A speed sensorless method based on variable coefficient feedforward frequency

Abstract

The present invention discloses a speed sensorless method based on a variable coefficient feedforward frequency, which belongs to the field of electromechanical control, including: converting the three-phase voltage and current of a permanent magnet motor into components in a two-phase stationary coordinate system; obtaining a back electromotive force voltage from a PMSM control equation, and filtering out harmonics through a low-pass filter; performing a rotation transformation on the back electromotive force voltage after low-pass filtering to obtain the dq axis components of the back electromotive force voltage; constructing a feedforward frequency observer control equation to obtain a phase-locked loop feedforward frequency; using the feedforward frequency as a phase-locked loop feedforward loop, and obtaining the rotor speed and angle through PI regulation. The present invention dynamically adjusts the feedforward phase-locked loop parameters according to the rotation speed, and can accurately track the rotor speed and position information.

Description

Technical Field

[0001] The invention relates to the field of electromechanical control, and in particular to a speed sensorless method based on variable coefficient feedforward frequency. Background Art

[0002] As the capacity of wind turbines continues to increase, higher requirements are placed on the control performance of converters. In order to establish an accurate and reliable permanent magnet synchronous motor (PMSM) control system, it is necessary to accurately obtain the motor rotor position and speed information. There are three main speed sensorless methods for estimating PMSM speed in wind power converters: a speed sensorless algorithm based on back-EMF voltage, a speed sensorless algorithm based on sliding mode observer, and a speed sensorless algorithm based on model reference adaptation.

[0003] Speed ​​sensorless algorithm based on back-EMF voltage: First, the back-EMF voltage Ud / Uq is calculated using the PMSM control equation in the dq coordinate system, and then the harmonic components of the back-EMF voltage are filtered out through a low-pass filter. Finally, the back-EMF voltage in the dq coordinate system is phase-locked by a traditional phase-locked loop to obtain the rotor angular velocity and angle. The back-EMF method for estimating the rotor angular velocity and angle is sensitive to the motor parameters. When the motor parameters change or the motor speed is low, the estimated angular velocity and angle will have large errors and poor robustness.

[0004] Speed ​​sensorless algorithm based on sliding mode observer: The sliding mode observer is designed based on the error between the given current and the feedback current, and the back electromotive force voltage of the motor is reconstructed by the error, and finally the rotor speed and angle information are estimated. The key to the implementation of this method lies in the selection of sliding mode surface function and sliding mode gain. The sliding mode surface function is difficult to select, the system has jitter, and the calculation process is complicated.

[0005] Speed ​​sensorless algorithm based on model reference adaptation: Model reference adaptation includes three parts: adjustable model, reference model, and adaptive law. The expression without unknown parameters is used as the reference model, and the expression with the parameters to be identified is used for the adjustable model, and the two models have outputs with the same physical meaning. The difference between the outputs of the two models is used to identify the speed parameters of the permanent magnet motor through a suitable adaptive law. However, the adaptive law is difficult to design. Since the system itself is asymptotically stable, the speed estimation error has a large value in the rising stage of the speed, but as the speed increases and stabilizes, the speed estimation error gradually decreases, and the rotor position estimation error also gradually decreases, and the calculation process is complicated. Summary of the invention

[0006] The present invention aims to solve the shortcomings of the speed sensorless method for estimating the speed of the converter in the prior art, and proposes a speed sensorless method based on a variable coefficient feedforward frequency, which has better dynamic performance and can accurately track the rotor speed and position information.

[0007] In order to achieve the above-mentioned object of the invention, the technical solution of the present invention is as follows:

[0008] A speed sensorless method based on variable coefficient feedforward frequency, characterized by comprising the following steps:

[0009] Convert the three-phase voltage and current of the permanent magnet motor into components in a two-phase stationary coordinate system;

[0010] The back EMF voltage is obtained from the PMSM control equation and the harmonics are filtered out by a low-pass filter;

[0011] Performing a rotation transformation on the low-pass filtered back electromotive force voltage to obtain the d-axis and q-axis components of the back electromotive force voltage;

[0012] Construct the feedforward frequency observer control equation and obtain the feedforward frequency of the phase-locked loop;

[0013] The feedforward frequency is used as the feedforward loop of the phase-locked loop, and the rotor speed and angle are obtained through PI regulation.

[0014] Furthermore, the construction of the feedforward frequency observer control equation and the acquisition of the feedforward frequency of the phase-locked loop include: determining the acceleration of the rotor estimated speed according to the feedforward frequency observer control equation; and obtaining the rotor estimated speed by discrete integration of the acceleration of the rotor estimated speed, which is the feedforward frequency.

[0015] Furthermore, the acceleration of the estimated rotor speed is equal to k1 times the estimated speed of the previous control cycle plus k0 times the angle error of the permanent magnet motor.

[0016] Furthermore, the estimated rotor speed obtained by discretely integrating the acceleration of the estimated rotor speed is low-pass filtered and then limited; the estimated rotor speed after the limiting process is integrated to obtain the rotor angle, and the rotor angle is used as the rotation angle in the rotation transformation.

[0017] Furthermore, in the process of obtaining the feedforward frequency, the k0 and k1 parameters are dynamically adjusted according to the current estimated speed until the estimated rotor speed and angle can stably track the actual rotor speed and angle.

[0018] Furthermore, the k0 and k1 parameters are dynamically adjusted according to the following relationship:

[0019] Let -k0=c0(1+3*f 估计 / f 额定), k0=-k1 2 , c0 is a constant, f 估计 is the estimated frequency, f 额定 is the rated frequency.

[0020] Furthermore, the angle error of the permanent magnet motor is transformed using the arc tangent of the q-axis component, which is equivalent to the q-axis component divided by the comprehensive vector amplitude during the dynamic tracking process.

[0021] Furthermore, the phase-locked loop is a three-phase software phase-locked loop based on dq transformation, whose input is q-axis voltage, and the compensation term is feedforward frequency. The phase-locked loop estimates the machine-side voltage frequency ω and angle θ to obtain the rotor speed and angle.

[0022] In summary, the present invention has the following advantages:

[0023] 1. The PMSM speed sensorless method with variable coefficient feedforward frequency of the present invention dynamically adjusts the parameters of the feedforward phase-locked loop according to the rotor speed to accurately track the rotor speed and position information. The calculation is simple and the effect is obvious. The parameter accuracy requirement is low and the robustness is good. It can adapt to control systems that cannot accurately estimate motor parameters and is more in line with practical engineering applications.

[0024] 2. In the method of the present invention, the feedforward frequency directly acts on the speed estimation, and the design principle based on the variable coefficient is to speed up the speed when the speed estimation is stable. Therefore, when the speed of the permanent magnet motor changes over a large range, the speed of the permanent magnet motor can be accurately and quickly detected; when the speed suddenly changes and fluctuates, the speed estimation can respond quickly, and the code implementation is small and the understanding is simple, which is of great significance for improving the performance of the control system and has broad application prospects.

[0025] 3. The present invention has been successfully applied to a 1140 voltage level wind power converter, which is simple to implement and has rapid dynamic response. At the same time, the present invention can also provide reference for the speed sensorless rotor speed and position estimation problem in other permanent magnet motor industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 A flow chart of a speed sensorless method based on a variable coefficient feedforward frequency according to the present invention;

[0027] Figure 2 The waveform is a comparison between the rotation speed estimated by the method of the present invention and the actual rotation speed;

[0028] Figure 3 A schematic flow chart of the rotor speed and angle estimation using a feedforward phase-locked loop frequency. DETAILED DESCRIPTION

[0029] In order to explain the present invention more clearly, the present invention is further described below in conjunction with preferred embodiments and drawings. It should be understood by those skilled in the art that the content described below is illustrative rather than restrictive, and should not be used to limit the scope of protection of the present invention.

[0030] In the converter control system, it is very important to accurately and quickly detect the speed of the permanent magnet motor and obtain the coordinate transformation angle for control, which directly affects the control performance. For control, it is necessary to estimate the speed of the permanent magnet motor in real time, integrate it to obtain the coordinate transformation angle for control, and then use the estimated speed and angle to control the power and speed of the motor.

[0031] The present invention proposes a speed sensorless method based on a variable coefficient feedforward frequency, and the specific implementation steps are as follows:

[0032] Step 1: Use hardware circuit sampling to obtain the ABC three-phase voltage and three-phase current of the permanent magnet synchronous motor, and convert them into a two-phase stationary coordinate system through Clark transformation;

[0033] Step 2: Obtain the back electromotive force voltage by constructing the PMSM control equation, and filter out the harmonics of the back electromotive force voltage through a low-pass filter;

[0034] Step 3: Convert the back electromotive force voltage in the two-phase stationary coordinate system after low-pass filtering to the rotating coordinate system through Park transformation to obtain the d-axis and q-axis components of the back electromotive force voltage;

[0035] Step 4: Construct a feedforward frequency observer control equation to obtain the phase-locked loop feedforward frequency;

[0036] In this step, the control equation of the constructed feedforward frequency observer is:

[0037] a * (t) = k1f * (t)+k0(θ * (t)-θ(t));

[0038] Where θ(t) represents the actual motor position, f * (t) is the estimated motor frequency, θ * (t) is the estimated motor position, a * (t) is the estimated acceleration.

[0039] According to the established feedforward frequency observer control equation, the acceleration of the estimated rotor speed is equal to k1 times the estimated speed of the previous control cycle plus k0 times the angular error of the permanent magnet motor. Among them, the angular error of the permanent magnet motor uses the arc tangent transformation of the q-axis component, which can be equivalent to the q-axis component divided by the comprehensive vector amplitude during dynamic tracking.

[0040] The estimated rotor speed, ie, the feedforward frequency, can be obtained by discretely integrating the acceleration of the estimated rotor speed.

[0041] The rotor estimated speed obtained by discretely integrating the acceleration of the rotor estimated speed is low-pass filtered and then limited. The rotor estimated speed after the limiting process is integrated to obtain the rotor angle, which is used as the rotation angle of the Park transform in step three.

[0042] In the process of obtaining the feedforward frequency, the k0 and k1 parameters are dynamically adjusted according to the current estimated speed until the estimated rotor speed and angle can stably track the actual rotor speed and angle (the error is within 3‰, which means stable tracking).

[0043] Among them, the k0 and k1 parameters are dynamically adjusted according to the following relationship:

[0044] -k0=c0(1+3×f 估计 / f 额定 ), In the formula, c0 is a constant, f 估计 is the estimated frequency, f 额定 is the rated frequency.

[0045] Step 5: Use the feedforward frequency as the phase-locked loop feedforward loop and obtain the rotor speed and angle through PI regulation.

[0046] In this embodiment, the phase-locked loop adopts a three-phase software phase-locked loop based on dq transformation: the input of the phase-locked loop is the q-axis voltage, the compensation term is the feedforward frequency, and the phase-locked loop is used to estimate the machine-side voltage frequency ω and angle θ, and finally obtain the rotor speed and angle.

[0047] The present invention dynamically adjusts the parameters of the feedforward phase-locked loop according to the rotor speed to accurately track the rotor speed and position information. The calculation is simple and the effect is obvious. The parameter accuracy requirement is low and the robustness is good. It can adapt to the control system that cannot accurately estimate the motor parameters. In addition, the feedforward frequency directly acts on the speed estimation, and the design principle based on the variable coefficient is to speed up the speed under the condition of stable speed estimation. Therefore, when the speed of the permanent magnet motor changes over a large range, the speed of the permanent magnet motor can be accurately and quickly detected; when the speed suddenly changes and fluctuates, the speed estimation can respond quickly.

[0048] The following is a discussion of the feedforward frequency mathematical model of the permanent magnet motor based on back electromotive force.

[0049] For the converter, the current reference direction is the motor convention. The equation for the instantaneous back EMF is:

[0050]

[0051] Among them, E is the back electromotive force, U is the motor terminal voltage, and I is the line current.

[0052] The three-phase voltage u at the output end of the permanent magnet motor can be obtained by sampling a u b u c and three-phase current i a i b i c The instantaneous value of. Through the transformation from three-phase to two-phase stationary coordinate system, that is:

[0053]

[0054]

[0055] Then the back electromotive force is obtained as follows:

[0056]

[0057] Among them, E α 、E β is the back electromotive force of the stationary coordinate system. The back electromotive force is low-pass filtered, and the Laplace expression is omitted in the expression.

[0058] Then, the q-axis voltage is obtained by transforming from the stationary two-phase coordinate system to the voltage synchronous rotating coordinate system, where θ is the back electromotive force voltage angle obtained by the phase-locked loop, namely:

[0059]

[0060] Among them, the feedforward frequency integration gives:

[0061] a(t)=k0*f(t)+k1*Eq(t);

[0062] f 前馈 (t) = ∫a(t)dt;

[0063] θ 前馈 (t) = ∫2*pi*f(t)dt;

[0064] like Figure 3 The figure shows a schematic flow chart of using the feedforward phase-locked loop frequency to estimate the rotor speed and angle. Its input terminal is the three-phase back electromotive force voltage, which is transformed from the three-phase to the two-phase stationary coordinate system to obtain E α 、E β ; Then, after the transformation from the two-phase stationary to the two-phase rotating coordinate system, we can get E d 、E q Then it enters the classic phase-locked loop (PLL) and uses the previous feedforward frequency as the reference frequency to estimate the frequency ω and angle θ of the grid voltage.

[0065] The mathematical model of the feedforward frequency variable coefficient of the permanent magnet motor is explained below.

[0066] According to the relationship between the motor position, speed, and acceleration, the following equations are available. The ones without asterisks are the real motor position, speed, and acceleration, and the ones with asterisks are the estimated motor position, speed, and acceleration. The control of the estimated acceleration is designed as follows:

[0067]

[0068] Differentiating the acceleration and substituting it into the first three equations yields the following relationship:

[0069]

[0070] Transforming the above equation from the time domain to the Laplace domain yields:

[0071]

[0072] make -k1=2*ξ*ω n ,but

[0073] Let -k0=c0(1+3*f 估计 / f 额定 ),

[0074] The above formula is a classic second-order system. According to the second-order system coefficient setting method of automatic control theory, the damping coefficient is usually 0.4~0.8, and when it is 0.5, the relationship between k0 and k1 can be obtained, k0=-k1*k1. It can be seen from the automatic control theory that -k0 is the square of the natural frequency. The larger the k0 is, the farther the pole of the system is from the y-axis, the faster the response is, and the larger the bandwidth is. It can be seen from the acceleration relationship that at low speeds, k0 is smaller, the fluctuation is small, and the dynamic performance is good. At high speeds, k0 is larger, the response is fast, otherwise it is estimated to diverge. Therefore, the size of k0 is designed to change dynamically with the speed, maintain a linear relationship, take into account both rapidity and stability, and the relationship between k0 and k1 remains unchanged.

[0075] Attached Figure 2 The comparison waveforms of the speed estimated by the method of the present invention and the actual speed are shown. It can be seen that the error between the two is within 2‰, which shows that the method can be used to accurately track the rotor speed and position information.

[0076] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any simple modification or equivalent change made to the above embodiment based on the technical essence of the present invention shall fall within the protection scope of the present invention.

Claims

1. A speed sensorless method based on variable coefficient feedforward frequency, characterized in that: The steps include: Convert the three-phase voltage and current of the permanent magnet motor into components in a two-phase stationary coordinate system; The back EMF voltage is obtained from the PMSM control equation and the harmonics are filtered out by a low-pass filter; Performing a rotation transformation on the low-pass filtered back electromotive force voltage to obtain the d-axis and q-axis components of the back electromotive force voltage; Construct the feedforward frequency observer control equation and obtain the feedforward frequency of the phase-locked loop, including: The acceleration of the estimated rotor speed is determined according to the control equation of the feedforward frequency observer. The acceleration of the estimated rotor speed is equal to k1 times the estimated motor frequency of the previous control cycle plus k0 times the angular error of the permanent magnet motor. The estimated rotor speed obtained by discrete integration of the acceleration of the estimated rotor speed is equivalent to the feedforward frequency. In the process of obtaining the feedforward frequency, the k0 and k1 parameters are dynamically adjusted according to the current estimated rotor speed until the estimated rotor speed and angle can stably track the actual rotor speed and angle. Among them, the angular error of the permanent magnet motor uses the inverse tangent transformation of the q-axis component, which is equivalent to the q-axis component divided by the comprehensive vector amplitude during the dynamic tracking process. The k0 and k1 parameters are dynamically adjusted according to the following relationship: make , , c0 is a constant, f 估计 To estimate the motor frequency, f 额定 is the rated frequency; The rotor estimated speed obtained by discrete integration of the acceleration of the rotor estimated speed is subjected to low-pass filtering and then subjected to amplitude limiting processing; the rotor estimated speed after amplitude limiting processing is integrated to obtain a rotor angle, and the rotor angle is used as a rotation angle in the rotation transformation; The feedforward frequency is used as the feedforward loop of the phase-locked loop, and the rotor speed and angle are obtained through PI regulation.

2. A speed sensorless method based on variable coefficient feedforward frequency according to claim 1, characterized in that: The phase-locked loop is a three-phase software phase-locked loop based on dq transformation. Its input is the q-axis voltage, and the compensation term is the feedforward frequency. The phase-locked loop estimates the machine-side voltage frequency. and angle , which is equivalent to the estimated rotor speed and angle.

Images