Induction motor driving method based on random switching frequency pulse width modulation

By using random switching frequency and Markov chain processing technology in the induction motor drive system, the current harmonic energy is dispersed, and the problems of electromagnetic vibration and noise of the induction motor are solved, and the high-frequency noise of the motor is effectively suppressed.

CN118971698APending Publication Date: 2024-11-15JIANGSU UNIV
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411025234.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The electromagnetic vibration and noise problems generated by induction motors in electric vehicles, especially the non-sine waveform and magnetic field non-uniformity caused by high-frequency current harmonics output by the inverter, increase the vibration and noise of the motor, and the existing random SVPWM Technology cannot effectively eliminate high-frequency current harmonics related to carrier frequency.

Method used

Using a pulse width modulation method based on random switching frequency, the switching signal is randomly changed by introducing random variables into the inverter control unit, combining Markov chain and asymmetric SVPWM technology to disperse harmonic energy to make it have a wider frequency bandwidth and optimize the distribution of random values ​​through sinusoidal signals.

Benefits of technology

Effectively suppress high-frequency electromagnetic vibration and narrowband noise of induction motors, reduce the current harmonic amplitude at the switching frequency and its frequency doubling, realize harmonic suppression of the current spectrum, and reduce the high-frequency noise of the motor.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118971698A_ABST
    Figure CN118971698A_ABST
Patent Text Reader

Abstract

The invention discloses an induction motor driving method based on random switching frequency pulse width modulation, and the method comprises the following steps: 1, deducing a switching period calculation formula according to the Fourier transform theory, and dividing a specific sector according to a judgment formula in combination with the current of an induction motor, the rotating speed of a rotor and the angle of the rotor; 2, generating a random number by adopting a two-state Markov chain, dividing the variation range of a random switching frequency into two state spaces, and enabling the switching frequency to be switched immediately in the two state spaces to realize asymmetric SVPWM (Space Vector Pulse Width Modulation); step 3, replacing the original fixed center frequency with a changing sinusoidal signal to form a switching frequency to realize small-range random change in a large range; and step 4, designing upper and lower limits of the switching frequency based on a random value generated by the Markov chain and an interaction relationship between the sinusoidal signal and the fixed carrier frequency. According to the invention, the high-frequency harmonic amplitude can be effectively suppressed, so that the motor has relatively low high-frequency noise.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of motor control application, and designs a control technology for an induction motor, in particular to a pulse width modulation method for an induction motor frequency converter. Background Art

[0002] As people pay more attention to the comfort of electric vehicles and electric drive noise, the low noise of electric drive systems has become one of the important indicators of the electric vehicle industry. Electric vehicles use induction motors as electric drive systems because of their high reliability, simple maintenance, and low cost, so they are widely used. Electric drive noise mainly comes from the electromagnetic vibration and noise of the induction motor. Therefore, reducing the electromagnetic vibration and noise of the induction motor has important engineering practical value and theoretical research significance. The induction motor used in electric vehicles is controlled by a frequency converter. The electromagnetic design of the motor body and the control strategy of the frequency converter are the key factors that determine the electromagnetic vibration and noise of the induction motor.

[0003] The electromagnetic noise and vibration of induction motors mainly come from the following aspects:

[0004] 1. Changes in the electromagnetic field: When the induction motor is running, the modulation voltage provided by the inverter will cause changes in the electromagnetic field inside the motor, resulting in fluctuations in the electromagnetic force, causing vibrations in the motor components, and then generating noise.

[0005] 2. Current harmonics: The PWM signal output by the inverter will introduce high-frequency current harmonics. These harmonic currents will cause non-sinusoidal waveforms of the current inside the motor, thereby generating non-uniformity of the magnetic field and increasing the vibration and noise of the motor.

[0006] 3. Motor structure and mechanical transmission: The internal structure and mechanical transmission system of the induction motor will also affect the vibration and noise, such as bearing friction, fan aerodynamic noise, etc.

[0007] As an important part of electric vehicles, the electric drive system consists of a drive motor and an inverter. Its losses include motor losses and inverter losses, of which the inverter losses consist of conduction losses and switching losses. With the increase of power level and switching frequency, the switching loss of the inverter gradually increases, resulting in reduced system efficiency. The main factor affecting the switching loss is the switching frequency. The higher the switching frequency, the greater the switching loss of the inverter. In order to reduce the switching loss and improve the system efficiency, the switching frequency needs to be reduced, but this will lead to an increase in current ripple, intensified high-frequency vibration, and increased motor noise.

[0008] The core of the induction motor drive system for electric vehicles is the inverter using pulse width modulation (PWM) technology, which is usually composed of a control chip and a power switch device. According to the principle of pulse width modulation, the current output by the inverter contains harmonics of the switching frequency and its sidebands. These harmonics have the characteristics of high frequency and narrow bandwidth, which will cause electromagnetic vibration and narrowband noise in the motor, and narrowband noise makes people feel uncomfortable. Increasing the switching frequency (for example, the switching frequency is higher than 20kHz) can suppress this noise, but it will increase switching losses and is not widely used.

[0009] There are many switching control techniques, such as sinusoidal pulse width modulation (SPWM), space vector pulse width modulation (SVPWM), carrier-based PWM, selective harmonic elimination PWM, harmonic band PWM, etc. SVPWM technology is the best method for variable frequency drive based on voltage source inverter because of its low switching loss and ability to minimize the harmonic output signal generated by the inverter. Most SVPWMs require complex online calculations, which makes real-time implementation difficult. Traditional SVPWM requires additional memory, which limits the choice of switching frequency, thereby reducing the accuracy of SVPWM.

[0010] The existing random SVPWM technology randomly changes the switching signal period, pulse position and other parameters to evenly distribute the harmonic power originally concentrated at the switching frequency and its integer multiples over a wider frequency range, thereby reducing the peak value in the power spectral density (PSD) diagram. It is currently a widely studied method for effectively reducing electromagnetic vibration and noise.

[0011] Random SVPWM technology can be divided into random switching frequency modulation and random pulse position modulation. It mainly studies two aspects: the generation of random constants and the determination of switching period and pulse position. Compared with random pulse position modulation, when random switching frequency modulation acts on the induction motor control system, the PSD of the motor output line voltage is more uniform, and the high-order harmonics are randomly distributed in a wider frequency range. The existing random SVPWM technology can make the original harmonic peaks evenly distributed in a wide frequency range, but it cannot eliminate the high-frequency current harmonics generated in the current of the induction motor that are related to the carrier frequency and are mainly distributed near the integer multiples of the carrier frequency and are relatively concentrated. Summary of the invention

[0012] In view of the above problems, the present invention starts with the inverter control strategy and proposes an induction motor driving method based on random switching frequency pulse width modulation to reduce vibration and noise of the induction motor. Random variables are introduced into the control unit of the inverter to make the switching signal change randomly. According to the statistical communication principle, the power device controlled by the switching signal improves the energy distribution characteristics concentrated in the switching frequency harmonic spectrum, disperses the harmonic energy, and makes it have a wider bandwidth. In order to further reduce the harmonic amplitude, asymmetric SVPWM is combined, Markov chain processes the carrier, and asymmetric SVPWM processes the signal wave. Thereby achieving the purpose of reducing the high-frequency electromagnetic vibration and narrowband noise of the induction motor.

[0013] The present invention adopts the following technical solutions to achieve the above-mentioned invention object:

[0014] An induction motor driving method based on random switching frequency pulse width modulation comprises the following steps:

[0015] 1. First, the switching period calculation formula is derived based on Fourier transform theory. Combined with the current, rotor speed and rotor angle of the induction motor, the specific sectors are divided according to the judgment formula.

[0016] 2. Then a two-state Markov chain is used to generate random numbers, and the range of random switching frequency is divided into two state spaces, so that the switching frequency is switched randomly in the two state spaces. The generated random number is used as the frequency of the inverter in the switching cycle calculation formula. The Markov chain random number generates a variable frequency carrier through a random number generation method, thereby generating a three-phase PWM pulse. By changing the vector order in the second half of the cycle, asymmetric SVPWM is achieved.

[0017] 3. Then, a changing sinusoidal signal is used to replace the original fixed center frequency, so that the switching frequency can be randomly changed in a small range within a large range.

[0018] 4. Then, based on the random value generated by the Markov chain and the interaction between the sinusoidal signal and the fixed carrier frequency, the upper and lower limits of the switching frequency are designed.

[0019] The present invention adopts the above technical solution and has the following beneficial effects:

[0020] (1) A random switching frequency modulation strategy is adopted to reduce the current harmonic amplitude at the switching frequency and its multiples, and finally a harmonic suppression strategy of the current spectrum is implemented to achieve the purpose of suppressing the vibration noise of the induction motor.

[0021] (2) The pseudo-random number algorithm in the random SVPWM modulation strategy of the present invention is replaced by a Markov chain, so that the random value jumps less and the distribution is more uniform; and the fixed center frequency is replaced by a changing sinusoidal signal to form a small range randomness within a large range, thereby realizing an optimized sinusoidal random signal.

[0022] (3) The present invention can effectively suppress the amplitude of high-frequency harmonics, so that the motor has lower high-frequency noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is the distribution diagram of the basic voltage vector of the three-phase inverter.

[0024] Figure 2 This is the topological structure diagram of the three-phase inverter.

[0025] Figure 3 It is the state transition diagram of the two-state Markov chain.

[0026] Figure 4 Flowchart of random number generation for a two-state Markov chain.

[0027] Figure 5 This is the state diagram of the three-phase switch action when the voltage vector is in the first sector.

[0028] Figure 6 Block diagram for random carrier frequency implementation.

[0029] Figure 7 Block diagram of induction motor vector control based on random switching frequency SVPWM DETAILED DESCRIPTION

[0030] The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0031] Firstly, the switching period calculation formula is derived from Fourier transform theory. Combined with the current, rotor speed and rotor angle of the induction motor, the specific sectors are divided according to the judgment formula.

[0032] The output parameters of the induction motor, stator current, rotor speed and rotor angle, are transformed into the reference voltage vector U ref ; Then, the specific sector where the induction motor output voltage vector is located is determined, such as Figure 1 As shown, the obtained 8 basic voltage vectors divide the complex plane into 6 sectors.

[0033] Figure 2The topological structure diagram of the three-phase inverter circuit is shown in Figure 1. The three bridge arms of the three-phase inverter have a total of 6 switch tubes. When the power device of the upper bridge arm is turned on, the corresponding power device of the lower bridge arm is turned off, and vice versa. Each switch device has two states, and the inverter bridge can be combined into 8 switch states, corresponding to 8 basic voltage vectors, including 6 non-zero vectors V1 (001), V2 (010), V3 (011), V4 (100), V5 (101), V6 (110) and two zero vectors V0 (000) and V7 (111). Taking the first sector as an example, when the reference voltage vector belongs to the first sector, the basic voltage vectors acting are V4 (100) and V6 (110).

[0034] The pulse voltage signal that turns the power devices on and off in the inverter is generated by comparing a triangular carrier wave with a fixed frequency with the modulation wave. After modulation, a pulse voltage signal with a fixed frequency is output, which in turn generates high-frequency current harmonics in the motor current that are related to the carrier frequency and are relatively concentrated, mainly distributed near integer multiples of the carrier frequency.

[0035] The switching frequency and the conduction position of the power device in the SVPWM modulation can affect the voltage in the input motor and the harmonic energy distribution of the motor output current. Therefore, in order to reduce the high-frequency switching noise of the motor, the random SVPWM modulation technology is adopted, which can disperse the concentrated high-frequency current harmonic energy into a wider spectrum, thereby reducing the current harmonic amplitude and achieving the purpose of suppressing the high-frequency vibration noise of the motor caused by the inverter power supply.

[0036] The switching frequency is randomized by randomly changing the frequency of the triangular carrier. In the inverter using the random pulse width modulation strategy, the frequency of the triangular carrier is

[0037] f c =f c0 +R i Df(1)

[0038] Among them, f c Indicates the instantaneous value of the switching frequency, f c0 represents the switching center frequency, i.e. the average value of the switching frequency. Δf represents the change in the switching frequency, which determines the upper and lower limits of the random switching frequency. R i Represents a random number distributed in the range [-1, +1]. Generated by the linear congruential method

[0039]

[0040] Among them, R n and R n+1 Represents the random numbers generated for the nth and (n+1)th times respectively. s Represents the number of random digits, a and b are two prime numbers.

[0041] The harmonic distribution of the inverter output voltage and current spectrum depends on the random switching frequency f c , f c The wider the range of change, the wider and more uniform the distribution of the voltage and current spectrum. However, it should also be noted that f c The upper frequency limit is limited by the performance of the power switching device and the switching loss. c If the lower limit of is too low, the ripple of the output voltage and current will be too large, which will affect the quality of the voltage and current waveforms. The random numbers generated consecutively are often distributed on the same side of the mathematical expectation, and the distribution is poor, which makes the harmonic distribution effect in the spectrum poor. Asymmetric SVPWM is achieved by changing the vector order of the second half cycle.

[0042] Furthermore, a two-state Markov chain is used to generate random numbers, and the range of random switching frequency is divided into two state spaces, so that the switching frequency is switched randomly in the two state spaces. The generated random number is used as the frequency of the inverter in the switching period calculation formula. The Markov chain random number generates a variable frequency carrier through a random number generation method, thereby generating a three-phase PWM pulse.

[0043] A random process {x(t), t∈T}, where the time set T={0,1,2,...}, satisfies:

[0044] P(x n ,t n |x n-1 ,t n ;x n-2 ,t n-2 ; ... ; x1, t1) = P(x n ,t n |x n-1 ,t n-1 )(3)

[0045] Where n = 0, 1, ..., x n Only with x n-1 The random process is called a Markov process.

[0046] The Markov process has no aftereffect. At the next moment t n+1 The state x n+1 Only with the current time t n Status x n The process is called a discrete Markov chain.

[0047] The Markov chain is composed of three parts: state space, transition probability and transition probability matrix. n A random sequence, each unit variable x i has many different possible values. i The set of all possible values ​​of x is called the “state space”. n The value represents the value at time t n The transition probability refers to the conditional probability of using the value at the previous moment to obtain the value at the current moment. The conditional probability model is used to describe the Markov chain.

[0048] P ij (m) = P{x m+1 =j|x m =i},i=1,2,...;j=1,2,...(4)

[0049] Indicates that at time m the system is in state x i With probability, at time m+1 the system is in state x j The conditional probability of . All the conditional probabilities form the transition probability matrix. It is a random matrix, and each random element of the transition probability matrix is ​​greater than or equal to 0, and the sum of the probabilities of any row is equal to 1.

[0050] The most basic of Markov chain is the switching of two states. The two-state Markov chain is introduced into the random number generation link of random switching frequency SVPWM. c The random switching frequency of the two-state Markov chain with the center frequency is the random number generation method using the two-state Markov chain algorithm. The switching frequency of formula (1) can be expressed as

[0051] f c =f c0 +S i ×Df(5)

[0052] Among them, S i It is a random value based on the Markov chain algorithm, and Δf is a fixed frequency band.

[0053] Considering the distribution of random switching frequency, two states are introduced, and S = S i gDf, the random switching frequency variation range [f c -S,f c +S] is divided into [f c -S,f c ] and [f c ,f c +S] two state spaces, corresponding to state 1 and state 2 respectively, so that the switching frequency is switched randomly in the two state spaces.

[0054] Draw the state transition diagram of the two-state Markov chain based on the one-step transfer matrix as follows Figure 3 The specific implementation idea is: when the carrier frequency changes randomly, its mathematical expectation is f c , divide the possible frequency range of the carrier into c and less than f c Two parts, one of which is less than f c The frequency is recorded as state 1, which is greater than f c The frequency is recorded as state 2. The purpose of introducing the Markov chain is to make the frequencies of two adjacent carriers switch between state 1 and state 2 as much as possible, reducing the adverse effects caused by the appearance of multiple frequencies on the desired side. Assuming that the system is currently in the frequency range of state 1, the probability of transferring to state 2 at the next moment is P1, and the probability of still being in state 1 at the next moment is 1-P1; the system is currently in the frequency range of state 2, the probability of transferring to state 1 at the next moment is P1, and the probability of still being in state 2 at the next moment is 1-P1. At this time, the corresponding one-step transfer matrix can be written as:

[0055]

[0056] In the formula, P1 and P2 are both greater than zero, and P1+P2=1.

[0057] Select a=29, b=37 and a=97, b=59 as prime numbers, and generate two sets of random numbers r1 and r2. Define the random number generated by the transition probability judgment as r1. The random number of the random switching frequency is r2. The present invention selects the transition probability P1=0.8, and the two-state Markov chain random number generation flow chart is as follows: Figure 4 As shown, the process requires switching between two sets of random numbers.

[0058] The theoretical random number range is 0.0 to 1.0, and the theoretical mathematical expectation value is 0.5. Compared with conventional random, Markov chain random numbers avoid being distributed near the theoretical mathematical expectation, which corresponds to integer multiples of the switching frequency. It is more inclined to disperse harmonics throughout the spectrum and avoid concentration near integer multiples of the switching frequency.

[0059] Asymmetric SVPWM reduces harmonic amplitude by increasing equivalent switching frequency, so the present invention realizes asymmetric SVPWM by adjusting the order of active vectors within a half cycle. The switching order of asymmetric SVPWM in each sector is shown in Table 1.

[0060] Taking the first sector as an example, the switching state of the asymmetric SVPWM in one cycle is as follows: Figure 5As shown in the figure, the vector sequence is V0(000)→V4(100)→V6(110)→V7(111)→V7(111)→V4(100)→V6(110)→V0(000). In the first half of a carrier cycle, the switching sequence of asymmetric SVPWM and conventional SVPWM is exactly the same. But in the next half cycle, the vector sequence of conventional SVPWM changes from V6→V4 to V4→V6. The change of switching state is reflected in the driving pulse. The middle phase bridge arm moves 4 times, and the other two phase bridge arms move 2 times. The total number of switching times in one cycle is 8 times. Random modulation based on Markov chain is to randomize the carrier frequency, thereby changing the distribution of harmonics on the spectrum determined by the carrier frequency.

[0061] Table 1 Vector order of asymmetric SVPWM in each sector

[0062] Vector The corresponding bridge arm Asymmetric SVPWM I(0≤θ<60°) <![CDATA[S a ]]> 0-4-6-7-7-4-6-0 II(60°≤θ<120°) <![CDATA[S b ]]> 0-2-6-7-7-2-6-0 III(120°≤θ<180°) <![CDATA[S c ]]> 0-2-3-7-7-2-3-0 IV(180°≤θ<240°) <![CDATA[S a ]]> 0-1-3-7-7-1-3-0 V(240°≤θ<300°) <![CDATA[S b ]]> 0-1-5-7-7-1-5-0 VI(300°≤θ<360°) <![CDATA[S c ]]> 0-4-5-7-7-4-5-0

[0063] Furthermore, a changing sinusoidal signal is used to replace the original fixed center frequency, so that the switching frequency can be randomly changed in a small range within a large range.

[0064] The center frequency of the carrier in the random switching frequency modulation technology is fixed. The Gaussian random signal is obtained by combining the random value generated by the pseudo-random number algorithm. The random performance of the random switching frequency is limited and the distribution is uneven. As a basic signal with obvious characteristics, the sine signal uses its periodic characteristics and is combined with the random value generated by the Markov chain algorithm mentioned above to achieve the coexistence of randomness and periodicity of the switching frequency of SVPWM modulation. The random signal is transformed from a Gaussian random signal to an optimized sinusoidal random signal, forming a small-range randomness within a large range.

[0065] The principle of realizing the random switching frequency centered on the sinusoidal signal is that during the action of the sinusoidal signal, the random value generated by the Markov chain algorithm changes with the sinusoidal signal. Taking a sinusoidal signal cycle as an example, the switching frequency first increases and then decreases in the positive half-cycle of the sinusoidal signal, and first decreases and then increases in the negative half-cycle. Such a large-scale cyclic change, that is, within one cycle of the sinusoidal signal, the Markov random value realizes four large-scale changes on its basis.

[0066] The expected center frequency can be expressed as the sum of the sinusoidal signal changing frequency and the wide range changing frequency of the random value within a single sinusoidal cycle. Therefore, the frequency selection of the sinusoidal signal is closely related to the expected center frequency, which is reflected in the proportional relationship that the sinusoidal signal changing frequency accounts for about one-fifth of the expected center frequency, and the frequency of the sinusoidal signal is positively correlated with the expected center frequency.

[0067] The implementation block diagram of random carrier frequency is as follows: Figure 6As shown, the expansion ratio of the random SVPWM modulation is defined as

[0068]

[0069] where f is the frequency of the sine signal; f c is the center frequency of the SVPWM modulation, and the range of m is 0 < m ≤ 1, and it expands within the linear region.

[0070] The expression of the random carrier frequency of the expanded random switching frequency pulse width modulation is

[0071] f s ' = [k1 + k2 sin(ωt)] + S i ·Δf s (8)

[0072] where sin(ωt) is the sine signal, where ω = 2πf; k1 and k2 are non-zero natural numbers, and Df s is the fixed frequency band.

[0073] Furthermore, based on the interaction relationship between the random numerical values generated by the Markov chain, the sine signal, and the fixed carrier frequency, the upper and lower limits of the switching frequency are designed.

[0074] The carrier frequency of the random modulation contains a sine signal, and the amplitude and frequency of the sine signal interact with the fixed carrier frequency band Δf s , so there are slight differences in designing the upper and lower limits of the switching frequency. The random numerical part uses the Markov chain algorithm to ensure the good random performance of the random numerical value S i , and it is necessary to limit the upper and lower limits of the switching frequency.

[0075] The random carrier frequency f s ' has an upper limit frequency f s ' max and a lower limit frequency f s ' min . When S i = -1 and the sine signal is at the peak of the negative half cycle, the lower limit of the random carrier frequency is reached, that is, f s ' min = k1 - k2 - Df s . In the current frequency spectrum, if there are no obvious harmonics in the range where the harmonic at the frequency f h min is separated from the fundamental wave f e by Df ε , then f h min = f e + Df ε . In the current harmonic spectrum, the harmonic closest to the fundamental wave is near the first harmonic of the carrier frequency, so take (f c-4f e ) is more reasonable, that is

[0076] f h min =f e +Df ε £f c -4f e (9)

[0077] When f h min When the minimum value is taken, f s ' min =f c , combined with formula (9), the lower limit of the random carrier frequency is obtained as

[0078] f s ' min 35f e +Df ε (10)

[0079] When S i =1, and when the sinusoidal signal is at the peak of the positive half cycle, the upper limit of the random carrier frequency is reached, i.e., f s ' max =k1+k2+Df s In the current spectrum, the harmonic peak is obvious and the harmonic farthest from the fundamental wave is located at the sideband frequency of the triple of the carrier frequency, so 93f is taken. c +4f e ) is more reasonable, that is

[0080] f h max 33f c +4f e (11)

[0081] When f h max When the minimum value is taken, f s ' max =f c , combined with formula (11), the upper limit of the random carrier frequency is

[0082]

[0083] In the modulation method, the carrier frequency is a combination of a sinusoidal signal and a random frequency band, where the amplitude of the sinusoidal signal contributes part of the carrier frequency, so the fixed frequency band Δf s The value range is smaller than that of traditional modulation methods and the performance is better.

[0084] According to the above, the vector control system of the induction motor based on random switching frequency pulse width modulation constructed by the present invention is as follows Figure 7 The overall double closed-loop structure is adopted, and the current acquisition module collects the three-phase current of the motor. a 、ib 、i c , the three-phase current is transformed by Clark and Park to obtain the current i in the two-phase rotating coordinate system d 、i q At the same time, the speed signal on the other side is collected and PI adjusted with the given speed, and then d 、i q Together they generate a voltage U d , U q After Park transformation, the input is sent to the pulse width modulation module to generate SVPWM wave, which controls the three-phase inverter to generate three-phase voltage and sends it to the induction motor to complete the control of the induction motor.

[0085] The pulse width modulation based on random switching frequency proposed by the present invention randomizes the carrier frequency, thereby changing the distribution of harmonics on the spectrum determined by the carrier frequency. The Markov chain has no aftereffect and independence of the state space, so that it obeys certain rules that can produce better parameter values, making the random values ​​generated by the Markov algorithm more uniform and having better random performance than those generated by the formula method. The Markov chain random number generates a variable frequency carrier through a random number generation method, thereby generating a three-phase PWM pulse. By adjusting the vector order of the second half cycle, asymmetric SVPWM can be realized. The Markov chain processes the carrier, and the asymmetric SVPWM processes the signal wave. The two algorithms are independent and uncoupled. According to the excellent random performance of the Markov chain algorithm, the random performance of the random SVPWM switching frequency can be made more ideal, thereby achieving the effect of avoiding the drawbacks of large random value jumps and uneven values ​​in the traditional pseudo-random number algorithm, and the random values ​​generated can be evenly distributed near the mathematical expectation. Avoiding concentration near the integer multiple switching frequency, the pulse width modulation module generates an SVPWM wave to control the inverter output current, thereby dynamically controlling the induction motor.

[0086] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention's specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An induction motor driving method based on random switching frequency pulse width modulation, characterized in that: The following steps are involved: Step 1. First, the switching cycle calculation formula is derived based on Fourier transform theory, and the specific sectors are divided according to the judgment formula based on the current, rotor speed and rotor angle of the induction motor; Step 2. Then use a two-state Markov chain to generate random numbers, divide the range of random switching frequency into two state spaces, so that the switching frequency is switched randomly in the two state spaces, and use the generated random numbers as the switching period calculation formula to adjust the inverter frequency. The Markov chain random number generates a variable frequency carrier through a random number generation method, thereby generating a three-phase PWM pulse, and realizing asymmetric SVPWM by changing the vector order of the second half cycle; Step 3. Then, a changing sinusoidal signal is used to replace the original fixed center frequency, so that the switching frequency can be randomly changed in a small range within a large range; Step 4. Then, based on the random value generated by the Markov chain and the interaction between the sinusoidal signal and the fixed carrier frequency, the upper and lower limits of the switching frequency are designed.

2. The induction motor driving method based on random switching frequency pulse width modulation according to claim 1, characterized in that: In step 1, the output parameters of the induction motor, the stator current, the rotor speed and the rotor angle, are transformed into a reference voltage vector U ref ; Then, the specific sector where the induction motor output voltage vector is located is determined, and 8 basic voltage vectors are obtained: including 6 non-zero vectors V1 (001), V2 (010), V3 (011), V4 (100), V5 (101), V6 (110) and two zero vectors V0 (000), V7 (111); the complex plane is divided into 6 sectors; The random SVPWM modulation technology is used, which can disperse the concentrated high-frequency current harmonic energy into a wider spectrum. The switching frequency is randomized by randomly changing the frequency of the triangular carrier. In the inverter using the random pulse width modulation strategy, the frequency of the triangular carrier is f c =f c0 +R i Δf (1) Among them, f c Indicates the instantaneous value of the switching frequency, f c0 represents the switching center frequency, i.e. the average value of the switching frequency. Δf represents the change in the switching frequency, which determines the upper and lower limits of the random switching frequency. R i Represents a random number distributed in the range [-1, +1]. Generated by the linear congruential method Among them, R n and R n+1 Respectively represent the random numbers generated for the nth and (n+1)th times, N s Represents the number of random digits, a and b are two prime numbers.

3. The induction motor driving method based on random switching frequency pulse width modulation according to claim 1, characterized in that: The specific process of step 2 is: A random process {x(t), t∈T}, where the time set T={0,1,2,...}, satisfies: P(x n ,t n |x n-1 ,t n ;x n-2 ,t n-2 ;...;x1,t1)=P(x n ,t n |x n-1 ,t n-1 ) (3) Where n = 0, 1, ..., x n Only with x n-1 About n is the nth moment; then the random process is called a Markov process; The Markov process has no aftereffect. At the next moment t n+1 The state x n+1 Only with the current time t n Status x n It is related to the state of the next moment, but not to the state of the past moment. It continuously generates the state of the next moment based on the current state to obtain random values ​​that are independent of each other and have better performance. When the independent variable t and state space {x} of the Markov process are both discrete values, the process is called a discrete Markov chain. The Markov chain is composed of three parts: state space, transition probability and transition probability matrix; where the Markov chain is x1,x2,x3,…,x n A random sequence, each unit variable x i has many different possible values; x i The set of all possible values ​​of is called the "state space", n The value represents the value at time t n The transition probability refers to the conditional probability of using the value at the previous moment to obtain the value at the current moment. The conditional probability model is used to describe the Markov chain. P ij (m)=P{x m+1 =j|x m =i},i=1,2,...;j=1,2,... (4) Indicates that at time m the system is in state x i With probability, at time m+1 the system is in state x j The conditional probability of , where all conditional probabilities form a transition probability matrix; it is a random matrix, and each random element of the transition probability matrix is ​​greater than or equal to 0, and the sum of the probabilities of any row is equal to 1; The most basic of Markov chain is the switching of two states. The two-state Markov chain is introduced into the random number generation link of random switching frequency SVPWM. c The random switching frequency of the two-state Markov chain with the center frequency is the random number generation method using the two-state Markov chain algorithm. The switching frequency of formula (1) can be expressed as f c =f c0 +S i ·Δf (5) Among them, S i is a random value based on the Markov chain algorithm, Δf is a fixed frequency band; f c0 Indicates the switching center frequency, that is, the average switching frequency. Considering the distribution of random switching frequency, two states are introduced, and S = S i ·Δf, the random switching frequency variation range [f c -S,f c +S] is divided into [f c -S,f c ] and [f c ,f c +S] two state spaces, corresponding to state 1 and state 2 respectively, so that the switching frequency is switched randomly in the two state spaces; The state transition diagram of the two-state Markov chain is drawn according to the one-step transfer matrix. The specific implementation idea is: when the carrier frequency changes randomly, its mathematical expectation is f c , divide the possible frequency range of the carrier into c and less than f c Two parts, one of which is less than f c The frequency is recorded as state 1, which is greater than f c The frequency is recorded as state 2. The purpose of introducing the Markov chain is to make the frequencies of two adjacent carriers switch between state 1 and state 2 as much as possible, reducing the adverse effects caused by the appearance of multiple frequencies on the desired side; assuming that the system is currently in the frequency range of state 1, the probability of transferring to state 2 at the next moment is P1, and the probability of still being in state 1 at the next moment is 1-P1; the system is currently in the frequency range of state 2, the probability of transferring to state 1 at the next moment is P1, and the probability of still being in state 2 at the next moment is 1-P1. At this time, the corresponding one-step transfer matrix can be written as: Wherein, P1 and P2 are both greater than zero, and P1+P2=1; Asymmetric SVPWM reduces the harmonic amplitude by increasing the equivalent switching frequency. Therefore, the present invention realizes asymmetric SVPWM by adjusting the order of active vectors within half a cycle. Taking the first sector as an example, the switching state of the asymmetric SVPWM within a cycle, the vector order is V0(000)→V4(100)→V6(110)→V7(111)→V7(111)→V4(100)→V6(110)→V0(000). In the first half of a carrier cycle, the switching order of the asymmetric SVPWM is exactly the same as that of the conventional one, but in the next half cycle, the vector order of the conventional SVPWM changes from V6→V4 to V4→V6. The change of the switching state is reflected in the driving pulse. The middle phase bridge arm acts 4 times, and the other two phase bridge arms act 2 times. The total number of switching times within a cycle is 8 times. The random modulation based on the Markov chain is to randomize the carrier frequency, thereby changing the distribution of harmonics on the spectrum determined by the carrier frequency.

4. The induction motor driving method based on random switching frequency pulse width modulation according to claim 1, characterized in that: The specific process of step 3 is: The realization principle of random switching frequency centered on the sinusoidal signal is that during the action of the sinusoidal signal, the random value generated by the Markov chain algorithm changes with the sinusoidal signal. If it is a sinusoidal signal cycle, the switching frequency increases first and then decreases in the positive half cycle of the sinusoidal signal, and decreases first and then increases in the negative half cycle. In such a large-scale cyclic change, that is, within one cycle of the sinusoidal signal, the Markov random value realizes four large-scale changes on its basis; The expected center frequency can be expressed as the sum of the changing frequency of the sinusoidal signal and the wide-range changing frequency of the random value within a single sinusoidal cycle. Therefore, the frequency selection of the sinusoidal signal is closely related to the expected center frequency. The proportional relationship is that the changing frequency of the sinusoidal signal accounts for about one-fifth of the expected center frequency, and the frequency of the sinusoidal signal is positively correlated with the expected center frequency. The expansion ratio of random SVPWM modulation is defined as: where f is the frequency of the sine signal; f c is the center frequency of SVPWM modulation, the range of m is 0 < m ≤ 1, and it is extended within the linear region; The random carrier frequency expression of the extended random switching frequency pulse width modulation is: f’ s =[k1+k2sin(ωt)]+S i ·Δf s (8) Wherein, sin(ωt) is a sinusoidal signal, where ω=2πf; k1 and k2 are non-zero natural numbers, Δf s It is a fixed frequency band.

5. The induction motor driving method based on random switching frequency pulse width modulation according to claim 4, characterized in that: The specific process of step 4 is: The randomly modulated carrier frequency contains a sinusoidal signal, the amplitude and frequency of the sinusoidal signal are in the same range as the fixed carrier frequency band Δf s There is interaction between them, so the upper and lower limits of the switching frequency are slightly different when designing. The random value part uses the Markov chain algorithm to ensure the random value S i Good random performance requires limiting the upper and lower limits of the switching frequency; Random carrier frequency f s 'has an upper frequency f' smax and lower frequency f' smin A random sequence, when S i = -1, and when the sinusoidal signal is at the peak of the negative half cycle, the lower limit of the random carrier frequency is reached, i.e. f' smin =k1-k2-Δf s , in the current spectrum, if at frequency f hmin The harmonics and fundamental wave f e Distance Δf ε If there is no obvious harmonic in the range of hmin =f e +Δf ε In the current harmonic spectrum, the harmonic closest to the fundamental wave is located near one times the carrier frequency, so (f c -4f e ) is more reasonable, that is f hmin =f e +Δf ε ≤f c -4f e (9) When f hmin When the minimum value is reached, f' smin =f c , combined with formula (9), the lower limit of the random carrier frequency is obtained as f’ smin ≥5f e +Δf ε (10) When S i =1, and when the sinusoidal signal is at the peak of the positive half cycle, the upper limit of the random carrier frequency is reached, i.e. f' smax =k1+k2+Δf s In the current spectrum, the harmonic peak is obvious and the harmonic farthest from the fundamental wave is located at the sideband frequency of the triple of the carrier frequency, so (3f c +4f e ) is more reasonable, that is f hmax ≥3f c +4f e (11) When f hmax When the minimum value is reached, f' smax =f c , combined with formula (11), the upper limit of the random carrier frequency is In the modulation method, the carrier frequency is a combination of a sinusoidal signal and a random frequency band, where the amplitude of the sinusoidal signal contributes part of the carrier frequency, so the fixed frequency band Δf s The value range is smaller than that of traditional modulation methods and the performance is better.

Citation Information

Cited By

  • Switch noise suppression method for motor controller

    CN119945227A