randomly injects four square wave motors of the same frequency but different phases into the rotor reference frame, eliminating the current discrete spectrum and reducing noise. However, in the two methods of random frequency square wave injection and random phase square wave injection, only the frequency or phase of the injected signal is random. The power spectrum density expansion range can be increased by fully randomizing the frequency and phase, thereby improving the noise reduction effect. Summary of the Invention

[0004] Purpose of the invention: To address the problems that mechanical position sensors increase costs and reduce reliability, and at the same time reduce the noise caused by high-frequency injection and the maximum noise at the injection frequency, a position sensorless noise reduction control method for permanent magnet synchronous motors with random frequency and phase high-frequency square wave voltage injection is proposed to reduce noise and realize position sensorless operation of permanent magnet synchronous motors at low speeds.

[0005] The technical solution is: a position sensorless noise reduction control method for a permanent magnet synchronous motor with random frequency and phase high frequency square wave voltage injection, comprising the following steps:

[0006] Step 1: Generate a random high-frequency square wave voltage signal. Eight high-frequency square wave voltage signals with different frequencies and phases are generated by combination, and the probability of each frequency square wave is adjusted to be equal.

[0007] Step 2: Inject the voltage signal into the estimated reference frame, and inject a high-frequency square wave voltage signal with completely random frequency and phase into the estimated d-axis;

[0008] Step 3: Demodulate the position angle and speed: The high-frequency response current is extracted through a high-pass filter (HPF). To reduce the impact of the injected amplitude on the position estimation, it is normalized and then passed through an orthogonal phase-locked loop to obtain the position error signal. Finally, the estimated electrical angular velocity is obtained through a proportional-integral (PI) controller. The estimated position angle is obtained by integrating the electrical angular velocity.

[0009] Furthermore, the square wave waveform in step 1 is:

[0010] Define the unit square wave function φ sqr :

[0011]

[0012] Where, is the injection signal phase, is the random operator for random signal injection; t r (t,T i ) is time t divided by T i The remainder of V i 、T i is the amplitude and period of the injected random square wave voltage, i={1,2}; u ori 、u fin are the injected signals before and after combination respectively.

[0013] Furthermore, the principle for generating the square wave voltage signal in step 1 is:

[0014] In order to ensure the consistency of the high-frequency induced current amplitude, the principle of equal volt-second area should be followed, that is, the product of the period and amplitude of different frequency signals is a fixed value.

[0015] Furthermore, in step 2, a random square wave voltage signal is injected into the estimated d-axis:

[0016]

[0017] Where, and To estimate the high frequency voltage of dq axis.

[0018] Furthermore, the high-frequency current response of step 3 is: sampling the motor phase current, and obtaining the current i in the two-phase stationary coordinate system through abc / αβ (Clark) transformation α 、i β , and then use the high-pass filter HPF to extract the high-frequency induced current i αh 、iβh:

[0019]

[0020] Where V i 、T i is the amplitude and period of the injected random square wave voltage; i inj is the induced current corresponding to the injected square wave; is the phase of the induced current, L dh and L qh is the dq axis high frequency incremental inductance; Δθ is the position angle estimation error, and θ e is the actual position angle, is the estimated position angle; φ tri is the unit triangle wave function, represent The remainder when divided by 2.

[0021] Furthermore, the processing process of the induced high-frequency current in step 3 is as follows:

[0022] Project the induced current into the measurement reference frame that lags the estimation reference frame by 45°:

[0023]

[0024] Where, ΔL=(L qh -L qh ) / 2, and To measure the high frequency current in the reference frame; is the Park transform, is the estimated position angle; i αh 、i βh is the extracted high-frequency induced current of αβ axis; L dh and L qh is the dq axis high frequency incremental inductance; Δθ is the position angle estimation error;

[0025] When the position angle error Δθ is small enough, the above formula can be simplified to:

[0026]

[0027] Taking the absolute value of the above formula, we get:

[0028]

[0029] Then project it back to the two-phase stationary coordinate system:

[0030]

[0031] Where, and is the demodulated high-frequency current of the αβ axis. When the position angle error Δθ is small enough, it can be simplified to:

[0032]

[0033] Where θ e is the actual position angle.

[0034] Furthermore, in order to reduce the influence of injection amplitude and other factors on position estimation, the demodulated signal is normalized:

[0035]

[0036] In Chinese, and is the normalized current signal; and is the demodulated αβ axis high frequency current.

[0037] Furthermore, the position error signal ε is:

[0038]

[0039] Where, and is the normalized current signal, is the estimated position angle, and Δθ is the position angle estimation error.

[0040] Finally, the estimated electrical angular velocity is obtained through the proportional-integral PI controller, and the estimated position angle can be obtained by integrating the electrical angular velocity.

[0041] Technical effect: Claim 1 seeks to protect a method for noise reduction control of a permanent magnet synchronous motor without position sensor by random frequency and phase high-frequency square wave voltage injection. The present invention realizes the calculation of estimated position angle and speed based on the principle of motor saturation salient polarity, eliminates the need for mechanically installed position sensors, and reduces costs. In order to solve the problem of obtaining the position angle at zero and low speed of the motor, a high-frequency square wave voltage injection method with completely random frequency and phase is adopted, and a random signal generator based on a probability modulation mechanism is designed. It can directly inject four basic square wave voltage signals with different frequencies and phases, or inject a square wave signal with an additional frequency generated by a specific combination of the basic square wave voltage signal, thereby widening the frequency range of the injected signal and making high-frequency signals of different frequencies have the same injection probability. By demodulating the high-frequency response current to obtain the estimated position angle and speed of the motor, the permanent magnet synchronous motor without position sensor control is realized. The power spectrum density of the current in this method has a wider range and is more evenly distributed, reducing the noise caused by the high-frequency injection method. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A structural block diagram of a dual three-phase motor high-frequency harmonic suppression strategy system provided by an embodiment of the present invention;

[0043] Figure 2 A combination module provided by an embodiment of the present invention;

[0044] Figure 3 The injected square wave before the combination provided by the embodiment of the present invention, T=T1 in (a), (b) T = T2, (c) T = T1, (d) T = T2,

[0045] Figure 4 The injected square wave after the combination provided by the embodiment of the present invention, (a) T = 2T1, (b) T = 2T2, (c) T = 2T1, (d) T = 2T2,

[0046] Figure 5 A structural block diagram of an observer provided in an embodiment of the present invention;

[0047] Figure 6 The PSD experimental result diagram provided by the embodiment of the present invention;

[0048] Figure 7 A diagram showing the position angle results of a motor provided by an embodiment of the present invention;

[0049] Figure 8 This is a diagram showing the speed results of the motor provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] like Figure 1 As shown, the present invention proposes a position sensorless noise reduction control method for a permanent magnet synchronous motor with random frequency and phase high frequency square wave voltage injection.

[0052] The specific implementation steps of the proposed method for position sensorless noise reduction control of a permanent magnet synchronous motor with random frequency and phase high-frequency square wave voltage injection include:

[0053] Step 1: Generate a high-frequency square wave voltage signal with random frequency and phase

[0054] like Figure 2 As shown, the present invention proposes a random signal generator designed based on a probability modulation mechanism, which can directly inject four basic square wave voltage signals with different frequencies and phases, or inject square wave signals with additional frequencies generated by specific combinations of basic square wave voltage signals, thereby broadening the frequency range of the injected signal and making high-frequency signals of different frequencies have the same injection probability.

[0055] First, two random numbers, R1 and R2, are generated by the linear congruential method. The function of R1 is to select a basic square wave, and the function of R2 is to determine whether to combine the basic square waves. The initial value of i is 0, and its function is to exit the combination program after the square wave combination is completed. First, set the initial probability parameters p1=3 / 4, p2=1 / 2, p3=1 / 4. This design can ensure that the probabilities of the four injected basic square waves selected for the first time are equal. After completing a cycle of square wave injection, it is necessary to judge by R2 whether to directly output the basic waveform or combine the basic waveforms. If a combination of basic waveforms is required, the probability parameters need to be adjusted accordingly based on the first selected basic waveform to combine the two basic waveforms. For example, when R1>3 / 4, first select Figure 3 (a) shows the basic square wave, at this time i = 1, if R2 < 1 / 3, then the basic square wave is directly output. Otherwise, it is necessary to combine the basic waveforms. The combination of basic square waves needs to meet the requirements of equal frequency and opposite phase. The injection signal of the previous cycle is Figure 3 Taking the basic square wave shown in (a) as an example, the probability parameters will be updated to p1=1, p2=1, p3=0, which means Figure 3 (a) with Figure 3 (c) Combine. After the two basic waveforms are combined, they will be directly output Figure 4(a) waveform, while setting \(i = 2\) and jumping out of the combination program, and starting to select a new injection signal. This random signal generator not only retains the original frequency components (\(T1\) and \(T2\)), but also introduces signal components with frequencies of \(2T1\) and \(2T2\), making the diffusion range of the phase current power spectral density (PSD) expand from \([f1, f2]\) to \([0.5f1, f2]\) (assuming \(f1 < f2\)), which is more conducive to noise suppression.

[0056] It should also be noted that in this article, \(0.5f2 = f1\), then the output signal frequencies include \(0.5f1\), \(f1\), and \(f2\). For the basic square wave signal with a period of \(T1\), Figure 3 (a) or Figure 3 (c) will be directly output with a probability of \(1 / 3\). At this time, the probability \(P(f1)\) of injecting a high-frequency signal with a frequency of \(f1\) is \(P(f1)=(1 / 2)×(1 / 3)=1 / 6\). When the basic square wave signal with a period of \(T1\) combines the current signal with its \(180°\) inverted square wave with a probability of \(2 / 3\), the probability of injecting a high-frequency signal with a frequency of \(0.5f1\) is: \(P(0.5f1)=(1 / 2)×(2 / 3)=1 / 3\). For example Figure 3 (a) and Figure 3 (c) are combined into Figure 4 (a). Similarly, for the basic square wave signal with a period of \(T2\), Figure 3 (b) or Figure 3 (d) is directly output with a probability of \(P(f2)=(1 / 2)×(2 / 3)=1 / 3\). When the current signal is combined with its \(180°\) inverted square wave with a probability of \(1 / 3\), the probability is \(P(0.5f2)=(1 / 2)×(1 / 3)=� / 6\). For example Figure 3 (b) and Figure 3 (d) are combined into Figure 4 (b), and the combined signal period is \(2T2\). After combination, the injection probabilities of the three frequency signals of \(0.5f1\), \(f1\), and \(f2\) are all \(1 / 3\).

[0057] Define the unit square wave function \(\varphi\) sqr :

[0058]

[0059]

[0060] In the formula, is the injection signal phase, is the random operator for random signal injection; \(t\) r (t, T i ) is the remainder of \(t\) divided by \(T\) i ; \(V\) i , \(T\) iis the amplitude and period of the injected random square wave voltage, i={1,2}; u ori 、u fin are the injected signals before and after combination respectively.

[0061] like Figure 3 As shown in , in order to ensure the consistency of the high-frequency induced current amplitude, the principle of equal volt-second area should be followed, that is, the product of the period and amplitude of different frequency signals is a fixed value. The combined signal is as follows Figure 4 shown.

[0062] Step 2: Inject a random square wave voltage signal on the estimated d-axis:

[0063]

[0064] Where, and To estimate the high frequency voltage of dq axis.

[0065] Step 3: Demodulate the position angle and speed, such as Figure 5 shown.

[0066] The motor phase current is sampled and the current i in the two-phase stationary coordinate system is obtained through abc / αβ (Clark) transformation. α 、i β , and then use the high-pass filter HPF to extract the high-frequency induced current i αh 、i βh :

[0067]

[0068]

[0069] Where V i 、T i is the amplitude and period of the injected random square wave voltage; i inj is the induced current corresponding to the injected square wave; is the phase of the induced current, L dh and L qh is the dq axis high frequency incremental inductance; Δθ is the position angle estimation error, and θ e is the actual position angle, is the estimated position angle; φ tri is the unit triangle wave function, represent The remainder when divided by 2.

[0070] Project the induced current into the measurement reference frame that lags the estimation reference frame by 45°:

[0071]

[0072] Where, ΔL=(L qh -L qh ) / 2, and To measure the high frequency current in the reference frame; is the Park transform, is the estimated position angle; i αh 、i βh is the extracted high-frequency induced current of αβ axis; L dh and L qh is the dq-axis high-frequency incremental inductance; Δθ is the position angle estimation error; when the position angle error Δθ is small enough, the above formula can be simplified to:

[0073]

[0074] Taking the absolute value of the above formula, we get:

[0075]

[0076] Then project it back to the two-phase stationary coordinate system:

[0077]

[0078] Where, and is the demodulated high-frequency current of the αβ axis. When the position angle error Δθ is small enough, it can be simplified to:

[0079]

[0080] Where θ e is the actual position angle.

[0081] Furthermore, in order to reduce the influence of the injection amplitude on the position estimation, the demodulated signal is normalized:

[0082]

[0083] Where, and is the normalized current signal; and is the demodulated αβ axis high frequency current.

[0084] The position error signal ε is:

[0085]

[0086] Where, and is the normalized current signal, is the estimated position angle, and Δθ is the position angle estimation error.

[0087] Finally, the estimated electrical angular velocity is obtained through the proportional-integral PI controller, and the estimated position angle can be obtained by integrating the electrical angular velocity.

[0088] Figure 6 This is the PSD analysis result. Compared with the random frequency injection method and the random phase injection method, the random frequency and phase square wave injection method has the largest expansion range of current power spectrum density, the smallest amplitude, and a more uniform distribution, which reduces noise.

[0089] Figure 7 The position angle experimental results of the proposed method are shown in Figure 2. The estimated error is about 7°, which is relatively low and has good control performance.

[0090] Figure 8 The speed test results of the proposed method are as follows. The estimated error is about 6r / min, which is relatively low and has good control performance.

[0091] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative uses of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0092] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A method for controlling the noise reduction of a permanent magnet synchronous motor without position sensor by injecting high frequency square wave voltage with random frequency phase, characterized in that: The following steps are involved: Step 1: Generate a random high-frequency square wave voltage signal by randomly combining four high-frequency square waves with different frequencies and phases into eight high-frequency square waves with different frequencies and phases, and adjust the probability of each frequency square wave to be equal; Step 2: Inject the voltage signal into the estimated reference frame, and inject a high-frequency square wave voltage signal with completely random frequency and phase into the estimated d-axis; Step 3: Demodulate the position angle and speed: The high-frequency response current is extracted through a high-pass filter (HPF). To reduce the impact of the injected amplitude on the position estimation, it is normalized and then passed through an orthogonal phase-locked loop to obtain the position error signal. Finally, the estimated electrical angular velocity is obtained through a proportional-integral (PI) controller. The estimated position angle is obtained by integrating the electrical angular velocity.

2. The method according to claim 1, characterized in that The waveform of the square wave in step 1 is: Define the unit square wave function φ sqr : Where, is the injection signal phase, is the random operator for random signal injection; t r (t,T i ) is time t divided by T i The remainder of V i 、T i is the amplitude and period of the injected random square wave voltage, i={1,2}; u ori 、u fin They are the injection signals before and after combination respectively. The combination module is introduced in detail in the specific implementation part.

3. The method according to claim 1, characterized in that The principle of generating the square wave voltage signal in step 1 is: In order to ensure the consistency of the high-frequency induced current amplitude, the principle of equal volt-second area should be followed, that is, the product of the period and amplitude of different frequency signals is a fixed value.

4. The method according to claim 1, wherein In step 2, a random square wave voltage signal is injected into the estimated d-axis: Where, and To estimate the high frequency voltage of dq axis.

5. The method according to claim 1, characterized in that The high-frequency response current of step 3 is: sampling the motor phase current, and obtaining the current i in the two-phase stationary coordinate system through abc / αβ (Clark) transformation α 、i β , and then use the high-pass filter HPF to extract the high-frequency induced current i αh 、i βh : Where V i 、T i is the amplitude and period of the injected random square wave voltage; i inj is the induced current corresponding to the injected square wave; is the phase of the induced current, L dh and L qh is the dq axis high frequency incremental inductance; Δθ is the position angle estimation error, and θ e is the actual position angle, is the estimated position angle; φ tri is the unit triangle wave function, represent The remainder when divided by 2.

6. The method according to claim 5, characterized in that The processing process of the high-frequency induced current in step 3 is as follows: Project the induced current into the measurement reference frame that lags the estimation reference frame by 45°: Where, ΔL=(L qh -L qh ) / 2, and To measure the high frequency current in the reference frame; is the Park transform, is the estimated position angle; i αh 、i βh is the extracted high-frequency induced current of αβ axis; L dh and L qh is the dq-axis high-frequency incremental inductance; Δθ is the position angle estimation error; when the position angle error Δθ is small enough, the above formula can be simplified to: Taking the absolute value of the above formula, we get: Then project it back to the two-phase stationary coordinate system: Where, and is the demodulated αβ axis high frequency current; when the position angle error Δθ is small enough, it can be simplified to: Where θ e is the actual position angle.

7. The method according to claim 1, characterized in that In order to reduce the influence of the injection amplitude on the position estimation, the demodulated signal is normalized: Where, and is the normalized current signal; and is the demodulated αβ axis high frequency current.

8. The method according to claim 1, characterized in that The position error signal ε is: Where, and is the normalized current signal, is the estimated position angle, and Δθ is the position angle estimation error.

Images