Parameter identification-based synchronous reluctance motor full-speed-domain position-sensorless control method

Abstract

The invention discloses a parameter identification-based synchronous reluctance motor full-speed-domain position-sensorless control method, which relates to the technical field of motor control, and is characterized in that when a motor is in a zero-low-speed domain, a pulsating square wave voltage signal injection method is adopted to realize effective observation of a rotor position and good dynamic response performance; when a rotating speed transition interval is entered, an improved weighted switching strategy is adopted, so that the motor can smoothly switch two control strategies in the interval; when the motor enters a medium-high speed domain, rotor position angle observation is carried out by adopting a model reference adaptive control method, and high steady-state operation precision is achieved; however, in the whole operation process, inductance parameters influenced by magnetic circuit saturation and cross saturation are nonlinear, online identification is carried out by adopting a model reference adaptive method, and the estimation precision of the rotor position angle is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of motor control, and in particular to a full-speed range position sensorless control method for a synchronous reluctance motor based on parameter identification. Background Art

[0002] Synchronous reluctance motors are widely used in various fields because they have no permanent magnets in their rotors, simple and strong structures, low manufacturing costs, and no risk of demagnetization at high temperatures. In order for the synchronous reluctance motor to achieve good vector control performance, it is necessary to obtain accurate rotor position information to achieve speed estimation and coordinate transformation, but the introduction of position sensors will undoubtedly increase system costs and reduce the reliability of the control system. Therefore, it is of great significance to study the position sensorless control method of synchronous reluctance motors.

[0003] According to different rotor position tracking strategies, the synchronous reluctance motor sensorless control algorithm can be divided into two categories: fundamental wave model and salient pole model. The principle of the fundamental wave model algorithm is to estimate the rotor position by analyzing and modeling the current and voltage signals. The principle of the salient pole model algorithm is based on the special rotor structure of the synchronous reluctance motor with strong salient polarity, and the rotor position is estimated by combining the model and observer. Both sensorless algorithms have limitations. The commonly used fundamental wave model algorithm essentially relies on back electromotive force solution, but the back electromotive force is very small when the motor runs at zero and low speed, and the signal-to-noise ratio of the available signal is very low, making it difficult to detect and separate. Therefore, the fundamental wave model algorithm has superior performance at medium and high speeds, but it is difficult to play a role at zero and low speeds. The salient pole model algorithm is just the opposite. Combining the two sensorless algorithms and building a composite observer is the key to completing the full-speed domain position sensorless control solution.

[0004] Due to the special rotor structure of the synchronous reluctance motor, the d-axis inductance and the q-axis inductance are strongly nonlinear. Online parameter identification of the inductance parameters is a prerequisite for achieving high-performance vector control. The real parameters of the motor can be obtained in real time without affecting the normal operation of the motor, thereby building a motor control model in real time and improving the accuracy of the control strategy for rotor position estimation during motor operation. Summary of the invention

[0005] In order to overcome the defects in the above-mentioned prior art, the present invention provides a full-speed range position sensorless control method for a synchronous reluctance motor based on parameter identification, which solves the defects that the traditional full-speed range position sensorless control system is difficult to meet the influence of motor parameter changes on the rotor position identification accuracy, and cannot smoothly switch between zero-speed and medium-speed control strategies.

[0006] To achieve the above object, the present invention adopts the following technical solutions, including:

[0007] The full-speed range position sensorless control method of synchronous reluctance motor based on parameter identification includes the following contents:

[0008] Zero-speed sensorless control: The pulse square wave voltage signal injection method without filter signal separation is adopted to inject a high-frequency square wave voltage signal into the direct axis of the estimated rotating coordinate system. The rotor position angle error is obtained by detecting the stator current and separating the high-frequency component of the current by algebraic operation without filter, thereby obtaining the estimated rotor position angle.

[0009] Sensorless control in the medium and high speed domain: The model reference adaptive control method is used, with the motor body as the reference model and the motor mathematical model containing the motor speed as the adjustable model. The adaptive law is designed to make the errors of the two models approach zero, thereby obtaining the estimated rotor position angle.

[0010] Online identification of inductance parameters: Using the model reference adaptive control method, the motor body is taken as the reference model, and the motor mathematical model including the motor speed is taken as the adjustable model. The adaptive law is designed to make the errors of the two models approach zero, so as to obtain the estimated inductance. The estimated inductance is used for position sensorless control in the zero and low speed domain and position sensorless control in the medium and high speed domain.

[0011] Preferably, the position sensorless control in the zero low speed range is specifically as follows:

[0012] Establish the position relationship between the coordinate systems: dq is the actual rotating coordinate system of the actual position of the rotor, θ e is the actual rotor position angle; is the estimated rotating coordinate system of the estimated rotor position, To estimate the rotor position angle; the rotor position angle error is Δθ e ,

[0013] Estimated Rotation Coordinate System The high-frequency square wave voltage signal injected into the shaft is:

[0014]

[0015] In the formula, They represent the estimated stator in the rotating coordinate system. High frequency component of shaft voltage, U in is the amplitude of the injected high-frequency square wave voltage signal, k is the discrete sampling time;

[0016] The high-frequency component of the current separated from the stator current is:

[0017]

[0018] In the formula, They represent the high-frequency components of the stator α and β axis currents in the stationary coordinate system, respectively. d , Lq are the d-axis and q-axis inductances in the actual rotating coordinate system; θ e is the actual rotor position angle, Δθ e is the rotor position angle error;

[0019] Extract the envelope of the high-frequency component of the current:

[0020]

[0021] In the formula, I αh ,I βh They represent the envelopes of the high-frequency components of the stator α and β axis currents in the stationary coordinate system, Δi αh , Δi βh They represent the changes of the high-frequency components of the stator α-axis and β-axis currents in the stationary coordinate system, respectively, and sign(·) is the sign function; Represents the estimated stator in the rotating coordinate system High frequency component of shaft voltage;

[0022] The rotor position angle error Δθ is obtained by vector cross multiplication e :

[0023]

[0024] Get the rotor position error signal Δθ e Then, the estimated rotor position angle is obtained using the phase-locked loop PLL. The phase-locked loop (PLL) compares the input signal (actual rotor position angle) with the output feedback signal (estimated rotor position angle) and generates a rotor position error signal Δθ e Perform PI control to estimate the motor electrical angular velocity, and then integrate the electrical angular velocity to obtain the estimated rotor position angle

[0025] Preferably, the high frequency component and the base frequency component are separated by a filter-free algebraic operation:

[0026]

[0027] In the formula, i dqf Represents the fundamental frequency component of the current, i dqh Represents the high-frequency component of current, i dq represents the sampling current of the stator, n and n-1 represent the sampling moments of the nth and n-1th discrete currents respectively;

[0028] The voltage equation of the synchronous reluctance motor in the actual rotating coordinate system is:

[0029]

[0030] In the formula, u d、u q They represent the stator d and q axis voltage components in the actual rotating coordinate system, respectively. d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system, respectively. d , L q are the d-axis and q-axis inductances in the actual rotating coordinate system, R S is the stator resistance, ω e is the motor electrical angular velocity;

[0031] When the motor speed is in the zero low speed domain, the voltage model under the excitation of the high-frequency square wave voltage signal is expressed as:

[0032]

[0033] In the formula, They represent the estimated stator in the rotating coordinate system. High frequency component of shaft current; They represent the estimated stator in the rotating coordinate system. High frequency component of shaft voltage;

[0034] The high-frequency component of the current in the stationary coordinate system is:

[0035]

[0036] In the formula, They represent the high-frequency components of the stator α and β axis currents in the stationary coordinate system, u dh 、u qh They represent the high-frequency components of the stator d-axis and q-axis voltages in the actual rotating coordinate system respectively;

[0037] The relationship between the actual rotation coordinate system and the estimated rotation coordinate system is:

[0038]

[0039] We can get:

[0040]

[0041] Preferably, the position sensorless control in the medium and high speed range is specifically as follows:

[0042] The adjustable models are:

[0043]

[0044] In the formula, They represent the estimated current components of the stator d and q axes in the actual rotating coordinate system in the adjustable model; u d 、u qThey represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system respectively; Represents the estimated motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; L d , L q represents the d-axis and q-axis inductances in the actual rotating coordinate system in the adjustable model; R S is the stator resistance;

[0045] The state error equation after the difference between the reference model and the adjustable model is expressed as:

[0046]

[0047] In the formula, C is the linear compensator matrix; ω e is the motor electrical angular velocity;

[0048] According to the state error equation, a nonlinear time-varying feedback system is obtained, including a feedforward linear model and a nonlinear feedback model;

[0049] The transfer function matrix of the feedforward linear model satisfies strictly positive real, that is:

[0050]

[0051] Where G(s) represents the transfer function of the feedforward linear model, s represents the complex frequency variable in Laplace transform, and I represents the identity matrix;

[0052] The nonlinear feedback model satisfies the Popov integral inequality, namely:

[0053]

[0054] In the formula, is a finite positive number; η(0,t) represents Popov integral inequality, t represents the time variable, and the superscript T represents the transpose;

[0055] The adaptive law is designed in the form of proportional integral, and the motor electrical angular velocity is estimated as:

[0056]

[0057] In the formula, To estimate the initial value of the motor electrical angular velocity; F 1 (τ), F 2 (τ) is the function to be solved, and τ is the independent variable;

[0058]

[0059] In the formula, k i, k p are all numbers greater than zero, representing the proportional gain and integral gain of the PI regulator respectively; the estimated motor electrical angular velocity is obtained and estimated rotor position angle

[0060]

[0061] Preferably, the online identification of the inductance parameters is specifically as follows:

[0062] The adjustable models are:

[0063]

[0064] In the formula, They represent the estimated current components of the stator d and q axes in the actual rotating coordinate system in the adjustable model; u d 、u q They represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system respectively; Represents the estimated motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; They represent the estimated inductance of the d and q axes in the actual rotating coordinate system in the adjustable model; R S is the stator resistance; ω e is the motor electrical angular velocity;

[0065] The state error equation after the difference between the reference model and the adjustable model is expressed as:

[0066]

[0067] In the formula, W′=(Bi * +Du), C is the linear compensator matrix;

[0068] According to the state error equation, a nonlinear time-varying feedback system is obtained, including a feedforward linear model and a nonlinear feedback model;

[0069] The transfer function matrix of the feedforward linear model satisfies strictly positive real, that is:

[0070] G(s)=C(sI-A) -1

[0071] Where G(s) represents the transfer function of the feedforward linear model, s represents the complex frequency variable in Laplace transform, and I represents the identity matrix;

[0072] The nonlinear feedback model satisfies the Popov integral inequality, namely:

[0073]

[0074] In the formula, is a finite positive number; η(0,t) represents Popov integral inequality, t represents the time variable, and the superscript T represents the transpose;

[0075] The adaptive law is designed in the form of proportional integral, and the estimated inductance is:

[0076]

[0077] In the formula, They represent the estimated initial values ​​of the inductance of the d and q axes in the actual rotating coordinate system in the adjustable model; F 1 (τ), F 2 (τ), F 3 (τ), F 4 (τ) is the function to be solved, and τ is the independent variable;

[0078]

[0079] In the formula, k i ′, k′ p , k i ″、k′ p ' are all numbers greater than zero;

[0080] The estimated inductance is:

[0081]

[0082] Preferably, the zero low speed range means that the motor speed is less than or equal to the set speed ω e1 The medium and high speed range refers to the motor speed being greater than the set speed ω e2 .

[0083] Preferably, the method further includes: a weighted switching strategy between the zero-speed domain and the medium-speed domain, as shown below:

[0084] The estimated rotor position angle obtained in the zero low speed domain is The estimated rotor position angle in the medium and high speed domain is Set the switching lower limit to ω e1 , the upper limit of switching is ω e2 , the switching weight coefficient is f(ω e ),

[0085]

[0086] In the formula, ω e is the motor electrical angular velocity;

[0087] In [ω e1,ω e2 ] is estimated within the range of:

[0088]

[0089] The present invention also provides a readable storage medium having a computer program stored thereon, and when the computer program is executed, the above-mentioned parameter identification-based full-speed range position sensorless control method for a synchronous reluctance motor is implemented.

[0090] The present invention also provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and run on the processor. When the processor executes the computer program, it implements the above-mentioned parameter identification-based full-speed range position sensorless control method for a synchronous reluctance motor.

[0091] The present invention also provides a computer program product, which includes a computer program / instruction, which, when executed by a processor, implements the above-mentioned full-speed range position sensorless control method of a synchronous reluctance motor based on parameter identification.

[0092] The advantages of the present invention are:

[0093] (1) The present invention provides a full-speed range position sensorless control method for a synchronous reluctance motor based on parameter identification, which solves the problem that a traditional full-speed range position sensorless control system is difficult to meet the impact of motor parameter changes on the accuracy of rotor position recognition, and cannot smoothly switch between zero-speed and medium-speed control strategies.

[0094] (2) In the zero-speed and low-speed sections, a pulse square wave voltage signal injection control strategy without filter signal separation is adopted. The processing flow is simple, the estimation accuracy is high, and the signal is separated by algebraic operations, eliminating the phase error caused by the use of filters.

[0095] (3) The model reference adaptive control method is used in the medium and high speed sections. It is simple to calculate, easy to implement, and has high steady-state operation accuracy.

[0096] (4) Considering the nonlinearity of the inductance parameters of the synchronous reluctance motor affected by magnetic circuit saturation and cross saturation, the same method, namely the model reference adaptive control method, is used to identify the parameters online, which reduces the complexity of the system, simplifies the control and has high identification accuracy. After identification, the motor model and control strategy model are constructed to improve the accuracy of rotor position estimation during the entire operation process.

[0097] (5) In the transition section, an improved switching control strategy with a continuous and smooth switching weight coefficient is proposed to improve the switching smoothness of the position sensorless control of the motor in the full speed range. BRIEF DESCRIPTION OF THE DRAWINGS

[0098] Figure 1 The present invention is a flow chart of the method.

[0099] Figure 2 is the relationship between the injected high-frequency square wave voltage signal and the high-frequency component in the current response.

[0100] Figure 3 This is a block diagram of the principle of separating the fundamental frequency component and the high-frequency component in the current response.

[0101] Figure 4 This is a schematic diagram of the coordinate position relationship of the pulse square wave voltage signal injection method.

[0102] Figure 5 Block diagram of the principle of estimating the rotor position angle for the zero low-speed domain.

[0103] Figure 6 It is a nonlinear time-varying feedback system that is equivalent to the model reference adaptive control method.

[0104] Figure 7 The schematic diagram of the weighted switching strategy in the zero low-speed domain and the medium-high-speed domain.

[0105] Figure 8 This is the principle block diagram of the full-speed range position sensorless control method of the synchronous reluctance motor based on parameter identification. DETAILED DESCRIPTION

[0106] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0107] Example 1

[0108] Depend on Figure 1 As shown in the figure, the full-speed range position sensorless control method of the synchronous reluctance motor based on parameter identification includes the following contents:

[0109] Zero low-speed domain position sensorless control: adopt the pulse square wave voltage signal injection method without filter signal separation, inject high-frequency square wave voltage signal into the direct axis of the estimated rotating coordinate system, detect the stator current and separate the high-frequency component of the current by algebraic operation without filter, obtain the rotor position angle error, and then estimate the rotor position angle through the phase-locked loop observer to obtain the estimated rotor position angle. Among them, the zero low-speed domain refers to the motor speed is below 10% of the rated speed, that is, the motor speed is less than or equal to the set speed ω e1 It is the zero low speed range.

[0110] Sensorless control in the medium and high speed domain: The model reference adaptive control method is used, with the motor body as the reference model and the motor mathematical model containing the motor speed as the adjustable model. The adaptive law is designed to make the errors of the two models approach zero, thereby obtaining the estimated rotor position angle. Among them, the medium and high speed domain refers to the motor speed greater than the set speed ω e2 .

[0111] Inductance identification: The model reference adaptive control method is adopted, with the motor body as the reference model and the motor mathematical model including the motor speed as the adjustable model. The adaptive law is designed to make the errors of the two models approach zero, so as to obtain the estimated inductance. The estimated inductance is used for position sensorless control in the zero and low speed domain and position sensorless control in the medium and high speed domain.

[0112] Weighted switching strategy in the zero low speed domain and the medium and high speed domain, the motor electrical angular velocity ω e In the switching range [ω e1 ,ω e2 ], the estimated rotor position angle is obtained by weighted method.

[0113] Specifically, the zero-speed sensorless control is as follows:

[0114] When the motor operates in the zero low speed domain, the position sensorless control method uses the strong salient polarity of the motor rotor structure for control, and the rotor position is estimated by the high-frequency signal injection method. Common high-frequency signal injection methods are divided into rotating high-frequency injection method and pulse high-frequency injection method. In order to achieve a higher injection frequency to weaken the noise, the pulse high-frequency injection method, which is widely used in engineering, is adopted. Among them, the pulse square wave voltage signal injection method is a type of pulse high-frequency injection method. It injects a high-frequency square wave voltage signal into the direct axis of the estimated rotating coordinate system, detects the stator three-phase current signal and separates the high-frequency component to obtain the rotor position angle error, and then decouples the rotor position angle error to achieve the estimation of the rotor position angle. The advantage of this method is that a high-frequency square wave voltage signal is injected into the direct axis of the estimated rotating coordinate system. Since the torque component of the high-frequency square wave voltage signal is very small and can be ignored, the motor torque pulsation caused by the injection of the high-frequency square wave voltage signal can be avoided.

[0115] In order to decouple the rotor position angle from the response signal of the pulse square wave signal injection method, it is necessary to first establish the position relationship between the coordinate systems, such as Figure 4 As shown, dq is the actual rotating coordinate system of the actual position of the rotor, θ e is the actual rotor position angle; is the estimated rotating coordinate system of the estimated rotor position, To estimate the rotor position angle; the rotor position angle error is Δθ e ,

[0116] To estimate the direct axis of the rotation coordinate system, The high-frequency square wave voltage signal injected into the shaft is:

[0117]

[0118] In the formula, They represent the estimated stator in the rotating coordinate system. High frequency component of shaft voltage, U in is the amplitude of the injected high-frequency square wave voltage signal, k is the discrete sampling time;

[0119] In a rotating coordinate system, a high-frequency square wave voltage signal is injected into the motor control system. After transformation, the high-frequency component in the sampled current needs to be separated. The high-frequency component and the fundamental frequency component in the current response are separated by a filter-free algebraic operation, which can reduce the delay and phase lag of the observer caused by the filter.

[0120] The frequency of the injected high-frequency square wave signal is set to half of the inverter switching frequency to ensure the spectrum separation space. The sampling frequency of the current signal is adjusted to twice the switching frequency, so that the fundamental frequency is much smaller than the current sampling frequency. It is assumed that the current fundamental frequency component remains basically unchanged in two adjacent sampling cycles. The relationship between the injected high-frequency square wave voltage signal and the high-frequency component in the response current is as follows: Figure 2 The current in two adjacent sampling periods can be expressed by superposition of high-frequency components and fundamental frequency components, that is:

[0121]

[0122] In the formula, i dq Represents the sampling current of the stator, i dqf Represents the fundamental frequency component of the sampling current, i dqh represents the high-frequency component of the sampled current, and n and n-1 represent the sampling moments of the nth and n-1th discrete currents, respectively.

[0123] The principle diagram of separation of high frequency component and fundamental frequency component is as follows: Figure 3 As shown, Figure 3 Middle Z -1 Indicates a delay of one sampling period; According to the above algebraic operation, the high-frequency component and the fundamental frequency component can be obtained:

[0124]

[0125] The voltage equation of the synchronous reluctance motor in the actual rotating coordinate system is:

[0126]

[0127] In the formula, u d 、u qThey represent the stator d and q axis voltage components in the actual rotating coordinate system, respectively. d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system, respectively. d , L q are the d-axis and q-axis inductances in the actual rotating coordinate system, R S is the stator resistance, ω e is the motor electrical angular velocity.

[0128] When the motor speed is low, the frequency of the injected high-frequency square wave voltage signal is much greater than the fundamental operating frequency, and the rotor voltage drop can be ignored when the speed is low and the injected voltage amplitude is large, so the voltage model under the excitation of the high-frequency square wave voltage signal can be expressed as:

[0129]

[0130] In the formula, They represent the estimated stator in the rotating coordinate system. High frequency component of shaft current; They represent the estimated stator in the rotating coordinate system. High frequency component of shaft voltage.

[0131] Injecting a high-frequency square wave voltage signal into the estimated rotating coordinate system, the high-frequency component of the current in the stationary coordinate system (α-β coordinate axis) is obtained as follows:

[0132]

[0133] In the formula, They represent the high-frequency components of the stator α and β axis currents in the stationary coordinate system, u dh 、u qh They respectively represent the high-frequency components of the stator d-axis and q-axis voltages in the actual rotating coordinate system.

[0134] Depend on Figure 4 It can be seen that the relationship between the actual rotation coordinate system and the estimated rotation coordinate system is:

[0135]

[0136] We can get:

[0137]

[0138] When the rotor position angle error Δθ between the estimated rotor position angle and the actual rotor position angle e =0, the above formula is:

[0139]

[0140] From the above formula, we can know that the inverse tangent function of the high-frequency component of the current can be directly used to obtain the rotor position angle information, but the direct method is easily affected by noise signals. The envelope of the high-frequency component of the current contains the position signal. Extracting the envelope and demodulating the rotor position angle information can avoid the influence of high-frequency noise.

[0141] The envelope of the high-frequency component of the current can be obtained by multiplying the change in the high-frequency component of the stator α and β axis current in the stationary coordinate system with the output of the high-frequency square wave voltage signal after the sign function is applied, that is:

[0142]

[0143] In the formula, I αh ,I βh They represent the envelopes of the high-frequency components of the stator α and β axis currents in the stationary coordinate system, Δi αh , Δi βh They represent the changes of the high-frequency components of the stator α and β axis currents in the stationary coordinate system, respectively. sign(x) is the sign function, that is, Represents the estimated stator in the rotating coordinate system High frequency component of shaft voltage.

[0144] After extracting the envelope of the high-frequency component of the current, the rotor position angle error can be obtained by vector cross multiplication, that is:

[0145]

[0146] Get the rotor position angle error Δθ e Then, a phase-locked loop (PLL) is used to obtain the estimated rotor position angle. The phase-locked loop (PLL) compares the input signal (actual rotor position angle) with the output feedback signal (estimated rotor position angle) and generates a rotor position error signal Δθ e Perform PI control to estimate the motor electrical angular velocity, and then integrate the electrical angular velocity to obtain the estimated rotor position angle The above principle is as follows Figure 5 shown.

[0147] Specifically, the mid- and high-speed sensorless control is as follows:

[0148] When the motor operates in the medium and high speed domain, the sensorless control method estimates the motor's electrical angular velocity and rotor position angle by studying the fundamental wave model of the motor and the interaction between the rotor angle and the back electromotive force. Common methods include model reference adaptive control and sliding mode observer. In the present invention, the model reference adaptive control method is selected, which has the advantages of simple structure, easy engineering implementation, strong robustness, etc.

[0149] The model reference adaptive control method mainly includes three parts: reference model, adjustable model and adaptive law. The combination forms include series type, parallel type and series-parallel type structure. The present invention adopts parallel type structure. The adjustable model and the reference model obtain their own output variables under the same input variables. The error obtained by subtracting the output variables is input into the adaptive law to adjust the adjustable model so that the errors of the two models are close to zero, thereby observing the state variables of the motor.

[0150] The present invention uses the motor body as a reference model and uses a mathematical model of a synchronous reluctance motor including the motor speed as an adjustable model.

[0151] The reference model is:

[0152]

[0153] In the formula, i d 、i q They represent the actual current components of the stator d and q axes in the actual rotating coordinate system in the reference model; u d 、u q Respectively represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system; ω e Indicates the actual motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; They represent the estimated inductance of the d and q axes in the actual rotating coordinate system of the reference model; R S is the stator resistance.

[0154] The adjustable models are:

[0155]

[0156] In the formula, They represent the estimated current components of the stator d and q axes in the actual rotating coordinate system in the adjustable model; u d 、u q They represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system respectively; Represents the estimated motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; They represent the estimated inductance of the d and q axes in the actual rotating coordinate system in the adjustable model; R S is the stator resistance;.

[0157] The difference between the reference model and the adjustable model is expressed as the state error equation:

[0158]

[0159] In the formula, C is the linear compensator matrix, usually set to the unit matrix; ω e is the motor electrical angular velocity.

[0160] The above state error equation transforms the model reference adaptive control system into an equivalent nonlinear time-varying feedback system, which includes a feedforward linear model and a nonlinear feedback model. The principle block diagram is shown in Figure 6 As shown, the adaptive law output is the estimated motor electrical angular velocity.

[0161] The key point of the model reference adaptive control system is to obtain an adaptive law that makes the error approach zero quickly. There are usually three design methods: ① based on Lyapunov stability theory; ② based on local parameter optimization theory; ③ based on Popov superstability theory. Comparing the three methods, the present invention selects the Popov superstability theory with higher accuracy and lower design difficulty, so Figure 6 The nonlinear time-varying feedback system is asymptotically stable if two conditions are met:

[0162] (1) The transfer function matrix of the feedforward linear model satisfies the strict positive real condition, that is:

[0163]

[0164] Where G(s) represents the transfer function of the feedforward linear model, s represents the complex frequency variable in Laplace transform, and I represents the identity matrix;

[0165] According to matrix correlation theory, the above transfer function matrix satisfies strictly positive real.

[0166] (2) The nonlinear feedback model satisfies the Popov integral inequality:

[0167]

[0168] In the formula, is a finite positive number; η(0,t) represents Popov's integral inequality, t represents the time variable, and the superscript T Indicates transpose.

[0169] Substitute the above y and W into the inequality:

[0170]

[0171] The model reference adaptive law is generally designed in the form of proportional integral, and the motor electrical angular velocity is estimated as:

[0172]

[0173] In the formula, To estimate the initial value of the motor electrical angular velocity; F 1 (τ), F 2 (τ) represents the function to be solved, and τ represents the independent variable.

[0174] In order to satisfy Popov's integral inequality, let:

[0175]

[0176] In the formula, k i , k p Both are numbers greater than zero, representing the proportional gain and integral gain of the PI regulator respectively.

[0177] From this, we can estimate the motor electrical angular velocity and estimated rotor position angle

[0178]

[0179] Specifically, the online identification of inductor parameters is as follows:

[0180] For the online identification of inductance parameters, there are methods such as model reference adaptation, recursive least squares method, Kalman filter method and intelligent algorithm. Among them, model reference adaptation has high accuracy and simple control, and since the same method used at the above-mentioned medium and high speeds is used to identify the rotor position, the complexity of the system can be reduced. Therefore, the present invention adopts the model reference adaptive control method to identify the inductance parameters online, and improves the accuracy of the above-mentioned rotor position angle estimation control strategy.

[0181] Since the current and inductance of the motor change much faster than the speed, the motor electrical angular velocity can be treated as a short-term constant when performing inductance parameter identification. The parameter identification still uses the motor body as the reference model and the synchronous reluctance motor mathematical model containing the motor speed as the adjustable model.

[0182] The adjustable models are:

[0183]

[0184] In the formula, They represent the estimated current components of the stator d and q axes in the actual rotating coordinate system in the adjustable model; u d 、u qThey represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system respectively; Represents the estimated motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; They represent the estimated inductance of the d and q axes in the actual rotating coordinate system in the adjustable model; R S is the stator resistance; ω e is the motor electrical angular velocity.

[0185] The difference between the reference model and the adjustable model is expressed as the state error equation:

[0186]

[0187] In the formula, W′=(Bi * +Du), C is the linear compensator matrix, which is usually set to the identity matrix.

[0188] The above state error equation also transforms the model reference adaptive control system into an equivalent nonlinear time-varying feedback system, including a feedforward linear model and a nonlinear feedback model. The principle block diagram is also as follows: Figure 6 as shown, but the adaptive law output is an estimated inductance.

[0189] The online identification of inductor parameters also uses Popov's superstability theory, so the above two conditions still need to be met:

[0190] (1) The transfer function matrix of the feedforward linear model G(s) = C(sI-A) -1 Satisfies strict positive reality.

[0191] (2) The nonlinear feedback model satisfies the Popov integral inequality:

[0192]

[0193] Substitute the above y and W' into the inequality:

[0194]

[0195] The model reference adaptive law is generally designed in the form of proportional integral, and the estimated inductance is:

[0196]

[0197] In the formula, They represent the estimated initial values ​​of the inductance of the d and q axes in the actual rotating coordinate system in the adjustable model; F 1 (τ), F 2 (τ), F3 (τ), F 4 (τ) represents the function to be solved, and τ represents the independent variable.

[0198] In order to satisfy Popov's integral inequality, let:

[0199]

[0200] In the formula, k i ′, k′ p , k i ″、k′ p ' are all numbers greater than zero.

[0201] From this, the estimated inductance is:

[0202]

[0203] Specifically, the weighted switching strategy for the zero low speed domain and the medium and high speed domain is as follows:

[0204] A switching weight coefficient f(ω e ) Continuously smooth weighted switching strategy eliminates sudden changes during switching, reduces oscillation, and improves the switching smoothness of full-speed control.

[0205] How to implement different speed control strategies in the speed transition range [ω e1 ,ω e2 ] is the key to full-speed sensorless control. The current methods mainly include hysteresis switching method and linear average weighted switching method. In order to improve the switching smoothness, the present invention proposes an improved weighted switching strategy, which adopts a continuous smooth cubic function weighted switching method to replace the traditional linear switching function, and fuses the estimated rotor position angles obtained by the two sensorless control methods to realize full-speed sensorless hybrid vector control.

[0206] The continuous smooth cubic function weight switching method proposed by the present invention: Assume that there is a high-order function curve f(ω e ), the switching weight coefficient f(ω e ) is continuous and smooth, let:

[0207] f(ω e )=a n ω e n +a n-1 ω e n-1 +...+a 2 ω e 2 +a 1 ω e 1+a 0

[0208] In the formula, a n (n=0,1,2,...) are the coefficients of each power, ω e Indicates the motor electrical angular velocity.

[0209] In order to make the motor in the speed transition range [ω e1 ,ω e2 ] does not have a sudden change phenomenon, and the switching weight coefficient f(ω e ) needs to meet the following conditions: Switching weight coefficient f(ω e ) is continuously differentiable in the speed range, and the switching weight coefficient f(ω e ) at the switching lower limit ω e1 The value at is 1 and the derivative is 0. In the speed transition range [ω e1 ,ω e2 ]The value at the midpoint is 0.5, and the switching upper limit ω e2 The value at is 0 and the derivative is 0. Then f(ω e ) should meet the following requirements:

[0210]

[0211] Theoretically, we can solve the equations up to the fourth power to get the corresponding coefficients. However, since the equations are difficult to solve, we can first let ω e1 =0,ω e2 =2,ω e1 The value f(ω e1 ) is 2, the value at the midpoint of the switching speed range =1,ω e2 The value f(ω e2 ) is 0, and then the coordinate transformation and scaling of the speed range switching function is performed, the equation f(ω e )for:

[0212] f(ω e )=0.5ω e 3 -1.5ω e 2 +2

[0213] Verify that it is continuously differentiable in the interval [0,2], so it can be known that a continuous smooth high-order function can be a cubic function. Then use the coordinate transformation to scale f(ω e )The domain is scaled to [ω e1 ,ω e2 ], the range is scaled to [0,1] to satisfy the above conditions, and the continuous smooth cubic function is solved as follows:

[0214]

[0215] Therefore, the switching weight coefficient f(ω) of the present invention can be expressed as:

[0216]

[0217] In the speed transition range [ω e1 ,ω e2 ] is estimated within the range of:

[0218]

[0219] The continuous smooth cubic function proportional switching method can be applied in any speed transition range as long as the upper and lower limits of the switching are known. Since the motor's zero low speed stage is less than 10% of the rated speed, and the medium and high speed stage is more than 10% of the rated speed, the speed switching range is selected as 300r / min-400r / min, and the switching function is obtained, that is, Figure 7 As shown, Figure 7 This is the principle block diagram of the improved weighted switching strategy. The blue line is f(ω) and the red line is 1-f(ω).

[0220] When the motor is in the zero-speed domain, the pulse square wave voltage signal injection method is used to achieve effective observation of the rotor position and good dynamic response performance; when entering the speed transition interval, an improved weighted switching strategy is adopted to enable the motor to smoothly switch between two control strategies in this interval; when the motor enters the medium and high-speed domain, the model reference adaptive control method is used to observe the rotor position angle, which has high steady-state operation accuracy; but during the entire operation process, due to the nonlinearity of the inductance parameters affected by magnetic circuit saturation and cross saturation, the model reference adaptive method is used for online identification to improve the accuracy of rotor position angle estimation.

[0221] The present invention proposes a full-speed range position sensorless control method based on parameter identification, and the control principle block diagram is as follows: Figure 8 As shown in the figure, the zero-speed domain control strategy is first considered, and the pulse square wave voltage signal injection method is used to estimate the rotor position. The filter-free method of algebraic operation is used to separate the signals during operation; then the medium- and high-speed domain control strategy is considered, and the model reference adaptive control method is used to estimate the rotor position; when switching between the two control strategies, an improved weighted switching function is used to improve the switching smoothness; and the inductance parameters are identified through the model reference adaptive method to improve the system accuracy and robustness.

[0222] Example 2

[0223] An electronic device comprises a processor, a memory and a computer program stored in the memory and executable on the processor, wherein the method of the above-mentioned embodiment 1 is implemented when the processor executes the computer program.

[0224] The electronic device of the embodiment of the present application can be the mobile device itself, or a stand-alone device independent of the mobile device, which can communicate with the mobile device to receive the collected input signals from them and send the selected target decision behavior to them.

[0225] The electronic device includes one or more processors and memory. The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. The memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor may run the program instructions to implement the decision behavior decision method and / or other desired functions of the various embodiments of the present application described above.

[0226] Electronic devices may also include input devices and output devices.

[0227] Example 3

[0228] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the decision-making behavior decision method according to various embodiments of the present application described in the above-mentioned embodiment 1 of this specification.

[0229] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0230] Example 4

[0231] An embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes the steps of the decision-making behavior decision method according to various embodiments of the present application described in the above-mentioned embodiment 1.

[0232] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0233] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A full-speed range position sensorless control method for a synchronous reluctance motor based on parameter identification, characterized in that: Includes the following: Zero-speed sensorless control: The pulse square wave voltage signal injection method without filter signal separation is adopted to inject a high-frequency square wave voltage signal into the direct axis of the estimated rotating coordinate system. The rotor position angle error is obtained by detecting the stator current and separating the high-frequency component of the current by algebraic operation without filter, thereby obtaining the estimated rotor position angle. Sensorless control in the medium and high speed domain: The model reference adaptive control method is used, with the motor body as the reference model and the motor mathematical model containing the motor speed as the adjustable model. The adaptive law is designed to make the errors of the two models approach zero, thereby obtaining the estimated rotor position angle. Online identification of inductance parameters: Using the model reference adaptive control method, the motor body is taken as the reference model, and the motor mathematical model including the motor speed is taken as the adjustable model. The adaptive law is designed to make the errors of the two models approach zero, so as to obtain the estimated inductance. The estimated inductance is used for position sensorless control in the zero and low speed domain and position sensorless control in the medium and high speed domain.

2. The method for controlling a synchronous reluctance motor in full speed range without position sensor based on parameter identification according to claim 1, characterized in that: The zero-speed sensorless control is specifically as follows: Establish the position relationship between the coordinate systems: dq is the actual rotating coordinate system of the actual position of the rotor, θ e is the actual rotor position angle; is the estimated rotating coordinate system of the estimated rotor position, To estimate the rotor position angle; The rotor position angle error is Δθ e , Estimated Rotation Coordinate System The high-frequency square wave voltage signal injected into the shaft is: In the formula, They represent the estimated stator in the rotating coordinate system. High frequency component of shaft voltage, U in is the amplitude of the injected high-frequency square wave voltage signal, k is the discrete sampling time; The high-frequency component of the current separated from the stator current is: In the formula, They represent the high-frequency components of the stator α and β axis currents in the stationary coordinate system, respectively. d , L q are the d-axis and q-axis inductances in the actual rotating coordinate system; θ e is the actual rotor position angle, Δθ e is the rotor position angle error; Extract the envelope of the high-frequency component of the current: In the formula, I αh ,I βh They represent the envelopes of the high-frequency components of the stator α and β axis currents in the stationary coordinate system, Δi αh , Δi βh They represent the changes of the high-frequency components of the stator α-axis and β-axis currents in the stationary coordinate system, respectively, and sign(·) is the sign function; Represents the estimated stator in the rotating coordinate system High frequency component of shaft voltage; The rotor position angle error Δθ is obtained by vector cross multiplication e : Get the rotor position error signal Δθ e Then, the estimated rotor position angle is obtained using the phase-locked loop PLL. The phase-locked loop (PLL) compares the input signal (actual rotor position angle) with the output feedback signal (estimated rotor position angle) and generates a rotor position error signal Δθ e Perform PI control to estimate the motor electrical angular velocity, and then integrate the electrical angular velocity to obtain the estimated rotor position angle 3. The method for controlling a synchronous reluctance motor in full speed range without position sensor based on parameter identification according to claim 2, characterized in that: The high-frequency and fundamental frequency components are separated by algebraic operations without filters: In the formula, i dqf Represents the fundamental frequency component of current, i dqh Represents the high-frequency component of current, i dq represents the sampling current of the stator, n and n-1 represent the sampling moments of the nth and n-1th discrete currents respectively; The voltage equation of the synchronous reluctance motor in the actual rotating coordinate system is: In the formula, u d 、u q They represent the stator d and q axis voltage components in the actual rotating coordinate system, respectively. d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system, respectively. d , L q are the d-axis and q-axis inductances in the actual rotating coordinate system, R S is the stator resistance, ω e is the motor electrical angular velocity; When the motor speed is in the zero low speed domain, the voltage model under the excitation of the high-frequency square wave voltage signal is expressed as: In the formula, They represent the estimated stator in the rotating coordinate system. High frequency component of shaft current; They represent the estimated stator in the rotating coordinate system. High frequency component of shaft voltage; The high-frequency component of the current in the stationary coordinate system is: In the formula, They represent the high-frequency components of the stator α and β axis currents in the stationary coordinate system, u dh 、u qh They represent the high-frequency components of the stator d-axis and q-axis voltages in the actual rotating coordinate system respectively; The relationship between the actual rotation coordinate system and the estimated rotation coordinate system is: We can get:

4. The method for controlling a synchronous reluctance motor in full speed range without position sensor based on parameter identification according to claim 1, characterized in that: The position sensorless control in the medium and high speed range is specifically as follows: The adjustable models are: In the formula, They represent the estimated current components of the stator d and q axes in the actual rotating coordinate system in the adjustable model respectively; u d 、u q They represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system respectively; Represents the estimated motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; L d , L q represents the d-axis and q-axis inductances in the actual rotating coordinate system in the adjustable model; R S is the stator resistance; The state error equation after the reference model and the adjustable model are expressed as: In the formula, C is the linear compensator matrix; ω e is the motor electrical angular velocity; According to the state error equation, a nonlinear time-varying feedback system is obtained, including a feedforward linear model and a nonlinear feedback model; The transfer function matrix of the feedforward linear model satisfies strictly positive real, that is: Where G(s) represents the transfer function of the feedforward linear model, s represents the complex frequency variable in Laplace transform, and I represents the identity matrix; The nonlinear feedback model satisfies the Popov integral inequality, namely: In the formula, is a finite positive number; η(0,t) represents Popov integral inequality, t represents the time variable, and the superscript T represents the transpose; The adaptive law is designed in the form of proportional integral, and the motor electrical angular velocity is estimated as: In the formula, To estimate the initial value of the motor electrical angular velocity; F1(τ) and F2(τ) are the functions to be solved, and τ is the independent variable; In the formula, k i , k p Both are numbers greater than zero, representing the proportional gain and integral gain of the PI regulator respectively; Get the estimated motor electrical angular velocity and estimated rotor position angle 5. The method for controlling a synchronous reluctance motor in full speed range without position sensor based on parameter identification according to claim 1, characterized in that: The online identification of the inductance parameters is specifically as follows; The adjustable models are: In the formula, They represent the estimated current components of the stator d and q axes in the actual rotating coordinate system in the adjustable model respectively; u d 、u q They represent the stator d-axis and q-axis voltage components in the actual rotating coordinate system respectively; Represents the estimated motor electrical angular velocity; i d 、i q They represent the stator d-axis and q-axis current components in the actual rotating coordinate system respectively; They represent the estimated inductance of the d and q axes in the actual rotating coordinate system in the adjustable model; R S is the stator resistance; ω e is the motor electrical angular velocity; The state error equation after the difference between the reference model and the adjustable model is expressed as: In the formula, W′=(Bi * +Du), C is the linear compensator matrix; According to the state error equation, a nonlinear time-varying feedback system is obtained, including a feedforward linear model and a nonlinear feedback model; The transfer function matrix of the feedforward linear model satisfies strictly positive real, that is: G(s)=C(sI-A) -1 Where G(s) represents the transfer function of the feedforward linear model, s represents the complex frequency variable in Laplace transform, and I represents the identity matrix; The nonlinear feedback model satisfies the Popov integral inequality, namely: In the formula, is a finite positive number; η(0,t) represents Popov integral inequality, t represents the time variable, and the superscript T represents the transpose; The adaptive law is designed in the form of proportional integral, and the estimated inductance is: In the formula, They represent the estimated initial values ​​of the inductance of the d-axis and q-axis in the actual rotating coordinate system in the adjustable model respectively; F1(τ), F2(τ), F3(τ), and F4(τ) are the functions to be solved, and τ is the independent variable; In the formula, k i ′, k′ p , k i ″、k′ p ' are all numbers greater than zero; The estimated inductance is:

6. The method for controlling a synchronous reluctance motor in full speed range without position sensor based on parameter identification according to claim 1, characterized in that: The zero low speed region means that the motor speed is less than or equal to the set speed ω e1 The medium and high speed range refers to the motor speed being greater than the set speed ω e2 .

7. The method for controlling a synchronous reluctance motor in full speed range without position sensor based on parameter identification according to claim 1 or 6, characterized in that: It also includes: weighted switching strategies for the zero low speed domain and the medium and high speed domain, as shown below: The estimated rotor position angle obtained in the zero low speed domain is The estimated rotor position angle in the medium and high speed domain is Set the switching lower limit to ω e1 , the upper limit of switching is ω e2 , the switching weight coefficient is f(ω e ), In the formula, ω e is the motor electrical angular velocity; In [ω e1 ,ω e2 ] is estimated within the range of:

8. A readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed, the method for controlling a synchronous reluctance motor in full speed range without a position sensor based on parameter identification as claimed in any one of claims 1 to 7 is implemented.

9. An electronic device, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the full-speed range position sensorless control method of a synchronous reluctance motor based on parameter identification as described in any one of claims 1 to 7.

10. A computer program product, characterized in that It includes a computer program / instruction, which, when executed by a processor, implements the full-speed range position sensorless control method for a synchronous reluctance motor based on parameter identification as described in any one of claims 1 to 7.

Images