CN103916065A - Estimation method for static initial position of electro-magnetic synchronous motor rotor of no-position sensor - Google Patents

Abstract

The invention relates to a method for estimating the static initial position of the rotor of an electrically excited synchronous motor without a position sensor. By applying an excitation voltage to the rotor of an electrically excited synchronous motor, the voltage space vector is applied to the stator multiple times, and the voltage on the stator is collected during each application process. The maximum current response value i _A , i _B , i _C , calculate the three-phase stator current response value and obtain the current i _d ′ in the d′q′ coordinate system through coordinate transformation, and then obtain the voltage corresponding to the current i _d ′ according to i _{d ′} The angle between the space vector and the A-axis of the static coordinate system of the motor is used to estimate the initial position value of the rotor of the electrically excited synchronous motor. Beneficial effect: it can reduce the error caused by the fluctuation of the rotor excitation flux chain when the electric excitation synchronous motor uses the inductance saturation effect to estimate the initial position of the rotor.

Description

A method for estimating the static initial position of the rotor of an electrically excited synchronous motor without a position sensor

technical field

The invention belongs to the technical field of AC motor drive control, and in particular relates to a method for estimating the static initial position of a rotor of an electrically excited synchronous motor without a position sensor.

Background technique

The initial position of the rotor (the position of the N pole) is a very important quantity in the drive control of the electric excitation synchronous motor. An inaccurate initial position will lead to problems such as the reduction of the motor's load capacity and reverse rotation. Commonly used rotor position detection devices such as photoelectric encoders, resolvers, and Hall elements cannot directly obtain the initial position of the rotor due to mechanical installation errors; the installation of position sensors will also increase the cost, volume and weight of the system, resulting in system reliability. The performance is reduced, and in some cases, the installation of the position sensor is not allowed or cannot be realized. Due to the limitations of the installation and use of the electric excitation synchronous motor, it is often necessary to obtain the initial position when the rotor is stationary.

At present, there are many methods for estimating the initial position of the motor rotor, but mainly for permanent magnet synchronous motors. There are two simple and practical methods:

(1) Applying a constant voltage vector method. Applying a constant voltage space vector to the stator causes the rotor to turn to a specific position. However, this method is no longer applicable when the motor is required to obtain the initial position of the rotor under static conditions. At the same time, a constant voltage space vector will generate a large stator current, which may cause damage to the motor body.

(2) It is a commonly used method to detect the initial position of the rotor by using the saturation effect of the stator inductance. This method is simple and easy to implement. A voltage space vector of equal amplitude (360° electrical angle) is applied to the stator, and the three-phase current of the stator is collected at the same time. The rotor position is estimated by calculating and comparing the corresponding current response values. However, since the flux linkage of the excitation winding of the electric excitation synchronous motor rotor is not as stable as that of the permanent magnet synchronous motor rotor, there will be a large estimation error when using this method directly.

Theoretically speaking, the estimation accuracy of this method increases with the subdivision degree of the voltage space vector. However, in actual use, due to the existence of detection element errors and switch tube dead time, it is difficult to further improve the estimation accuracy only through voltage space vector subdivision, so it is not suitable for applications that require high initial rotor positions.

Contents of the invention

technical problem to be solved

In order to avoid the deficiencies of the prior art, the present invention proposes a method for estimating the static initial position of the electric excitation synchronous motor rotor without a position sensor, and solves the problem of using the inductance saturation effect to detect the initial position of the permanent magnet synchronous motor rotor for electric excitation synchronous For the problem of low motor detection accuracy, while reducing the estimation error caused by the fluctuation of the rotor excitation flux linkage of the electrically excited synchronous motor, the estimation accuracy of the initial rotor position can be further improved under a certain voltage space vector subdivision condition.

Technical solutions

A method for estimating the static initial position of the rotor of an electrically excited synchronous motor without a position sensor, characterized in that the steps are as follows:

Step 1: Apply an excitation voltage to the rotor of an electrically excited synchronous motor;

Step 2: Apply a voltage space vector of a certain magnitude to the stator, wait for the current on the stator to be zero, apply the voltage space vector of this magnitude again, and the number of cycles is greater than n and greater than 24 times;

The application time is the cycle of the PWM wave sent by the microcontroller; the amplitude is guaranteed not to damage the motor body;

Step 3: Collect the maximum current response values i _A , i _B , and i _C on the stator during each application process, calculate the three-phase stator current response values, and obtain the current i _d ′ in the d′q′ coordinate system through coordinate transformation, and obtain n i _d ';

Among them: θ is the angle between the voltage space vector applied each time and the A-axis of the motor's static coordinate system

Step 4: Calculate all i _d ′ currents greater than or equal to i _d ′ _max -Δi _d ′ among n i d ′, and find out the θ value corresponding to i _{d ′} current, and obtain the maximum value θ _max among all θ values with the minimum value θ _min ;

The Δi _d 'is 5-10A;

Step 5: Preliminary estimation of the initial rotor position value θ _{r_temp} of the electrically excited synchronous motor:

1. When θ _max - θ _min <180°, θ _{r_temp} = (θ _max + θ _min )/2;

2. When θ _max - θ _min ≥ 180°,

If 360°-θ _max ≥ _{θ min} , then θ _{r_temp} = (θ _max + θ _min + 360°)/2,

If 360°-θ _max < θ _min , then θ _{r_temp} = (θ _max + θ _min - 360°)/2;

Step 6: When θ _{r_temp} -90°≥0°, calculate the intermediate variable θ _temp = θ _{r_temp} -90°; when θ _{r_temp} -90°<0°, calculate the intermediate variable θ _temp = θ _{r_temp} -90°+360°;

Step 7: Let θ _k = 360k/n, the corresponding d′ axis current is i _d ′(θ _k ), when θ _k ≥ 360°, i _d ′(θ _k )=i _d ′(θ _{k －n} ), when θ _k <0°, i _d ′(θ _k )=i _d ′(θ _k+n ); find the k value m that minimizes |θ _k －θ _temp |, and the corresponding d′ axis current is i _d ′(θ _m ); where k is an integer;

Step 8: Let Δi _d ′(θ _m )=i _d ′(θ _m )－i _d ′(θ _m+n/2 ), Δi _d ′(θ _m+1 )=i _d ′(θ _m+1 )－i _d ′(θ _m+1+n/2 ),…, Δi _d ′(θ _m+n/2－1 )=i _d ′(θ _m+n/2－1 )－i _d ′( θ _{m+n/2－1+n/2} );

Step 9: For discrete points (θ _m , Δi _d ′(θ _m )), (θ _m+1 , Δi _d ′(θ _m+1 )), …, (θ _m+n/2－1 , Δi _d ′(θ _m+n/2－1 )) for Gaussian curve fitting, the Gaussian function is y=Aexp(-(x-μ) ² /(2σ ² )), and the fitting coefficient μ is obtained; since the Gaussian function is the highest Point theoretically represents the difference between the maximum and minimum response current of the rotor, corresponding to the position of the N pole of the rotor, so the estimated rotor position θ _r =μ, if μ≥360°, the estimated rotor position θ _r =μ-360 °; if μ<0°, the estimated rotor position θ _r =μ+360°.

Said 24≤n≤120.

Beneficial effect

A method for estimating the static initial position of the rotor of an electrically excited synchronous motor without a position sensor proposed by the present invention. By applying an excitation voltage to the rotor of an electrically excited synchronous motor, the voltage space vector is applied to the stator multiple times, and the voltage space vector on the stator is collected during each application process. The maximum current response value i _A , i _B , i _C , calculate the three-phase stator current response value and obtain the current i _d ′ in the d′q′ coordinate system through coordinate transformation, and then obtain the corresponding i _d ′ current according to i _{d ′} The angle between the voltage space vector and the A-axis of the motor's static coordinate system is used to estimate the initial position value of the rotor of the electrically excited synchronous motor. Beneficial effect: it can reduce the error caused by the fluctuation of the rotor excitation flux chain when the electric excitation synchronous motor uses the inductance saturation effect to estimate the initial position of the rotor.

Description of drawings

Figure 1: Schematic diagram of the amplitude selection requirements of the voltage space vector

Figure 2: Schematic diagram of the space vector of the voltage applied to the stator of the motor

Figure 3: Schematic diagram of the d'q' coordinate system

Figure 4: Stator three-phase current response curve

Figure 5: Schematic diagram of the 72 applied voltage space vectors

Figure 6: Stator three-phase current response curves at different positions

Figure 7: Curves of current i _d ′ at different positions

Figure 8: Schematic diagram of θ _max and θ _min

Figure 9: Drawing curves from coordinate points

Figure 10: Gaussian Fitting Curve

Detailed ways

Now in conjunction with embodiment, accompanying drawing, the present invention will be further described:

(1) Apply an excitation voltage to the rotor of an electrically excited synchronous motor.

(2) Determine the magnitude of the applied voltage space vector, the application time and the time of the zero vector inserted between two adjacent voltage space vectors. The amplitude of the voltage space vector increases gradually from 0, and the application time is the cycle of the PWM wave sent by the microcontroller, and the time of the zero vector inserted between two adjacent voltage space vectors depends on the amplitude of the voltage space vector and the application time. Generally, it can be gradually increased from 1ms. This step can be completed by observing with an oscilloscope to ensure that the above parameters can meet the requirements of not damaging the motor body. At the same time, it should be satisfied that the three-phase current of the stator has decayed to zero before the voltage space vector is applied each time, as shown in Figure 1.

In this embodiment, the amplitude U _m of the applied voltage space vector is determined to be 0.3 (modulation degree), the application time is 200 us, and the time for inserting a zero vector between two adjacent voltage space vectors is 6 ms.

(3) Select the number n of voltage space vectors applied (24≤n≤120), as shown in Figure 2 (taking n=24 as an example), and set the voltage space vector applied each time to the A-axis of the static coordinate system of the motor The included angle is θ. The number of voltage space vectors should not be too small (more than 24), otherwise there will be large errors in estimation. At the same time, the number of voltage space vectors is required to be an even number.

In this embodiment, the number of applied voltage space vectors is selected to be n=72, as shown in FIG. 5 .

(4) As shown in Figure 3, the αβ coordinate system is fixed on the stator winding of the motor, the dq coordinate system is a synchronous rotating coordinate system, the d axis is in the same direction as the N pole of the rotor, _θr represents the actual position of the rotor, and d′q ' The coordinate system is the estimated coordinate system. Let the angle between the d'q' coordinate system and the A axis of the static coordinate system be θ, and the d' axis is in the same direction as the voltage space vector U _s applied to the stator. In order to prevent the motor from rotating, the voltage space vector is applied to the motor stator according to the sequence 1-24 shown in Figure 2 (to ensure that the three-phase current of the stator has been reduced to 0 before each voltage space vector is applied), and at the same time, it is collected by the current sensor Stator three-phase current, as shown in Figure 4, take out the stator three-phase current response values i _A , i _B , and i _C corresponding to each time the voltage space vector is applied.

In this embodiment, the voltage space vector is applied to the stator of the motor according to the sequence 1 to 72 shown in Figure 5, and at the same time, the stator three-phase current is collected through the current sensor, and the corresponding stator three-phase current response value is taken out when the voltage space vector is applied each time. i _A , i _B , i _C . The current response values for all positions are shown in Fig. 6.

(5) The obtained three-phase stator current response value is obtained by coordinate transformation to obtain the current i _d ′ in the d′q′ coordinate system, and the coordinate transformation is shown in the following formula.

The results of this embodiment are shown in Figure 7.

(6) As shown in Figure 3, the angle between the d′q′ coordinate system and the d-axis where the rotor axis is located is δ. As |δ| decreases, the excitation flux linkage generated by the rotor winding under the excitation voltage will be There are more and more interlinkages with the stator d′ axis winding, making L _d ′ smaller and smaller. Therefore, compared with the voltage space vectors of the same amplitude at other angles, the d′ axis coincides with the d axis This position has the greatest rate of change of current. Due to the fluctuation of the rotor flux linkage of the electrically excited synchronous motor, there is a shaded part as shown in Fig. 3, and the current response i _d ' obtained by the applied voltage space vector in this part area is almost the same within the error range. By comparison, the maximum d′-axis current i _d ′ _max calculated in step (5) is obtained, and the error Δi _d ′ is determined by the current acquisition accuracy, and all i _d ′ currents greater than or equal to i _d ′ _max -Δi _d ′ are obtained, And find out the corresponding θ value, and compare the maximum value θ _max and the minimum value θ _min among them.

In this embodiment, i _d ′ _max = 255A, determine Δi _d ′=6A, find out the angle values corresponding to all i d ′ currents greater than or equal to i _d ′ _max -Δi _d ′=249A _, and find the maximum The value θ _max =95° and the minimum value θ _min =75° are shown in FIG. 8 .

(7) Preliminary estimation of the initial position value of the electric excitation synchronous motor rotor: θ _{r_temp} = (θ _max + θ _min )/2, when θ _max - _{θ min} > 180°, if 360° - θ _max ≥ θ _min , then θ _{r_temp} = (θ _max + θ _min + 360°)/2, if 360° - θ _max < θ _min , then θ _{r_temp} = (θ _max + θ _min - 360°)/2.

This embodiment preliminarily estimates the initial position value of the rotor of the electrically excited synchronous motor: θ _{r_temp} = (θ _max + θ _min )/2 = 85°.

(8) Calculate θ _temp = θ _{r_temp} -90°, if θ _{r_temp} -90°<0°, then θ _temp = θ _{r_temp} -90°+360°.

In this embodiment, θ _temp =θ _{r_temp} −90°+360°=355° is calculated.

(9) Let θ _k = 360k/n, where n is the number of voltage space vectors and k is an integer. The corresponding d′ axis current is i _d ′(θ _k ), when θ _k ≥360°, i _d ′(θ _k )=i _d ′(θ _k－n ), when θ _k <0°, i _d '(θ _k )=i _d '(θ _k+n ). Find the k value m that minimizes |θ _k －θ _temp |, and the corresponding d′-axis current is i _d ′(θ _m ).

In this embodiment, the k value m=71 that minimizes |θ _k - θ _temp | is obtained, and the corresponding d'-axis current is i _d '(θ ₇₁ ).

(10) Let Δi _d ′(θ _m )=i _d ′(θ _m )－i _d ′(θ _m+n/2 ), Δi _d ′(θ _m+1 )=i _d ′(θ _m+1 )－i _d ′(θ _m+1+n/2 ),…, Δi _d ′(θ _{m+n/2 －1} )=i _d ′(θ _m+n/2－1 )－i _d ′( θ _{m+n/2－1+n/2} ).

In this embodiment, Δi _d ′(θ ₇₁ )=i _d ′(θ ₇₁ )-i _d ′(θ ₁₀₇ ), Δi _d ′(θ ₇₂ )=i _d ′(θ ₇₂ )-i _d ′(θ ₁₀₈ ),..., Δi _d ′(θ ₁₀₆ )=i _d ′(θ ₁₀₆ )−i _d ′(θ ₁₄₂ ).

(11) For discrete points (θ _m , Δi _d ′(θ _m )), (θ _m+1 , Δi _d ′(θ _m+1 )), …, (θ _m+n/2－1 , Δi _d ′(θ _m+n/2－1 )) for Gaussian curve fitting (Gaussian curve fitting method can be found in numerical analysis textbooks), the Gaussian function is y=Aexp(-(x-μ) ² /(2σ ² ) ) to find the fitting coefficient μ. Since the highest point of the Gaussian function theoretically represents the difference between the maximum and minimum response current of the rotor, corresponding to the position of the N pole of the rotor, the estimated rotor position θ _r =μ, if μ≥360°, the estimated rotor position θ _r =μ−360°; if μ<0°, the estimated rotor position θ _r =μ+360°.

From (θ ₇₁ , Δi _d ′(θ ₇₁ )), (θ ₇₂ , Δi _d ′(θ ₇₂ )), …, (θ ₁₀₆ , Δi _d ′(θ ₁₀₆ )) to draw a plot, as shown in Figure 9 , it can be seen that its shape is similar to the Gaussian distribution, so the Gaussian function is used to fit the data. Through Gaussian curve fitting, μ=444.8, σ=35.1, A=47.07 are obtained, and the fitting results are shown in Fig. 10 . Since μ≥360°, the estimated rotor position θ _r =μ−360°=84.8°.

The difference between the rotor position estimated by the embodiment of the present invention and the actual motor rotor position of 84.381° is 0.419 electrical angles, which fully meets the requirements of drive control.

Claims (2)

1. an electric excitation synchronous motor stationary rotor initial position evaluation method for position-sensor-free, is characterized in that step is as follows:

Step 1: electric excitation synchronous motor rotor is applied to exciting voltage;

Step 2: to applying a certain amplitude voltage space vector on stator, wait for that the electric current on stator is at 1 o'clock, again apply the space vector of voltage of this amplitude, cycle-index is greater than n and is greater than 24 times;

Described application time is that microcontroller is sent out PWM wave period; Described amplitude guarantees not damage motor body;

Step 3: gather and apply the maximum current response i on stator in process at every turn _a, i _b, i _c, calculate threephase stator current response value and obtain the current i under d ' q ' coordinate system by coordinate transform _d', obtain n i _d';

Wherein: θ is the space vector of voltage that at every turn applies and the angle of motor rest frame A axle

Step 4: obtain n i _d' in all i that are more than or equal to _d' _max-Δ i _d' i _d' electric current, and find out corresponding i _dthe θ value of ' electric current, obtains the maximum θ in all θ values _maxwith minimum value θ _min;

Described Δ i _d' be 5～10A;

Step 5: preresearch estimates electric excitation synchronous motor initial position of rotor value θ _{r_temp}:

1, work as θ _max-θ _minwhen 180 ° of <, θ _{r_temp}=(θ _max+ θ _min)/2;

2, work as θ _max-θ _min>=180 ° time,

If 360 °-θ _max>=θ _min, θ so _{r_temp}=(θ _max+ θ _min+ 360 °)/2,

If 360 °-θ _max< θ _min, θ so _{r_temp}=(θ _max+ θ _min-360 °)/2;

Step 6: work as θ _{r_temp}-90 °>=0 °, calculate intermediate variable θ _temp=θ _{r_temp}-90 °; Work as θ _{r_temp}0 ° of-90 ° of <, calculates intermediate variable θ _temp=θ _{r_temp}-90 °+360 °;

Step 7: make θ _k=360k/n, corresponding d ' shaft current is i _d' (θ _k), work as θ _k>=360 ° time, i _d' (θ _k)=i _d' (θ _{k -n}), work as θ _kwhen 0 ° of <, i _d' (θ _k)=i _d' (θ _k+n); Obtain and make | θ _k-θ _temp| minimum k value m, corresponding d ' shaft current is i _d' (θ _m); Wherein k is integer;

Step 8: make Δ i _d' (θ _m)=i _d' (θ _m)-i _d' (θ _m+n/ ₂), Δ i _d' (θ _m+1)=i _d' (θ _m+1)-i _d' (θ _m+1+n/ ₂) ..., Δ i _d' (θ _m+n/2-1)=i _d' (θ _m+n/2-1)-i _d' (θ _m+n/2-1+n/ ₂);

Step 9: to discrete point (θ _m, Δ i _d' (θ _m)), (θ _m+1, Δ i _d' (θ _m+1)) ..., (θ _m+n/2-1, Δ i _d' (θ _m+n/2-1)) carrying out gaussian curve approximation, Gaussian function is y=Aexp ((x-μ) ²/ (2 σ ²)), obtain fitting coefficient μ; Because Gaussian function peak is representing the difference of rotor maximum, minimum response electric current in theory, corresponding the position of the rotor N utmost point, so the rotor position of estimation _r=μ, if μ>=360 °, the rotor position of estimation _r=μ-360 °; If 0 ° of μ <, the rotor position of estimation _r=μ+360 °.

2. the electric excitation synchronous motor stationary rotor initial position evaluation method of position-sensor-free according to claim 1, is characterized in that: described 24≤n≤120.