Sensorless control method for induction motor based on rotor back electromotive force observation

Abstract

The invention discloses an induction motor sensorless control method based on rotor back electromotive force observation, and belongs to the technical field of power generation, power transformation or power distribution. The crucial rotor back electromotive force can be calculated in one sampling period, and no integral and differential operation is involved, so that the problems of integral initial value, error accumulation, differential noise and the like in a traditional flux observer are solved. The invention respectively provides an alpha-beta coordinate system rotor back electromotive force observation method, a d-q coordinate system rotor back electromotive force observation method and a corresponding rotor rotating speed extraction scheme, the simplest method only needs to calculate the d-axis rotor back electromotive force, and the requirement on the performance of a processor is greatly reduced. In addition, the invention further provides a sensorless control method for low-speed operation, and the sensorless control of the full speed domain can be achieved by combining the sensorless control method. When the method is used for sensorless control of the induction motor, the operation complexity of a control system can be effectively reduced, and good dynamic performance and steady-state precision are achieved.

Description

TECHNICAL FIELD

[0001] The application relates to a speed sensorless control method of a three-phase induction motor (IM), in particular to an induction motor sensorless control method based on rotor back electromotive force observation, and belongs to the technical field of power generation, power transformation or power distribution. BACKGROUND

[0002] In the field of modern motor control, permanent magnet synchronous motor (PMSM) and induction motor are the two most widely used motor types. PMSM has the advantages of high efficiency and high power density, but at the same time, PMSM uses permanent magnet material to generate magnetic field, and the magnetic conductivity is low, which is not conducive to the control of the rotor magnetic field, so the speed regulation range of PMSM is limited; the permanent magnet flux of the rotor of PMSM will decrease with the increase of temperature, and even irreversible demagnetization will occur, which reduces the performance of the motor and even makes the motor unusable, so the use scene of PMSM is limited; the permanent magnet of PMSM generally uses rare earth material, and the price is high, so the cost of PMSM driving system is high. Compared with PMSM, induction motor has the advantages of low cost, high reliability, wide speed regulation range, strong overload capacity and tolerance to large temperature changes, and is widely used in industrial field and electric vehicles.

[0003] The induction motor control generally adopts the rotor flux oriented control (FOC), which needs to obtain accurate flux position information. In addition, in order to realize the control of the speed of the induction motor, the induction motor control system can use a position sensor to obtain the speed of the motor. The common position sensors are mainly encoders and rotary transformers, but such mechanical position sensors have application limitations and increase the cost of the motor system. Therefore, the induction motor sensorless control technology emerges as the times require, hereinafter referred to as sensorless control. At present, scholars at home and abroad have proposed various induction motor sensorless control methods, and they have been applied in large direct-drive wind power systems, ship propulsion systems, railway vehicle traction systems and coal mine belt conveyors and other drive systems. However, the induction motor sensorless vector control system still has many difficulties and challenges, such as the requirement that the induction motor can stably run at low speed in some occasions. However, with the decrease of the terminal voltage of the motor, the current sensorless control system will have a resonance problem, which will cause the problem that the rotor speed cannot be observed. According to whether the rotor flux position and speed observer is closed loop, the speed observer can be divided into two types: open loop observer and closed loop observer. The structure of the open loop observer is generally simple, but the traditional open loop observer has integral or differential operation. The integral operation has the problems of initial value and error accumulation, while the differential operation amplifies the noise, and the use of a filter limits the dynamic performance of the control system; the closed loop observer is based on the mathematical model of the motor, and generally needs to design an adaptive rate to estimate the rotor speed and flux, and according to the different adaptive algorithms, the adaptive rate design method can be divided into: model reference adaptive method, sliding mode observation method, extended Kalman filter algorithm and artificial neural network algorithm, etc., but these methods are usually complex, increase the frequency coupling between the closed loop observer and the controller, and to some extent, aggravate the occurrence of resonance phenomenon.

[0004] The most effective method often has the simplest form structure, that is, the "less is more" principle in the industry. In accordance with the principle, the present application aims to propose an open-loop rotor back-EMF observation method applicable to induction motors, in order to avoid the inherent shortcomings of the traditional open-loop and closed-loop flux observers. SUMMARY

[0005] The purpose of this invention is to address the shortcomings of existing sensorless control technologies for induction motors by proposing a sensorless control method for induction motors based on rotor back electromotive force (EMF) observation. The proposed back EMF observer is an open-loop observer that does not require any differentiation or integration calculations. By combining normalized phase-locked loop (PLL) technology and slip frequency calculation methods, rotor flux linkage position and speed information are extracted from the back EMF obtained by the back EMF observer. This solves the technical problems of initial integral value, error accumulation, and noise amplification caused by integration or differentiation operations in traditional open-loop flux linkage observers, thus achieving the invention objective of sensorless control of induction motors across the entire speed range.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0007] A sensorless control method for induction motors based on rotor back EMF observation includes: a sensorless control strategy in the medium-to-high speed domain and a sensorless control strategy in the low-speed domain. The medium-to-high speed domain sensorless control strategy constructs a rotor back EMF observer for the induction motor, considering rotor inductance and stator-rotor mutual inductance. The observed rotor back EMF value output by the observer is normalized using a quadrature phase-locked loop (PLL) to obtain estimated rotor flux linkage position angle and rotor flux linkage rotational angular velocity. The rotational speed is extracted from the estimated rotor flux linkage rotational angular velocity, considering slip frequency. Current inner-loop control and speed outer-loop control are performed based on the estimated rotor flux linkage position angle and the extracted rotational speed. The low-speed domain sensorless control strategy extracts the estimated rotor flux linkage position angle from a given rotational speed, using a rotor back EMF observer and a normalized PLL to obtain the estimated rotor flux linkage rotational angular velocity. The rotational speed is extracted from the estimated rotor flux linkage rotational angular velocity, considering slip frequency. Current inner-loop control and speed outer-loop control are performed based on the estimated rotor flux linkage position angle and the extracted rotational speed.

[0008] As a further optimization of the sensorless control method for induction motors based on rotor back EMF observation, the sensorless control strategy in the medium-to-high speed domain considers the rotor inductance and the mutual inductance between the stator and rotor to construct a rotor back EMF observer for the induction motor. Specifically, it considers the rotor inductance and the mutual inductance between the stator and rotor to establish a rotor back EMF observer for the induction motor in the α-β coordinate system. ,in, , The observed value of the rotor back electromotive force in the α-β coordinate system. , , , These are the equivalent self-inductance of the stator winding, the equivalent self-inductance of the rotor winding, the equivalent mutual inductance between the stator and rotor, and the stator winding resistance, respectively. The leakage flux coefficient of the motor. , , , The stator voltage and current are projected onto the α-β axis. This is an estimated value for the angular velocity of the rotor flux linkage.

[0009] As a further optimization of the sensorless control method for induction motors based on rotor back EMF observation, the sensorless control strategy in the medium-to-high speed domain involves normalizing the rotor back EMF observation value output by the rotor back EMF observer using a quadrature phase-locked loop to obtain the estimated rotor flux linkage position angle and the estimated rotor flux linkage rotational angular velocity. Specifically: ,in, This is the estimated value of the rotor flux linkage position angle. This represents the normalized orthogonal phase-locked loop operation.

[0010] As a further optimization of the sensorless control method for induction motors based on rotor back EMF observation, the sensorless control strategy in the medium-to-high speed domain considers the rotor inductance and the mutual inductance between the stator and rotor to construct a rotor back EMF observer for the induction motor. Specifically, it considers the rotor inductance and the mutual inductance between the stator and rotor to establish and construct a rotor back EMF observer for the induction motor in the estimated dq coordinate system. ,in, and The back electromotive force of the rotor is the observed value in the dq coordinate system for estimation. , , , These are the equivalent self-inductance of the stator winding, the equivalent self-inductance of the rotor winding, the equivalent mutual inductance between the stator and rotor, and the stator winding resistance, respectively. The leakage flux coefficient of the motor. , , , The stator voltage and current are projected onto the dq axis. This is an estimated value for the angular velocity of the rotor flux linkage. Rotor back electromotive force in α-β coordinate system , amplitude, This represents the actual rotor flux linkage position. This is the estimated value of the rotor flux linkage position angle.

[0011] As a further optimization of the sensorless control method for induction motors based on rotor back EMF observation, the sensorless control strategy in the medium-to-high speed domain normalizes the rotor back EMF observation value output by the rotor back EMF observer using a quadrature phase-locked loop to obtain the estimated rotor flux linkage position angle and the estimated rotor flux linkage rotational angular velocity. Specifically:

[0012] The input phase-locked loop is subjected to PI regulation so that is kept near 0, the rotor flux position angle estimate value and the rotor flux rotational angular velocity estimate value are obtained; or, The

[0013] input phase-locked loop is subjected to PI regulation so that is kept near 0, the rotor flux position angle estimate value and the rotor flux rotational angular velocity estimate value are obtained; or, The input phase-locked loop is subjected to PI regulation so that

[0014] is kept near 0, the rotor flux position angle estimate value and the rotor flux rotational angular velocity estimate value are obtained. The input phase-locked loop is subjected to PI regulation so that is kept near 0, the rotor flux position angle estimate value and the rotor flux rotational angular velocity estimate value are obtained.

[0015] As a further optimization scheme of the induction motor sensorless control method based on rotor back-EMF observation, in the no-speed-sensor control strategy in the medium and high speed domain, the slip frequency is extracted from the rotor flux rotational angular velocity estimate value, specifically:

[0016] The rotor resistance estimated slip frequency , , is the rotor electromagnetic time constant, , is the rotor resistance, is the rotor flux d-axis component;

[0017] The rotor rotational angular velocity is obtained by using the relationship among the slip frequency , the rotor rotational angular velocity , ;

[0018] The rotor rotational angular velocity is converted into the rotor rotational mechanical angular velocity , , is the number of pole pairs;

[0019] The rotor rotational mechanical angular velocity is converted into the rotor speed , .

[0020] ​As a further optimization of the sensorless control method for induction motors based on rotor back EMF observation, the sensorless control strategy in the low-speed domain considers extracting the rotor flux linkage position angle estimate from the given speed using the slip frequency. Specifically, the rotor rotational angular velocity corresponding to the given speed is obtained, and the rotor rotational angular velocity corresponding to the given speed and the slip frequency are summed and integrated to obtain the rotor flux linkage position angle estimate.

[0021] As a further optimization of the sensorless control method for induction motors based on rotor back EMF observation, a speed hysteresis control method is adopted to switch the sensorless control strategy in the low-speed domain to a sensorless control strategy in the medium- and high-speed domains.

[0022] The present invention, by adopting the above technical solution, has the following beneficial effects:

[0023] 1. This invention first proposes a back EMF observer in the α-β coordinate system. This back EMF observer is an improvement on the traditional rotor flux linkage voltage model of an induction motor. In addition to considering the stator inductance and resistance, it also needs to consider the rotor inductance, resistance, and mutual inductance between the stator and rotor. Since the back EMF is the differential of the flux linkage, it also contains information about the rotor position and speed. By combining this observer with a normalized orthogonal phase-locked loop, the position angle and speed of the rotor flux linkage are estimated. This avoids the problems of initial integral value, error accumulation and noise amplification caused by integration and differentiation operations in traditional open-loop observers such as rotor flux linkage voltage model observers.

[0024] 2. Based on the back EMF observer in the α-β coordinate system, this invention further proposes a back EMF observer in the rotor flux estimation coordinate system. Furthermore, considering the complexity of the algorithm for extracting rotor flux position and speed, three methods are presented for… Different processing methods will vary in terms of computational load and system performance. Therefore, the appropriate method can be selected based on actual conditions such as cost budget and performance requirements. There are methods with low computational load that can run on low-end controllers, as well as methods with high performance and high computational load that are suitable for running on high-end controllers.

[0025] 3. This invention has simple requirements for sensorless operation of motor at low speeds. The combination of the two can realize sensorless control in the full speed range. The control method is simple and can be applied to MCUs with poor performance to reduce the cost of the control system. It can also be applied to MCUs with better performance to achieve higher control performance. Attached Figure Description

[0026] Figure 1 This is a block diagram of the method for estimating the rotor flux linkage position and speed of an induction motor in the α-β coordinate system proposed in this invention.

[0027] Figure 2 This is a block diagram of the method for estimating the position and speed of the rotor flux linkage of an induction motor in the rotor flux linkage estimation coordinate system proposed in this invention.

[0028] Figure 3 This is a block diagram of the rotor flux linkage position calculation method in the low-speed domain proposed in this invention.

[0029] Figure 4 This is an overall block diagram of the sensorless control strategy for induction motors proposed in this invention.

[0030] Figure 5 The waveform is a simulated observation of the rotor back electromotive force at a motor speed of 450 rpm using the method of this invention.

[0031] Figure 6 The waveform is a simulated observation of the rotor back electromotive force at a motor speed of 1400 rpm using the method of this invention.

[0032] Figure 7 The actual speed waveform of the motor under full-speed range variable speed and variable load using the method of this invention is shown.

[0033] Figure 8 The method of this invention is used to estimate the speed simulation waveform when the motor is changing speed and load across the entire speed range. Figure 9 The simulation waveform shows the speed estimation error when the motor is subjected to full-speed and load changes using the method of this invention.

[0034] Explanation of the labels in the diagram: It is the stator voltage, which can be the α-β axis stator voltage or the dq axis voltage, depending on the coordinate system used; It is the stator current, which can be the α-β axis stator current or the dq axis stator current, depending on the coordinate system used; It is the rotor back electromotive force, which can be the α-β axis or the dq axis back electromotive force depending on the coordinate system used. Detailed Implementation

[0035] To more clearly illustrate the implementation method and technical advantages of the present invention, the technical solution of the invention will be described in detail below with reference to the accompanying drawings.

[0036] If we ignore the effects of stator core saturation and motor parameter changes during the operation of the induction motor, and assume that the three-phase windings of the motor are identical and spatially symmetrical with a phase difference of 120°, then the traditional rotor flux linkage voltage model observer is as shown in equation (1):

[0037]

[0038] In equation (1), , These are the observed values ​​of the rotor flux linkage in the α-β coordinate system. , , , These are the equivalent self-inductance of the stator winding, the equivalent self-inductance of the rotor winding, the equivalent mutual inductance between the stator and rotor, and the stator winding resistance. It is the leakage flux coefficient of the motor. , , , These are the stator voltage and current projected onto the α-β axis.

[0039] After obtaining the rotor flux linkage from equation (1), the position angle of the rotor flux linkage is generally obtained by calculating the arctangent of equation (2):

[0040]

[0041] Traditional sensorless control implemented using equations (1) and (2) is relatively simple, requiring only the sampling of the stator current. The stator voltages along the α-β axes can be given values, thus eliminating the need for phase voltage sampling. Furthermore, the calculation does not require rotor resistance, but only the stator winding resistance. It is also easy to obtain. However, this method has the following disadvantages: 1. Equation (1) contains a pure integral element, which inevitably introduces problems such as the initial value of the flux linkage, error accumulation, and integral drift, resulting in inaccurate flux linkage observation; 2. Using the arctangent function will amplify the noise in the rotor flux linkage observation value, causing the obtained rotor flux linkage position and speed to fluctuate greatly, thereby affecting the performance of the control system.

[0042] Therefore, by differentiating both sides of equation (1) and defining the differential of the rotor flux linkage as the rotor back electromotive force, we can construct the rotor back electromotive force observer shown in the following equation:

[0043]

[0044] In equation (3), , It is the observed value of the rotor back electromotive force in the α-β coordinate system. This is an estimate of the rotating electrical angular velocity of the rotor flux linkage, which is obtained through a normalized quadrature phase-locked loop (PLL) as shown in the following equation. The structural diagram of the normalized quadrature PLL is included in... Figure 1 Compared to the traditional arctangent calculation method, the normalized quadrature phase-locked loop (PLL) can obtain a relatively smooth speed estimate and also achieves decoupling between PLL performance and speed, further improving the overall system performance.

[0045]

[0046] Transforming equation (3) into the estimated synchronous rotating coordinate system (estimated dq coordinate system) leads to a second no-speed control scheme, where the back electromotive force observation value in the estimated dq coordinate system can be expressed as:

[0047]

[0048]

[0049] In equations (5) to (6), , , , These are the stator voltage and current projected onto the dq axis; It is the rotor back electromotive force , The amplitude is proportional to the synchronous electric angular velocity; This is the actual rotor flux linkage position. Clearly, when the estimated rotor flux linkage position... hour, Therefore, all you need to do is... The relevant data is processed and used as input to the phase-locked loop (PLL), and then regulated by a PI controller to... By keeping the position near 0, the rotor flux linkage position can be estimated, and the phase-locked loop structure at this time is as follows: Figure 2 As shown.

[0050] Furthermore, the above-mentioned The relevant data processing methods can be specifically divided into the following three approaches.

[0051] Type 1 The method for performing relevant data processing: directly... As the input to the phase-locked loop. This method has the least computational cost; however, due to the rotor back EMF amplitude... The dynamic response is proportional to the synchronous electric angular velocity. Therefore, under this processing method, the performance of the phase-locked loop (PLL) varies with the motor speed. It exhibits poor dynamic response but good steady-state performance at low speeds, and fast dynamic response but poor steady-state performance at high speeds. However, for applications with a small speed range, this processing method has the advantage of simple computation, making it very suitable for applications where the requirements for system dynamic performance are not high and the controller performance is limited.

[0052] Type 2 Methods for performing relevant data processing: As the input to the phase-locked loop (PLL). Under the premise that the PLL parameters are appropriate, this method... When converging to near 0, the dynamic and steady-state performance of the phase-locked loop shows smaller differences with changes in rotational speed. Compared to the first method, this method requires less computation due to the need to divide by... The computational load will increase slightly, but the control performance will be significantly improved.

[0053] The third type of pair Methods for performing relevant data processing: This method serves as the input to the phase-locked loop (PLL). When using this method, the dynamic and steady-state performance of the PLL remains unchanged regardless of motor speed variations, completely decoupling the PLL performance from the speed. It is equivalent to the method in the α-β coordinate system. Compared to the first method, this method adds square root and division operations, making it suitable for applications requiring high dynamic performance, a wide speed range, and high controller performance.

[0054] After obtaining the rotor flux position and angular velocity, the slip frequency of the motor is estimated using equation (7). :

[0055]

[0056] In equation (7), It is the rotor electromagnetic time constant. , It is the rotor resistance. It is the d-axis component of the rotor flux linkage.

[0057] Using slip frequency Rotor rotational angular velocity and rotor flux rotational angular velocity The relationship between these three factors can be used to obtain the rotor's angular velocity. :

[0058]

[0059] The angular velocities in equation (8) are all electrical angular velocities. To obtain the mechanical angular velocity of the rotor rotation, the number of pole pairs of the motor also needs to be considered. The mechanical angular velocity of the rotor can be obtained using equation (9). :

[0060]

[0061] The method proposed in this invention is only suitable for sensorless control in the high-speed domain of induction motors. For zero-speed start-up, the position of the rotor flux linkage... Rotor rotational angular velocity corresponding to a given rotor speed With slip frequency It is obtained by adding and integrating, as shown in the following method. Figure 3 As shown; the speed closed-loop control still uses Figure 1 or Figure 2The rotor flux linkage rotational angular velocity is estimated using a rotor back EMF observer and a normalized quadrature phase-locked loop. The rotational speed is then extracted from this estimated value, taking into account the slip frequency. The zero-speed start-up method can be switched to the mid-to-high-speed domain speed-free control method described in this patent via speed hysteresis control.

[0062] The overall block diagram of the sensorless control method for a three-phase induction motor proposed in this invention is as follows: Figure 4 As shown, the control system mainly consists of two current PI regulators, one speed PI regulator, current sampling, coordinate transformation, SVPWM, a three-phase inverter, a back EMF observer, and a PLL. The q-axis current setpoint... This is the output of the speed PI regulator, and the d-axis current setpoint. Then the calculation is performed using equation (10):

[0063]

[0064] In equation (10), The d-axis current is the given value of the rotor flux fundamental amplitude. The d-axis current is calculated from the rotor flux given value using equation (10) and is used to control the rotor flux magnitude of the motor. The q-axis current is the output of the speed PI regulator and is used to control the torque output of the motor. The d-axis and q-axis current given values ​​obtained above are compared with the d-axis and q-axis sampled currents after CLARK and PARK transformations. , The difference between the d-axis and q-axis currents is calculated separately and then fed into their respective current PI regulators, thereby achieving control of the inner current loop of the motor. The d-axis and q-axis current PI regulators generate the setpoint voltages for the d-axis and q-axis, respectively. , The given voltage is first transformed to the α-β coordinate system by inverse PRAK, and then sent to the SVPWM module to generate a switching control signal to control the inverter to output the corresponding voltage vector.

[0065] Since this is a sensorless control system, the motor speed and rotor flux position need to be obtained using the method described in this paper. Specifically, in the medium and high-speed domains, a back-EMF observer is used to obtain back-EMF observations, and then a normalized PLL is used to extract rotor position information. The rotor back-EMF observer mentioned in this invention has two types, established in the α-β coordinate system and the rotor flux estimation coordinate system, respectively, with the following structures: Figure 1 , 2 As shown. Depending on the coordinate system used, Figure 4 symbols in , , The meaning will be different when using Figure 1When observing the back electromotive force in the α-β coordinate system as shown, Stator voltage in α-β coordinate system , , Represents the stator current in the α-β coordinate system , , Represents the back electromotive force in the α-β coordinate system , When using Figure 2 When the back EMF observer in the rotor flux estimation coordinate system is shown, Represents the stator voltage in the rotor flux estimation coordinate system. , , Represents the stator current in the rotor flux estimation coordinate system. , , Represents the back electromotive force in the rotor flux estimation coordinate system. , .

[0066] The following section uses an α-β coordinate system back EMF observer as an example to detail the process of obtaining the rotor back EMF from the stator voltage and current, and then obtaining the rotor flux linkage position and speed:

[0067] First, use equation (3) to calculate the α-β axis back electromotive force of the motor. For easier viewing of equation (3), it is rewritten as follows:

[0068]

[0069] As can be seen from the above equation, the synchronous electric angular velocity estimate of the motor is required when calculating the back electromotive force along the α-β axis. However, as can be seen from equation (4), to obtain Therefore, the back electromotive force along the α-β axis must be obtained first. , This forms an algebraic ring. Therefore, when implementing this method in an MCU, the back electromotive force along the α-β axis should be calculated first in the first control cycle. An initial value can be provided, which is usually set to 0. The calculation is complete. , The PLL is used to calculate the synchronous angular velocity of the motor according to equation (4). Estimated location of magnetic flux Then, in the second control cycle, the calculation of equation (11) uses the synchronous electric angular velocity calculated by the PLL in the first control cycle. The calculations for other times follow the same pattern, that is, the back electromotive force at time k is calculated according to equation (11). , In practice, the synchronous electric angular velocity calculated at time (k-1) is used. Therefore, equation (11) can be specifically written in the following form:

[0070]

[0071] Because of the large mechanical inertia of the motor, it can generally be assumed that the motor speed does not change much within a control cycle; therefore, the synchronous speed from the previous moment is used. It is feasible.

[0072] Using equations (12) and (4), the corresponding synchronous electric angular velocity estimate can be calculated in each control cycle. and the estimated position of rotor flux Estimated position of rotor flux This is used to transform the sampled current in the abc phase coordinate system to the dq coordinate system, and to transform the dq-axis setpoint voltage output from the current PI regulator to the α-β coordinate system, for SVPWM modulation to control the motor. Based on the calculated synchronous electric angular velocity estimate... The synchronous speed of a motor can be calculated from the number of pole pairs. However, due to the slip of an induction motor, its speed will be slightly lower than the synchronous speed, so it cannot be directly used for speed feedback in a speed loop PI regulator. To obtain the actual speed of the motor, it is also necessary to calculate the real-time slip of the motor using equation (7), and then calculate the rotor rotational angular velocity using the relationship between the synchronous electrical angular velocity, slip, and rotor rotational angular velocity as shown in equation (8). The rotor rotational angular velocity obtained by equation (8) is still an electrical angular velocity, which is a fraction of the rotor's mechanical rotational angular velocity. The controlled object is generally the motor's rotational speed, not its angular velocity, therefore, the rotor's mechanical angular velocity also needs to be calculated using the following formula. Converted to motor speed:

[0073]

[0074] Obtain an estimate of the motor speed. Then, it is used in the speed loop to calculate the difference with the given speed and then fed into... Figure 4 The speed PI regulator shown generates the q-axis current setpoint.

[0075] Thus, the execution details of the sensorless control method for induction motors based on a back EMF observer in the α-β coordinate system proposed in this invention within one control cycle have been described in detail. Following the above-described sensorless control method execution flow, simulation calculations were performed, and the corresponding simulation waveforms are as follows: Figure 5 , Figure 6 , Figure 7 ,Figure 8 , Figure 9 As shown. Among them, Figure 5 , Figure 6 The waveforms are the observed back EMF at motor speeds of 500 rpm and 1400 rpm, respectively. It can be seen that the observed back EMF value at 500 rpm contains more high-order harmonics, causing some distortion. Moreover, this distortion gradually increases as the speed decreases further.

[0076] like Figure 5 , Figure 6 As shown, the higher harmonics are always relatively small at medium and high speeds in the motor, and the normalized PLL transfer function shown in equation (4) is:

[0077]

[0078] A normalized PLL is essentially a second-order low-pass filter, which can filter out the high-order harmonics contained in the back electromotive force and reduce their influence.

[0079] To address the issue that the method proposed in this invention is only suitable for sensorless control in the high-speed domain of induction motors, the rotor flux linkage position is used at low speeds. Figure 3 The method shown is used to obtain and estimate the rotational speed. Figure 1 or Figure 2 To achieve stable switching between the low-speed domain and the medium- and high-speed domains, this invention proposes using a speed hysteresis method for strategy switching.

[0080] To demonstrate that the sensorless control method proposed in this invention can achieve sensorless control of an induction motor across the entire speed range, dynamic simulations were performed under varying speed and load conditions. The inverter dead time was set to 2 μs, the switching frequency to 10 kHz, and the viscous friction coefficient of the motor during rotation was B = 0.005752 Nms. The simulation waveforms are shown below. Figure 7 , Figure 8 , Figure 9 As shown. Figure 7 , Figure 8 These are the waveforms of the actual and estimated motor speeds, respectively. The motor is first controlled to run at an extremely low speed of 45 rpm, then sequentially accelerated to a medium speed of 500 rpm and a high speed of 1400 rpm. The speed is further increased at t=1s and t=2s. and reduce The load was used to verify the dynamic tracking performance and disturbance rejection performance of the sensorless control proposed in this invention. The simulation results show that the sensorless control method mentioned in this invention has good tracking performance and disturbance rejection performance, and can achieve sensorless control of the induction motor across the entire speed range. To further analyze the performance of the sensorless control method mentioned in this invention, the speed estimation error throughout the above simulation is as follows:Figure 9 As shown, under the conditions of large load increases and decreases at t=1s and t=2s, the speed error quickly converges to near 0, indicating that the sensorless control method mentioned in this invention estimates a large speed bandwidth, which can meet the application scenarios with high dynamic performance requirements. Furthermore, at the extremely low speed of 45rpm, the speed estimation error fluctuates significantly, but its average value remains 0, thus achieving stable low-speed operation of the motor. Figure 7 As shown, the motor speed fluctuates significantly at 45 rpm and is only suitable for light-load operation, which is a common problem for all model-based sensorless control methods.

[0081] This invention proposes back EMF observers in both the stationary coordinate system and the rotor flux estimation coordinate system. In fact, the back EMF observers in the two coordinate systems are equivalent because equation (5) is obtained by performing the following coordinate transformation on equation (3):

[0082]

[0083] Since the rotor flux linkage position is obtained through the method proposed in this invention, it is not the actual rotor flux linkage position but rather its estimated value. Therefore, the back EMF observer after the coordinate transformation shown in Equation (15) is named "back EMF observer in rotor flux estimation coordinate system". The back EMF observer shown in Equation (3) and the back EMF observer shown in Equation (5) are established in the α-β coordinate system and the synchronous rotating coordinate system, respectively, and have the same electrical performance.

Claims

1. A sensorless control method for induction machines based on rotor back electromotive force observation, characterized in that, The application relates to a sensorless control strategy for a high-speed field-oriented induction motor, and a sensorless control strategy for a low-speed field-oriented induction motor. The sensorless control strategy for the high-speed field-oriented induction motor comprises the following steps: constructing a rotor back electromotive force observer of the induction motor by considering rotor inductance and stator-rotor mutual inductance; normalizing a rotor back electromotive force observation value output by the rotor back electromotive force observer by using a normalized quadrature phase-locked loop; obtaining a rotor flux position angle estimation value and a rotor flux rotation angular velocity estimation value; extracting a rotating speed from the rotor flux rotation angular velocity estimation value by considering a slip frequency; and performing current inner loop control and rotating speed outer loop control according to the rotor flux position angle estimation value and the extracted rotating speed. The sensorless control strategy for the low-speed field-oriented induction motor comprises the following steps: extracting a rotor flux position angle estimation value from a given rotating speed by considering a slip frequency; obtaining a rotor flux rotation angular velocity estimation value by using the rotor back electromotive force observer and the normalized quadrature phase-locked loop; extracting a rotating speed from the rotor flux rotation angular velocity estimation value by considering the slip frequency; and performing current inner loop control and rotating speed outer loop control according to the rotor flux position angle estimation value and the extracted rotating speed.

2. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 1, characterized in that, In the no-speed-sensor control strategy of the medium-high speed domain, rotor inductance and stator-rotor mutual inductance are considered to construct a rotor back electromotive force observer of the induction motor, specifically: rotor inductance and stator-rotor mutual inductance are considered to construct a rotor back electromotive force observer of the induction motor in an alpha-beta coordinate system, wherein, , is an observed value of the rotor back electromotive force in the alpha-beta coordinate system, , , , are equivalent self-inductance of the stator winding, equivalent self-inductance of the rotor winding, equivalent mutual inductance between the stator and the rotor, and resistance of the stator winding respectively, is a leakage coefficient of the motor, , , , are stator voltage and current projected to the alpha-beta axis, is an estimated value of the rotor flux rotation angular velocity.

3. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 2, characterized in that, In the no-speed-sensor control strategy of the medium-high speed range, the rotor back electromotive force observation value output by the rotor back electromotive force observer is normalized quadrature phase-locked loop, rotor flux position angle estimation value and rotor flux rotation angular velocity estimation value are obtained, specifically: Wherein, is the rotor flux position angle estimation value, represents the normalized quadrature phase-locked loop operation.

4. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 1, characterized in that, The rotor inductance and the rotor-stator mutual inductance are considered to construct a rotor back electromotive force observer of the induction motor, specifically: the rotor inductance and the rotor-stator mutual inductance are considered to construct a rotor back electromotive force observer of the induction motor in the estimation of the d-q coordinate system, wherein, and are observation values of the rotor back electromotive force in the estimation of the d-q coordinate system, , , , are the equivalent self-inductance of the stator winding, the equivalent self-inductance of the rotor winding, the equivalent mutual inductance between the stator and the rotor, and the resistance of the stator winding, is the magnetic leakage coefficient of the motor, , , , are the stator voltage and current projected onto the d-q axis, is an estimated value of the rotational angular velocity of the rotor flux linkage, is the amplitude of the rotor back electromotive force in the α-β coordinate system , , is the actual rotor flux linkage position, is an estimated value of the rotor flux linkage position angle.

5. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 4, characterized in that, In the sensorless control strategy for the high-speed field-oriented induction motor, the rotor back electromotive force observation value output by the rotor back electromotive force observer is normalized by using the normalized quadrature phase-locked loop, and the rotor flux position angle estimation value and the rotor flux rotation angular velocity estimation value are obtained. The input phase-locked loop is locked to the input signal, and the output signal is a signal with the same frequency and phase as the input signal. The input phase-locked loop is locked to the input signal, and the output signal is a signal with the same frequency and phase as the input signal. The PI regulation is performed so that the error is maintained near 0, and the rotor flux position angle estimate and the rotor flux rotation angular velocity estimate are obtained. The input phase-locked loop is locked to the input signal, and The input phase-locked loop is locked to the input signal, and the output signal is a signal with the same frequency and phase as the input signal. The input phase-locked loop is locked to the input signal, and the output signal is a signal with the same frequency and phase as the input signal. The PI regulation is performed so that the error is maintained near 0, and the rotor flux position angle estimate and the rotor flux rotation angular velocity estimate are obtained. The input phase-locked loop is locked to the input signal, and Will Input phase-locked loop, for PI adjustment makes Keep it near 0, and obtain the estimated values ​​of the rotor flux position angle and the rotor flux rotation angular velocity.

6. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 3 or 5, characterized in that, In the sensorless control strategy for the high-speed field-oriented induction motor, the rotating speed is extracted from the rotor flux rotation angular velocity estimation value by considering the slip frequency. Consider slip frequency of rotor resistance estimation , , is the rotor electromagnetic time constant, , is the rotor resistance, is the rotor flux d-axis component; Utilizing slip frequency , rotor rotational angular velocity and rotor flux rotational angular velocity between the three to obtain the rotor rotational angular velocity , ; rotor rotation angular velocity converted into a mechanical angular velocity of rotor rotation , , is the number of pole pairs; Mechanical angular velocity of the rotor Converting into rotational speed , .

7. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 6, characterized in that, In the sensorless control strategy for the low-speed field-oriented induction motor, the rotor flux position angle estimation value is extracted from the given rotating speed by considering the slip frequency, and the specific method is as follows: a rotor rotation angular velocity corresponding to the given rotating speed is obtained according to the given rotating speed; and the rotor rotation angular velocity corresponding to the given rotating speed and the slip frequency are summed and integrated to obtain the rotor flux position angle estimation value.

8. The sensorless control method of an induction motor based on rotor back-EMF observation according to claim 7, characterized in that, The sensorless control strategy for the low-speed field-oriented induction motor is switched to the sensorless control strategy for the high-speed field-oriented induction motor by using a rotating speed hysteresis loop control mode.

Images