Sensorless control method for permanent magnet synchronous motor

Abstract

The present invention relates to a sensorless control method for a permanent magnet synchronous motor. There are two main methods for obtaining rotor information in the sensorless case of a permanent magnet synchronous motor. One method is to construct a state observer to obtain rotor position information from a back electromotive force signal. Since the back electromotive force signal is weak when the motor is running at a low speed, this type of method has poor performance in estimating the rotor position at a low speed. At zero speed, the permanent magnet synchronous motor does not generate a back electromotive force and cannot estimate the rotor position, but has excellent performance in the medium and high speed range. The present invention includes the following steps: establishing a permanent magnet synchronous motor model, constructing a rotor flux observer to observe the rotor position information, performing initial value estimation and compensation on the rotor flux observer, constructing a speed observer to convert the rotor position information into speed information, and constructing an ID controller to complete error compensation. The present invention is used for a sensorless control method for a permanent magnet synchronous motor.

Description

Technical field:

[0001] The present invention relates to the field of motor control, and in particular to a sensorless control method for a permanent magnet synchronous motor. Background technology:

[0002] Permanent magnet synchronous motor is a synchronous motor that generates a synchronous rotating magnetic field by permanent magnet excitation. Compared with asynchronous motors, permanent magnet synchronous motors have high power density, small moment of inertia, high energy conversion efficiency, and low maintenance costs. They are widely used in new energy vehicles, aerospace, CNC machine tools and other fields. Traditional PMSM control systems usually use mechanical position sensors to obtain rotor position information and instantaneous speed. Such as photoelectric encoders, rotary transformers and Hall sensors, etc. However, in complex environments, mechanical sensors may have high environmental requirements, slow response, weak anti-electromagnetic interference ability, and aging. Therefore, the research on position sensorless drive control is of great significance and value to improving the reliability of the system.

[0003] There are two main methods for obtaining rotor information in the sensorless permanent magnet synchronous motor. One is to construct a state observer to obtain rotor position information from the back electromotive force signal. Since the back electromotive force signal is weak when the motor is running at low speed, this method has poor performance in estimating rotor position at low speed. At zero speed, the permanent magnet synchronous motor does not generate back electromotive force and cannot estimate the rotor position, but it has excellent performance in the medium and high speed range. The most common one is to construct a sliding mode observer. The sliding mode observer has excellent performance in estimating rotor position and speed, good system robustness, and easy digital programming of the algorithm, but the essential characteristics of sliding mode motion will cause the system to vibrate. The second method is based on the idea of ​​salient pole tracking, injecting high-frequency signals into the stator of the motor to obtain rotor position information from the response current. This method can solve the problem of motor parameter identification at low speed or zero speed. Since this method requires the motor to have a certain salient pole effect, it can be directly implemented on embedded and plug-in synchronous motors, but it is more difficult to implement for motors such as surface-mounted synchronous motors where the salient pole effect is not obvious. It needs to be implemented by modifying the internal structure of the motor or through saturation effect, and this method will cause greater noise and additional losses. Summary of the invention:

[0004] The object of the present invention is to provide a sensorless control method for a permanent magnet synchronous motor, by which the operation problem of a surface mounted permanent magnet synchronous motor at low speed and zero speed is solved.

[0005] The above purpose is achieved through the following technical solutions:

[0006] A sensorless control method for a permanent magnet synchronous motor, the algorithm comprises the following steps:

[0007] Step 1, establishing a mathematical model of a surface-mounted permanent magnet synchronous motor, that is, a motor model of the permanent magnet synchronous motor based on a stationary coordinate system (α-β coordinate system);

[0008] (1)

[0009] (2)

[0010] Where ψ=[ψ α , ψ β ] T is the stator flux in the stationary coordinate system, ψ m is the permanent magnet flux constant, L S is the stator winding inductance, i ab =[i α , i β ]T and is the stator current in the stationary coordinate system, u ab =[u α ,u β ] T is the stator voltage in the stationary coordinate system, R is the stator coil winding resistance, and θ is the rotor position;

[0011] Step 2: In order to obtain the rotor position information, a mathematical model of the rotor flux is derived;

[0012] Combining the above two formulas, we can get the formula for the rotor flux:

[0013] (3)

[0014] where Φ=[Φ α ,Φ β ] T , since ψ=[ψ α , ψ β ] T It is difficult to measure, but by differentiating it, it can be transformed into the voltage and current based on the stationary coordinate system, which are easily obtained on the right side of formula (2); thus, differentiating both sides of the equation can obtain

[0015] (4)

[0016] In order to estimate the rotor position information θ from the above formula, as long as the rotor flux Φ=[Φ α ,Φ β ] T , the rotor position information θ can be obtained from the following formula

[0017] (5)

[0018] Step 3: Based on step 2, introduce zero initial quantity and construct the rotor flux observer;

[0019] Create a reference model with zero initial quantities

[0020] (6)

[0021] where φ=[φ α ,φ β ] T , note that Φ and φ have the same structure, but different initial conditions; for the convenience of notation, ξ≡ψ(0)-Li(0) is introduced here, and it can be obtained that

[0022] (7)

[0023] Step 4: Construct an initial parameter estimation observer;

[0024] According to formula (7), we can derive the algebraic relationship ||φ+ξ|| 2 =||Φ|| 2 , according to formula (3), we can derive the algebraic relationship ||Φ|| 2 =ψ m 2 According to these two formulas, we can deduce

[0025] (8)

[0026] Apply the filter as / (s+a) to both ends of the formula and we get

[0027] (9)

[0028] (10)

[0029] (11)

[0030] ||φ|| 2 -ψ m 2 is a constant, and the error part in formula (9) is derived as follows

[0031] (12)

[0032] Both Y(φ) and A(φ) in formula (9) can be obtained by calculating φ, so formula (9) can be regarded as a linear equation with unknown parameter ξ;

[0033] The error (Y(φ)-A(φ) T ξ) 2 Minimize, and you can get

[0034] (13)

[0035] Where K is the gain matrix; According to the relationship in formula (13), the initial parameter observer can be constructed, and the rotor flux observer and the initial parameter estimation observer can be combined according to formula (7) to obtain the rotor position information;

[0036] Step 5: Construct a speed observer with a phase-locked loop;

[0037] The rotor position estimated by the inverse tangent function has a jump between -π and π. In order to eliminate the error caused by this jump, a phase-locked loop is introduced in the speed estimation part to construct a speed observer. The trigonometric function difference is used to replace the angle difference, as shown in formula (14):

[0038] (14)

[0039] The speed can be obtained by differentiating the rotor position, as shown in formula (15):

[0040] (15)

[0041] Where ω is the motor speed, the speed observer, the closed-loop transfer function can be expressed as

[0042] (16)

[0043] where ω n is the natural oscillation frequency, ζ is the damping coefficient, and the PI coefficient of the phase-locked loop can be obtained by the following formula;

[0044] (17)

[0045] Where kp, ki are the PI coefficients of the phase-locked loop;

[0046] Step 6: Build an ID controller to eliminate rotor position error;

[0047] The error between the current rotor position and the estimated rotor position is represented by ∆θ;

[0048] (18)

[0049] In order to eliminate the rotor position error, the ID injection method is adopted when defining the ID controller. dr and I qr is the reference current for synchronizing the actual speed of the d-axis and q-axis,

[0050] (19)

[0051] (20)

[0052] Add the structure in the above steps to the system, transmit the voltage and current signals measured in the motor to the rotor flux observer, obtain the observed value of the rotor position information by compensating the initial value, send the obtained observed value to the speed observer, and the obtained rotor speed information is fed into the ID controller to form a speed closed loop. The inner loop adopts the pi current closed loop, and then the signal processed by the inner loop is sent to the inverter bridge in the form of SVPWM.

[0053] Beneficial effects:

[0054] 1. The present invention is a sensorless control method for a permanent magnet synchronous motor. A rotor flux observer is added to the control system to solve the shortcomings of traditional mechanical sensors, such as high environmental requirements, slow response, weak anti-electromagnetic interference ability, and aging.

[0055] 2. The present invention solves the problem that medium and low speeds cannot be measured and zero-speed starting cannot be achieved in an observer that uses a back-electromotive force method to measure rotor position information by defining d-axis and q-axis currents.

[0056] 3. The present invention solves the noise problem caused by the high-frequency injection method for measuring the rotor position information and the problem of being unable to detect the rotor position of the surface-mounted permanent magnet synchronous motor. The control method is simple to implement and does not require the injection of high-frequency signals, thereby avoiding high-frequency noise. There is no need to observe the back electromotive force, thereby improving the performance of the motor at zero and low speeds. Description of the drawings:

[0057] Figure 1 It is the system diagram of sensorless control of permanent magnet synchronous motor;

[0058] Figure 2 It is the framework diagram of the flux observer;

[0059] Figure 3 It is a diagram of a speed observer with a phase-locked loop;

[0060] Figure 4 It is a simplified diagram of the speed observer. Specific implementation method:

[0061] Embodiment 1:

[0062] A sensorless control method for a permanent magnet synchronous motor, the algorithm comprises the following steps:

[0063] Step 1, establishing a mathematical model of a surface-mounted permanent magnet synchronous motor, that is, a motor model of the permanent magnet synchronous motor based on a stationary coordinate system (α-β coordinate system);

[0064] (1)

[0065] (2)

[0066] Where ψ=[ψ α , ψ β ] T is the stator flux in the stationary coordinate system, ψ m is the permanent magnet flux constant, L S is the stator winding inductance, i ab =[i α , i β ]T and is the stator current in the stationary coordinate system, u ab =[u α ,u β ] T is the stator voltage in the stationary coordinate system, R is the stator coil winding resistance, and θ is the rotor position;

[0067] Step 2: In order to obtain the rotor position information, a mathematical model of the rotor flux is derived;

[0068] Combining the above two formulas, we can get the formula for the rotor flux:

[0069] (3)

[0070] where Φ=[Φ α ,Φ β ] T , since ψ=[ψ α , ψ β ] T It is difficult to measure, but by differentiating it, it can be transformed into the voltage and current based on the stationary coordinate system, which are easily obtained on the right side of formula (2); thus, differentiating both sides of the equation can obtain

[0071] (4)

[0072] In order to estimate the rotor position information θ from the above formula, as long as the rotor flux Φ=[Φ α ,Φ β ] T , the rotor position information θ can be obtained from the following formula

[0073] (5)

[0074] Step 3: Based on step 2, introduce zero initial quantity and construct the rotor flux observer;

[0075] Create a reference model with zero initial quantities

[0076] (6)

[0077] where φ=[φ α ,φ β ] T , note that Φ and φ have the same structure, but different initial conditions; for the convenience of notation, ξ≡ψ(0)-Li(0) is introduced here, and it can be obtained that

[0078] (7)

[0079] Step 4: Construct an initial parameter estimation observer;

[0080] According to formula (7), we can derive the algebraic relationship ||φ+ξ|| 2 =||Φ|| 2 , according to formula (3), we can derive the algebraic relationship ||Φ|| 2 =ψ m 2 According to these two formulas, we can deduce

[0081] (8)

[0082] Apply the filter as / (s+a) to both ends of the formula and we get

[0083] (9)

[0084] (10)

[0085] (11)

[0086] ||φ|| 2 -ψ m 2 is a constant, and the error part in formula (9) is derived as follows

[0087] (12)

[0088] Both Y(φ) and A(φ) in formula (9) can be obtained by calculating φ, so formula (9) can be regarded as a linear equation with unknown parameter ξ;

[0089] The error (Y(φ)-A(φ) T ξ) 2 Minimize, and you can get

[0090] (13)

[0091] Where K is the gain matrix; According to the relationship in formula (13), the initial parameter observer can be constructed, and the rotor flux observer and the initial parameter estimation observer can be combined according to formula (7) to obtain the rotor position information;

[0092] Step 5: Construct a speed observer with a phase-locked loop;

[0093] The rotor position estimated by the inverse tangent function has a jump between -π and π. In order to eliminate the error caused by this jump, a phase-locked loop is introduced in the speed estimation part to construct a speed observer. The trigonometric function difference is used to replace the angle difference, as shown in formula (14):

[0094] (14)

[0095] The speed can be obtained by differentiating the rotor position, as shown in formula (15):

[0096] (15)

[0097] Where ω is the motor speed, the speed observer, the closed-loop transfer function can be expressed as

[0098] (16)

[0099] where ω n is the natural oscillation frequency, ζ is the damping coefficient, and the PI coefficient of the phase-locked loop can be obtained by the following formula;

[0100] (17)

[0101] Where kp, ki are the PI coefficients of the phase-locked loop;

[0102] Step 6: Build an ID controller to eliminate rotor position error;

[0103] The error between the current rotor position and the estimated rotor position is represented by ∆θ;

[0104] (18)

[0105] In order to eliminate the rotor position error, the ID injection method is adopted when defining the ID controller. dr and I qr is the reference current for synchronizing the actual speed of the d-axis and q-axis,

[0106] (19)

[0107] (20)

[0108] Add the structure in the above steps to the system, transmit the voltage and current signals measured in the motor to the rotor flux observer, obtain the observed value of the rotor position information by compensating the initial value, send the obtained observed value to the speed observer, and the obtained rotor speed information is fed into the ID controller to form a speed closed loop. The inner loop adopts the pi current closed loop, and then the signal processed by the inner loop is sent to the inverter bridge in the form of SVPWM.

Claims

1. A sensorless control method for a permanent magnet synchronous motor, characterized in that: The control method comprises the following steps: Step 1, establishing a mathematical model of a surface-mounted permanent magnet synchronous motor, that is, a motor model of the permanent magnet synchronous motor based on a stationary coordinate system α-β coordinate system; ； in, is the stator flux in the stationary coordinate system, is the permanent magnet flux constant, is the stator winding inductance, and are the stator currents in the stationary coordinate system, is the stator voltage in the stationary coordinate system, is the stator coil winding resistance, is the rotor position; Step 2, establishing a rotor flux model; (2) in ; Step 3, in order to obtain the rotor position information, derive the mathematical model of the rotor flux; Combining the above two formulas, we can get the formula for the rotor flux: ； in ,because It is difficult to measure, but by differentiating it, it can be transformed into the voltage and current based on the stationary coordinate system, which are easily obtained on the right side of formula (2); thus, differentiating both sides of the equation gives ； In order to estimate the rotor position information θ from the above formula, as long as the rotor flux is obtained , the rotor position information θ can be obtained from the following formula ； Step 3: Based on step 2, introduce zero initial quantity and construct the rotor flux observer; Create a reference model with zero initial quantities ； in ,Note that Φ and φ have the same structure, but different initial conditions; for the convenience of notation, we introduce , it can be concluded that ； Step 4: Construct an initial parameter estimation observer; According to formula (7), we can get the algebraic relationship According to formula (3), we can get the algebraic relationship According to these two formulas, we can deduce ； Apply the filter as / (s+a) to both ends of the formula, and we get ； ； is a constant, and the error part in formula (9) is derived as follows ； Both Y(φ) and A(φ) in formula (9) can be obtained by calculating φ, so formula (9) can be regarded as a linear equation with unknown parameter ξ; The error Minimize, and you can get ； Where K is the gain matrix; According to the relationship in formula (13), the initial parameter observer can be constructed, and the rotor flux observer and the initial parameter estimation observer can be combined according to formula (7) to obtain the rotor position information; Step 5: Construct a speed observer with a phase-locked loop; The rotor position estimated by the inverse tangent function has a jump between -π and π. In order to eliminate the error caused by this jump, a phase-locked loop is introduced in the speed estimation part to construct a speed observer, and the trigonometric function difference is used to replace the angle difference, as shown in formula (14): ； The speed can be obtained by differentiating the rotor position, as shown in formula (15): ； Where ω is the motor speed, the speed observer, the closed-loop transfer function can be expressed as ； Where ω n is the natural oscillation frequency, ζ is the damping coefficient, and the PI coefficient of the phase-locked loop can be obtained by the following formula; ； Where kp, ki are the PI coefficients of the phase-locked loop; Step 6: Build an ID controller to eliminate rotor position error; The error between the current rotor position and the estimated rotor position is represented by ∆θ; ； In order to eliminate the rotor position error, the ID injection method is used when defining the ID controller. and is the reference current for the actual speed synchronization of the d-axis and q-axis, ； Add the structure in the above steps to the system, transmit the voltage and current signals measured in the motor to the rotor flux observer, obtain the observed value of the rotor position information by compensating the initial value, send the obtained observed value to the speed observer, and the obtained rotor speed information is fed into the ID controller to form a speed closed loop. The inner loop adopts the pi current closed loop, and then the signal processed by the inner loop is sent to the inverter bridge in the form of SVPWM.

Images