Permanent magnet synchronous motor random injection signal extraction method based on virtual coordinate transformation

Abstract

The application discloses a random injection signal extraction method of a permanent magnet synchronous motor based on virtual coordinate transformation and belongs to the technical field of position sensorless control of the permanent magnet synchronous motor. First, a voltage signal generator of a high-frequency pulse signal with randomly selected different pulse sequences is designed and generated, a general expression form of a high-frequency current response in a high-frequency virtual coordinate system is derived, and a high-frequency pulse voltage with different pulse sequences in the d-axis is estimated. Second, the high-frequency current response is demodulated in the high-frequency virtual coordinate axis, and position information and position estimation are extracted. The high-frequency pulse voltage signals with different pulse sequences are superimposed on the direct axis of the estimated two-phase rotating coordinate system, the high-frequency current response containing rotor position information is detected, the response signal is demodulated, the rotor speed and position information are acquired, and the dynamic performance and reliability are improved on the premise of retaining the position sensorless steady-state performance based on the high-frequency signal injection.

Description

Technical Field

[0001] This invention belongs to the field of sensorless control technology for permanent magnet synchronous motors, and particularly relates to a method for extracting randomly injected signals for permanent magnet synchronous motors based on virtual coordinate transformation. Background Technology

[0002] Due to their advantages such as simple structure, high power density, and high efficiency, permanent magnet synchronous motors have been gradually applied in many industrial fields, including electric vehicles, home appliances, and robotics. Precise rotor position is crucial for closed-loop control and can be extracted using mechanical position sensors or sensorless actuators. Sensorless control technology helps reduce system cost and size while improving system robustness, and this technology has been widely adopted. In particular, the high-frequency injection method is used in the low-speed range and under stationary conditions by injecting a high-frequency excitation voltage signal into the motor stator windings and demodulating the rotor position from the high-frequency current response.

[0003] However, the response current acquired from the current sensor contains both fundamental and high-frequency components. The high-frequency current response signal containing rotor position information needs to be separated using a bandpass filter. The introduction of a traditional bandpass filter limits the bandwidth of the system, thereby degrading the dynamic performance of the system and reducing the convergence speed and observation accuracy of the rotor position observer. Summary of the Invention

[0004] Purpose of the invention: To address the shortcomings and defects of existing technologies, this invention proposes a random high-frequency injection signal extraction method for permanent magnet synchronous motors based on virtual coordinate transformation. This method enables sensorless control of permanent magnet synchronous motors, solves the problems of high-frequency noise and phase delay caused by traditional filters, improves dynamic performance, and enhances estimation accuracy.

[0005] Technical solution: The present invention provides a method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation, comprising the following steps:

[0006] Step 1: Generate random numbers using a random number generator, and select the basic injection pulse voltage of the permanent magnet synchronous motor based on different pulse sequences according to different random numbers;

[0007] Step 2: Establish a mathematical model of the permanent magnet synchronous motor using the high-frequency signal injection method;

[0008] Step 3: Based on the basic injection pulse voltage in Step 1 and the mathematical model of the permanent magnet synchronous motor in Step 2, calculate the high-frequency current response of the high-frequency virtual coordinate axis through multiple coordinate transformations.

[0009] Step 4: Extract the DC component converted from the high-frequency response signal using an improved second-order generalized integrator. Replace the low-pass filter in the current loop with an improved second-order generalized integrator to extract the fundamental frequency current signal.

[0010] Step 5: Input the position deviation signal to the phase-locked loop and adjust it so that the rotor position estimation error is zero, thereby making the rotor position estimation value equal to the true value, and realizing sensorless control.

[0011] Further, step 1 specifically involves: generating random numbers using a random number generator, and selecting basic injection pulse voltages for different pulse sequences based on the different random numbers. In this invention, the two different basic injection pulse voltage sequences are as follows: Figure 1 As shown. Where "P" represents positive voltage injection, "N" represents negative voltage injection, and "0" represents zero voltage injection. Its expression is:

[0012]

[0013] in, This represents the basic injection pulse voltage, n = 0, 1, 2, ... To represent the period of the injected pulse signal, the operator R[] denotes random selection within the brackets. For simplicity, Figure 1 The sequences “PN0NP0” and “NP0PN0” shown in (a)-(b) are defined as sequences ab. This represents the pulse voltage signal within a period of an injected pulse signal, where x represents the sequence number, taken as a, b. Described as:

[0014]

[0015] in, This indicates the amplitude of the injected pulse signal. This represents time t divided by The remainder.

[0016] Furthermore, step 2 specifically involves establishing the voltage equation for the permanent magnet synchronous motor as follows:

[0017]

[0018] in, , These are the stator voltages. Axial components, , These are the stator currents. Axial components, R is the stator inductance, and R is the stator resistance. Represents permanent magnet flux linkage. This indicates the actual electric speed of the rotor.

[0019] The voltage equation simplifies to:

[0020]

[0021] In the formula, , , , These represent the high-frequency voltage and current of the dq axis, respectively.

[0022] Furthermore, step 3 specifically involves the transformation formula for converting the current in the stationary coordinate system to the current in the virtual coordinate system:

[0023]

[0024] In the formula, The rotation angle of the high-frequency virtual coordinate system. The estimated electric speed of the rotor; , These are the stationary axis components of the stator current. , These are the virtual axis components of the stator current;

[0025] Substituting into the high-frequency virtual transformation formula, we obtain the high-frequency response current in the virtual coordinate system:

[0026]

[0027] In the formula, , They are respectively High-frequency virtual axis high-frequency current component, sign() represents the sign function. This indicates the amplitude of the injected signal. For PWM modulation period, The angular frequency of the injected signal, It is an inductor;

[0028] at this time and It is still a high-frequency alternating component, but since the rotational angular frequency of the high-frequency virtual coordinate system is synchronized with the frequency of the injected signal, and the rotational speed of the coordinate system is consistent with the alternating speed of the high-frequency response current, the two are relatively stationary. Therefore, the high-frequency pulsation term is canceled out, and a stable DC component is finally obtained.

[0029] Further, step 4 specifically involves: accurately extracting the DC component converted from the high-frequency response signal using an improved second-order generalized integrator, thereby replacing the original bandpass filter and low-pass filter in the system; and then replacing the low-pass filter in the current loop with an improved second-order generalized integrator to extract the fundamental frequency current signal.

[0030] After improvement, the new state equation is:

[0031]

[0032] Where k represents the gain of the improved second-order generalized integrator. express Response current in virtual coordinate system This represents the high-frequency components containing rotor position extracted by the low-pass filter of the second-order generalized integrator. It is a state variable.

[0033] Performing a Laplace transform on it to eliminate intermediate variables yields the transfer function:

[0034]

[0035] When the cutoff angular frequency is set At that time, the DC component of the high-frequency response signal can be accurately extracted. After filtering, the component containing rotor position information corresponding to the retained high-frequency response current is represented as follows:

[0036]

[0037] in, For the estimated rotor position, ( ) represents the position estimation error.

[0038]

[0039] Furthermore, in step 5, a phase-locked loop (PLL) is selected to estimate the rotor position and speed. The position deviation signal is input to the PLL, and adjustments are made to ensure the rotor position estimation error is zero, thus making the obtained rotor position estimate equal to the true value, achieving sensorless control. The closed-loop transfer function of the PLL observer can be expressed as:

[0040]

[0041] In the formula and These are the proportional gain and integral gain of the PLL, respectively.

[0042] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0043] 1. It avoids the phase delay problem caused by traditional filters, reduces the bandwidth of the sensorless control system of the permanent magnet synchronous motor, and significantly improves the convergence speed and observation accuracy of the rotor position observer.

[0044] 2. High-frequency virtual coordinate system transformation is used instead of low-pass filtering, which reduces the cost and complexity of algorithm development. Attached Figure Description

[0045] Figure 1 This is a block diagram of the overall control system structure of the present invention;

[0046] Figure 2 This diagram illustrates the positional relationship between a two-phase stationary coordinate system, an actual two-phase synchronous rotating coordinate system, an estimated two-phase synchronous rotating coordinate system, and a high-frequency virtual coordinate system.

[0047] Figure 3 A block diagram illustrating the principle of a signal generator for high-frequency pulse signals with different pulse sequences;

[0048] Figure 4 The following are simulation waveforms of the estimated and actual speed of a permanent magnet synchronous motor: (a) is the simulation waveform of the estimated and actual speed of a permanent magnet synchronous motor with a traditional random signal injection scheme when the load changes abruptly from 0 to 1 Nm; (b) is the simulation waveform of the estimated and actual speed of a permanent magnet synchronous motor according to the present invention.

[0049] Figure 5 The following are simulation waveforms of the speed error of a permanent magnet synchronous motor: (a) is the simulation waveform of the speed error of the permanent magnet synchronous motor under the traditional random signal injection scheme when the load changes abruptly from 0 to 1 Nm; (b) is the simulation waveform of the speed error of the permanent magnet synchronous motor according to the present invention.

[0050] Figure 6 The following are simulation waveforms of the position error of a permanent magnet synchronous motor: (a) is the simulation waveform of the position error of a permanent magnet synchronous motor with a traditional random signal injection scheme when the load changes abruptly from 0 to 1Nm; (b) is the simulation waveform of the position error of a permanent magnet synchronous motor according to the present invention. Detailed Implementation

[0051] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0052] Figure 1 The control block diagram for a random high-frequency injection signal extraction method for permanent magnet synchronous motors based on virtual coordinate transformation includes the following steps:

[0053] Step 1: The signal generator randomly selects high-frequency pulse signals of different pulse sequences, as follows:

[0054] Random numbers are generated by a random number generator, and different basic injection pulse voltages for different pulse sequences are selected based on these random numbers. The two different basic injection pulse voltages in this invention are as follows: Figure 1 As shown. Where "P" represents positive voltage injection, "N" represents negative voltage injection, and "0" represents zero voltage injection. Its expression is:

[0055]

[0056] in, This represents the basic injection pulse voltage, n = 0, 1, 2, ... To represent the period of the injected pulse signal, the operator R[] denotes random selection within the brackets. For simplicity, Figure 1 The sequences “PN0NP0” and “NP0PN0” shown in (a)-(b) are defined as sequences ab. This represents the pulse voltage signal within a period of an injected pulse signal, where x represents the sequence number, taken as a, b. Described as:

[0057]

[0058] in, This indicates the amplitude of the injected pulse signal. This represents time t divided by The remainder.

[0059] Step 2: Establish the mathematical model of the permanent magnet synchronous motor using the high-frequency signal injection method. The process is as follows:

[0060] First, establish such as Figure 2 The coordinate relationship diagram shown is as follows. For actual synchronous rotating coordinate system, For a two-phase stationary coordinate system, For a high-frequency virtual coordinate system, To estimate the rotating coordinate system. and These represent the actual electric speed of the rotor and the position of the rotor magnetic pole d-axis, respectively. The superscript "^" indicates the estimated component. ( ) represents the position estimation error. denoted as the rotation angle of the high-frequency virtual coordinate system. The angular frequency of the high-frequency injected signal. The initial phase of the high-frequency signal must be completely consistent with the phase of the actual injected high-frequency voltage signal to avoid signal distortion caused by phase shift.

[0061] The voltage equation for a permanent magnet synchronous motor in a real synchronous rotating coordinate system is:

[0062]

[0063] in, , These are the stator voltages. Axial components, , These are the stator currents. Axial components, R is the stator inductance, and R is the stator resistance. Represents permanent magnet flux linkage.

[0064] In permanent magnet synchronous motor control systems, the injected high-frequency signal frequency is generally significantly higher than the motor's fundamental frequency, thus the influence of the motor's phase winding resistance can be approximately ignored. Furthermore, since high-frequency signal injection is suitable for zero-speed operation of permanent magnet synchronous motors, the motor's rotational angular frequency is very small, and the back electromotive force and cross-coupling terms in the voltage equation can be neglected.

[0065] The voltage equation can be simplified to:

[0066]

[0067] In the formula, , , , These represent the high-frequency voltage and current of the dq axis, respectively.

[0068] Step 3: Calculate the high-frequency current response in the high-frequency virtual coordinate system. The process is as follows:

[0069] according to Figure 1 The coordinate system relationship diagram shown is transformed with equation (4) to obtain the following result.

[0070]

[0071] Substituting formulas (1) and (2) into the equations, we can further derive...

[0072]

[0073] In the formula, sign() represents the sign function. , Each is a high-frequency α-β stationary axis current. The amount of change over a period of time.

[0074] Performing a high-frequency virtual coordinate transformation on the above equation yields the high-frequency response current in the virtual coordinate system.

[0075]

[0076] In the formula, , They are respectively High-frequency virtual axis high-frequency current component;

[0077] As can be seen from equation (7), at this time and It remains a high-frequency alternating component, but because the rotational angular frequency of the high-frequency virtual coordinate system is synchronized with the frequency of the injected signal, the rotational speed of the coordinate system is consistent with the alternating speed of the high-frequency response current. Since they are relatively stationary, the high-frequency pulsation term... The components are completely canceled out, and a stable DC component is finally obtained through low-pass filtering.

[0078] Step 4, in high-frequency virtual The high-frequency current response is demodulated in a coordinate system to extract position information, as follows:

[0079] An improved second-order generalized integrator accurately extracts the DC component converted from the high-frequency response signal, thus replacing the original bandpass and low-pass filters in the system. Furthermore, the low-pass filter in the current loop is replaced with an improved second-order generalized integrator to extract the fundamental frequency current signal. The simplified block diagram of the improved system is shown below. Figure 3 As shown.

[0080] After improvement, the new state equation is:

[0081]

[0082] Where k represents the gain of SOGI-QSG, express Response current in virtual coordinate system This represents the high-frequency components containing rotor position extracted by the low-pass filter of the second-order generalized integrator. It is a state variable.

[0083] Performing a Laplace transform on it to eliminate intermediate variables yields the transfer function:

[0084]

[0085] When the cutoff angular frequency is set At this time, the transfer function is basically equivalent to a second-order Butterworth low-pass filter. It retains the robustness of the traditional second-order generalized integrator to parameter changes and has the excellent filtering characteristics of a low-pass filter, which can accurately extract the DC component of the high-frequency response signal. After the above filtering algorithm, the component of the high-frequency response current containing rotor position information is retained, and its expression is shown in equation (10).

[0086]

[0087] in, For the estimated rotor position, ( ) represents the position estimation error.

[0088] Based on the stable DC component of equation (10), the equivalent position error The expression is shown in (11). Wherein, This is a proportionality coefficient, which needs to be fine-tuned according to the amplitude of the high-frequency response signal.

[0089]

[0090] Step 5, Location Estimation

[0091] The position deviation signal is input to the phase-locked loop (PLL), and adjustments are made until the rotor position estimation error is zero. This ensures that the obtained rotor position estimate equals the true value, achieving sensorless control.

[0092]

[0093] In the formula and These are the proportional gain and integral gain of the phase-locked loop, respectively.

[0094] In the simulation model of this embodiment, the PWM frequency of the inverter is set to 10kHz, and the frequency of the high-frequency injection signal is set to 0.5kHz to 2kHz, which is much higher than the fundamental frequency of the motor.

[0095] Example 1

[0096] Figures 4-6 The waveforms obtained from the simulation model built in MATLAB Simulink are shown in Table 1, taking a 2×3 phase surface-mounted permanent magnet synchronous motor as an example.

[0097] Table 1 Parameters of 2×3 phase surface-mounted permanent magnet synchronous motor

[0098]

[0099] In this embodiment, the motor speed is given as 50 r / min, and the load is 1 Nm. From Figure 4 As can be seen, compared with traditional bandpass and lowpass filter extraction schemes, the dynamic response performance of the rotational speed is significantly improved. From... Figure 5 As can be seen, compared with traditional bandpass and lowpass filter extraction schemes, the dynamic speed error is significantly reduced. From... Figure 6 As can be seen, compared with the traditional bandpass and lowpass filter extraction schemes, the transient and steady-state position errors are significantly reduced. Therefore, compared with the traditional combination of bandpass and lowpass filters, the random high-frequency injection signal extraction method for permanent magnet synchronous motors based on virtual coordinate transformation of this invention improves the dynamic response performance after virtual coordinate transformation and reduces position and speed errors. The simulation results prove the effectiveness of the method of this invention.

[0100] In summary, this invention discloses a sensorless scheme for random signal injection into a permanent magnet synchronous motor based on high-frequency virtual coordinate transformation. This scheme employs a high-frequency virtual coordinate transformation strategy, converting the high-frequency response signal into a DC component, which is then accurately extracted using an improved second-order generalized integrator. This invention superimposes high-frequency pulse voltage signals of different pulse sequences onto the direct axis of an estimated two-phase rotating coordinate system. By detecting the high-frequency current response containing rotor position information and demodulating this response signal, rotor speed and position information are obtained. This improves dynamic performance and reliability while preserving the steady-state performance of sensorless systems based on high-frequency signal injection.

[0101] The above description is merely a preferred embodiment of the present invention. For those skilled in the art, various improvements and optimizations can be made to the present invention without departing from its basic principles, and these improvements and optimizations should also be included within the scope of protection of the present invention.

Claims

1. A method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation, characterized in that, Includes the following steps: Step 1: Generate random numbers using a random number generator, and select the basic injection pulse voltage of the permanent magnet synchronous motor based on different pulse sequences according to different random numbers; Step 2: Establish a mathematical model of the permanent magnet synchronous motor using the high-frequency signal injection method; Step 3: Based on the basic injection pulse voltage in Step 1 and the mathematical model of the permanent magnet synchronous motor in Step 2, calculate the high-frequency current response of the high-frequency virtual coordinate axis through multiple coordinate transformations. Step 4: Extract the DC component converted from the high-frequency response signal using an improved second-order generalized integrator. Replace the low-pass filter in the current loop with an improved second-order generalized integrator to extract the fundamental frequency current signal. Step 5: Input the position deviation signal to the phase-locked loop and adjust it so that the rotor position estimation error is zero, thereby making the rotor position estimation value equal to the true value, and realizing sensorless control.

2. The method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation according to claim 1, characterized in that, Step 1 specifically involves: generating random numbers using a random number generator, and selecting the basic injection pulse voltage for different pulse sequences based on the different random numbers. The two basic injection pulse voltages for different pulse sequences in this invention are shown in Figure 1. Here, "P" represents positive voltage injection, "N" represents negative voltage injection, and "0" represents zero voltage injection. Its expression is: ； in, This represents the basic injection pulse voltage, n = 0, 1, 2, ... The operator R[] represents the period of the injected pulse signal, indicating random selection within the brackets. This represents the pulse voltage signal within one period of an injected pulse signal, where x represents the sequence number, taken as a, b; Described as: ； in, This indicates the amplitude of the injected pulse signal. This represents time t divided by The remainder.

3. The method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation according to claim 1, characterized in that, Step 2 specifically involves establishing the voltage equation for the permanent magnet synchronous motor as follows: ； in, , These are the stator voltages. Axial components, , These are the stator currents. Axial components, R is the stator inductance, and R is the stator resistance. Represents permanent magnet flux linkage. This indicates the actual electrical speed of the rotor; The voltage equation simplifies to: ； In the formula, , , , These represent the high-frequency voltage and current of the dq axis, respectively.

4. The method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation according to claim 1, characterized in that, Step 3 specifically involves the transformation formula for converting current in the stationary coordinate system to current in the virtual coordinate system: ； In the formula, The rotation angle of the high-frequency virtual coordinate system. The estimated electric speed of the rotor; , These are the stationary axis components of the stator current. , These are the virtual axis components of the stator current; Substituting into the high-frequency virtual transformation formula, we obtain the high-frequency response current in the virtual coordinate system: ； In the formula, , They are respectively High-frequency virtual axis high-frequency current component, sign() represents the sign function. This indicates the amplitude of the injected signal. For PWM modulation period, The angular frequency of the injected signal, It is an inductor; at this time and It is still a high-frequency alternating component, but since the rotational angular frequency of the high-frequency virtual coordinate system is synchronized with the frequency of the injected signal, and the rotational speed of the coordinate system is consistent with the alternating speed of the high-frequency response current, the two are relatively stationary. Therefore, the high-frequency pulsation term is canceled out, and a stable DC component is finally obtained.

5. The method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation according to claim 4, characterized in that, Step 4 specifically involves: accurately extracting the DC component converted from the high-frequency response signal using an improved second-order generalized integrator, thereby replacing the original bandpass and low-pass filters in the system; and then replacing the low-pass filter in the current loop with an improved second-order generalized integrator to extract the fundamental frequency current signal. After improvement, the new state equation is: ； Where k represents the improved second-order generalized integrator gain, express Response current in virtual coordinate system This represents the high-frequency components containing rotor position extracted by the low-pass filter of the second-order generalized integrator. For state variables; Performing a Laplace transform on it to eliminate intermediate variables yields the transfer function: ； When the cutoff angular frequency is set At that time, the DC component of the high-frequency response signal is extracted, and after filtering, the component containing rotor position information corresponding to the retained high-frequency response current is represented as follows: ； in, For the estimated rotor position, ( ) represents the position estimation error.

6. The method for extracting randomly injected signals from a permanent magnet synchronous motor based on virtual coordinate transformation according to claim 1, characterized in that, In step 5, a phase-locked loop is used to estimate the rotor position and speed.

Images