SVPWM dead zone compensation method and system based on artificial intelligence

Abstract

The present invention relates to the field of inverter control technology, and in particular to an SVPWM dead zone compensation method and system based on artificial intelligence. The method comprises acquiring first data; inputting a rated dead zone time into an SVPWM controller, starting a three-phase six-bridge-arm inverter, driving a load motor to run, collecting a second harmonic voltage amplitude sequence to a first harmonic voltage amplitude sequence; N Harmonic voltage amplitude sequence; establish a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, recorded as the first simulation model; construct a first objective function and a first constraint condition, and use a particle swarm optimization algorithm to calculate the optimal opening time compensation value; calculate the optimal dead time; input the optimal dead time into the SVPWM controller. The present invention can reduce the size of the harmonic voltage at the input end of the load motor while ensuring the dead time compensation effect by optimizing the dead time of the SVPWM controller.

Description

Technical Field

[0001] The present invention relates to the technical field of inverter control, and in particular to an SVPWM dead zone compensation method and system based on artificial intelligence. Background Art

[0002] The three-phase six-bridge-arm inverter drives the load motor to achieve variable frequency speed regulation and is widely used in the field of industrial production. As an excellent control strategy, SVPWM (space vector pulse width modulation) is widely adopted because of its advantages such as high DC bus voltage utilization and low output harmonic content. However, in actual engineering applications, in order to prevent the upper and lower IGBT switches of the three-phase six-bridge arm of the inverter from being turned on at the same time and causing a short circuit, a dead time must be introduced between the complementary switches. The existence of this dead time will cause the inverter output voltage to be distorted, causing harmonics in the input voltage of the load motor. In situations such as textiles that require precise control, the harmonics caused by the dead time effect will affect the motor speed, which will significantly affect the product quality.

[0003] At present, the commonly used dead zone compensation method is the fixed compensation method. The fixed compensation method superimposes a fixed compensation time on the established dead zone time, thereby offsetting some of the adverse effects of the dead zone time and preventing the inverter switch tube from short-circuiting due to insufficient dead zone time. Although this method is simple to implement, the fixed compensation time can only be obtained through theoretical calculations or based on practical experience, which is difficult to adapt to the complex and changeable working conditions on site, resulting in poor compensation effect. In recent years, with the development of digital twin and artificial intelligence technology, combining digital twin modeling with measured data to optimize the dead zone time and thereby reduce the harmonics caused by the dead zone has become a feasible technical means to solve this problem. Summary of the invention

[0004] (1) Technical issues to be resolved

[0005] The purpose of the present invention is to provide an SVPWM dead zone compensation method and system based on artificial intelligence, so as to achieve a three-phase six-bridge-arm inverter driving a load motor while ensuring the dead zone compensation effect and reducing the magnitude of the harmonic voltage at the input end of the load motor.

[0006] (2) Technical solution

[0007] To achieve the above object, the present invention provides an SVPWM dead zone compensation method based on artificial intelligence, the method comprising the following steps:

[0008] S1, obtaining first data, wherein the first data includes a preset rated dead time, a first sampling time interval, a first interval time, a second interval time, and a truncated harmonic order N.

[0009] S2, input the rated dead time into the SVPWM controller, start the three-phase six-bridge arm inverter, and drive the load motor to run; within the first sampling time interval, use the first interval time as the sampling time interval to collect the input terminal voltage of the load motor, and perform discrete Fourier transform with the second interval time as the time window according to the truncated harmonic order to obtain the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence.

[0010] S3, establish a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, input a preset simulation dead time sequence, and run to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence; use a nonlinear fitting algorithm to fit the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence, recorded as the second function to the Nth function; construct a first objective function and a first constraint condition according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and use a particle swarm optimization algorithm to calculate the optimal opening time compensation value.

[0011] S4, calculating the optimal dead time according to the rated dead time and the optimal on-time compensation value; inputting the optimal dead time into the SVPWM controller.

[0012] Furthermore, the method of establishing a simulation model of a three-phase six-bridge-arm inverter and a load motor by using a digital twin algorithm, inputting a preset simulation dead time sequence, and running to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence includes:

[0013] According to the circuit connection topology of the three-phase six-bridge-arm inverter and the load motor, a simulation model of the three-phase six-bridge-arm inverter and the load motor is established in the simulation platform, which is recorded as the first simulation model; according to the SVPWM controller, an SVPWM controller simulation model is established; and the SVPWM controller simulation model is embedded in the first simulation model.

[0014] The preset simulation dead time sequence is input into the SVPWM controller simulation model respectively to obtain the first simulation scene to the Mth simulation scene; wherein M represents the number of data in the simulation dead time sequence; the first simulation scene to the Mth simulation scene are set into the first simulation model, and the first simulation model is run; the second harmonic component to the Nth harmonic component of the input terminal voltage of the load motor under the first simulation scene to the Mth simulation scene are extracted to obtain the simulation harmonic component matrix ; The simulation harmonic component matrix It is expressed as:

[0015] ;

[0016] in, Indicates The voltage at the input terminal of the load motor in the simulation scenario Subharmonic components; The value range is 1 to integer variable of ; The value is 2 to An integer variable.

[0017] The simulation harmonic component matrix Split into second harmonic simulation sequence to Nth harmonic simulation sequence by column.

[0018] Furthermore, the method of fitting the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence using a nonlinear fitting algorithm, recorded as the second function to the Nth function, includes:

[0019] Establish the second harmonic simulation sequence to the Nth harmonic simulation sequence from the first data to the The data are the same as the first to the second data in the simulation dead zone time series. The mapping relationship of the data.

[0020] The simulation dead zone time series is from the first data to the The data are taken as independent variables, and the first data to the Nth data in the second harmonic simulation sequence to the Nth harmonic simulation sequence are taken as independent variables. The data are taken as dependent variables, and the nonlinear least squares method is used to fit the functional relationship between the second harmonic simulation value to the Nth harmonic simulation value and the simulation dead time, which are recorded as the second function to the Nth function respectively.

[0021] Furthermore, the method of constructing a first objective function and a first constraint condition according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and using a particle swarm optimization algorithm to calculate the optimal opening time compensation value includes:

[0022] A first objective function is constructed according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and the first objective function is expressed as:

[0023] ;

[0024] in, represents the first objective function; Indicates the opening time compensation value in seconds; Indicates the second interval time in seconds; Indicates the first sampling time interval, in seconds; The value is 2 to integer variable of ; The value range is 1 to integer variable of ; For the The first in the harmonic voltage amplitude sequence Data, in volts; Indicates function; Indicates the simulation dead time in seconds; Indicates the rated dead time in seconds.

[0025] Construct a first constraint condition, which is expressed as:

[0026] ;

[0027] in, Indicates the preset lower limit of the dead zone, in seconds; Indicates the preset upper limit of the dead zone, in seconds.

[0028] Taking the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value.

[0029] Further, the method for calculating the optimal dead time according to the rated dead time and the optimal opening time compensation value includes:

[0030] The optimal dead time is calculated according to the rated dead time and the optimal opening time compensation value. The calculation formula of the optimal dead time is:

[0031] ;

[0032] in, represents the optimal opening time compensation value, Indicates the optimal dead time.

[0033] Based on the same inventive concept, on the other hand, the present invention also provides an artificial intelligence-based SVPWM dead zone compensation system, the system comprising:

[0034] The data reading module is used to obtain first data, wherein the first data includes a preset rated dead time, a first sampling time interval, a first interval time, a second interval time, and a truncated harmonic number N.

[0035] The data acquisition module is connected to the data reading module and is used to input the rated dead time into the SVPWM controller, start the three-phase six-bridge arm inverter, and drive the load motor to run; within the first sampling time interval, the first interval time is used as the sampling time interval, the input terminal voltage of the load motor is collected, and according to the truncated harmonic number, the second interval time is used as the time window for discrete Fourier transform, and the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence are obtained.

[0036] The optimization calculation module is connected to the data acquisition module and is used to establish a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, input a preset simulation dead time sequence, and run to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence; a nonlinear fitting algorithm is used to fit the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence, which are recorded as the second function to the Nth function; a first objective function and a first constraint condition are constructed according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and the optimal opening time compensation value is calculated using a particle swarm optimization algorithm.

[0037] The result output module is connected to the optimization calculation module and is used to calculate the optimal dead time according to the rated dead time and the optimal opening time compensation value; and input the optimal dead time into the SVPWM controller.

[0038] Furthermore, the optimization calculation module includes:

[0039] A simulation model building module is used to establish a simulation model of a three-phase six-bridge-arm inverter and a load motor in a simulation platform according to a circuit connection topology structure of the three-phase six-bridge-arm inverter and the load motor, which is recorded as a first simulation model; establish an SVPWM controller simulation model according to the SVPWM controller; and embed the SVPWM controller simulation model into the first simulation model.

[0040] The simulation calculation module is connected to the simulation model construction module, and is used to input the preset simulation dead time sequence into the SVPWM controller simulation model respectively to obtain the first simulation scene to the Mth simulation scene; wherein M represents the number of data in the simulation dead time sequence; the first simulation scene to the Mth simulation scene are set in the first simulation model, and the first simulation model is run; the second harmonic component to the Nth harmonic component of the load motor input terminal voltage under the first simulation scene to the Mth simulation scene are extracted to obtain the simulation harmonic component matrix ; The simulation harmonic component matrix It is expressed as:

[0041] ;

[0042] in, Indicates The voltage at the input terminal of the load motor in the simulation scenario Subharmonic components; The value range is 1 to integer variable of ; The value is 2 to An integer variable.

[0043] A harmonic simulation sequence building module is connected to the simulation calculation module and is used to convert the simulation harmonic component matrix Split into second harmonic simulation sequence to Nth harmonic simulation sequence by column.

[0044] Furthermore, the optimization calculation module also includes:

[0045] The data mapping module is used to establish the first data to the second harmonic simulation sequence to the Nth harmonic simulation sequence. The data are the same as the first to the second data in the simulation dead zone time series. The mapping relationship of the data.

[0046] The function fitting module is connected to the data mapping module and is used to simulate the dead zone time series from the first data to the The data are taken as independent variables, and the first data to the Nth data in the second harmonic simulation sequence to the Nth harmonic simulation sequence are taken as independent variables. The data are taken as dependent variables, and the nonlinear least squares method is used to fit the functional relationship between the second harmonic simulation value to the Nth harmonic simulation value and the simulation dead time, which are recorded as the second function to the Nth function respectively.

[0047] Furthermore, the optimization calculation module also includes:

[0048] The first objective function construction module is used to construct a first objective function according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, wherein the first objective function is expressed as:

[0049] ;

[0050] in, represents the first objective function; Indicates the opening time compensation value in seconds; Indicates the second interval time in seconds; Indicates the first sampling time interval, in seconds; The value is 2 to integer variable of ; The value range is 1 to integer variable of ; For the The first in the harmonic voltage amplitude sequence Data, in volts; Indicates function; Indicates the simulation dead time in seconds; Indicates the rated dead time in seconds.

[0051] The first constraint condition building module is connected to the first objective function building module and is used to construct a first constraint condition, wherein the first constraint condition is expressed as:

[0052] ;

[0053] in, Indicates the preset lower limit of the dead zone, in seconds; Indicates the preset upper limit of the dead zone, in seconds.

[0054] Taking the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value.

[0055] Furthermore, the result output module includes:

[0056] The optimal dead time calculation module is used to calculate the optimal dead time according to the rated dead time and the optimal opening time compensation value. The calculation formula of the optimal dead time is:

[0057] ;

[0058] in, represents the optimal opening time compensation value, Indicates the optimal dead time.

[0059] (3) Beneficial effects

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The voltage at the input end of the load motor is collected, and the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence is calculated by discrete Fourier transform, thereby reflecting the real harmonic level of the load motor driven by the three-phase six-bridge arm inverter. Then, the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence is obtained by simulation, which is recorded as the second function to the Nth function. The theoretical influence of a small change in the dead time on the harmonic level of the load motor can be obtained by taking the derivative of the second function to the Nth function. The measured data and the simulation data are combined to construct the first objective function and the first constraint condition, and the optimization calculation is performed to obtain the optimal dead time. Since the first objective function reflects the adverse effect of the dead time on the harmonic voltage at the input end of the load motor, and the first constraint condition ensures the dead time compensation effect, the optimal dead time is input into the SVPWM controller, which can reduce the size of the harmonic voltage at the input end of the load motor while ensuring the dead time compensation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 14 is a flowchart of an SVPWM dead zone compensation method based on artificial intelligence according to Embodiment 1 of the present invention;

[0063] Figure 2 This is a schematic diagram of the module composition of the artificial intelligence-based SVPWM dead-zone compensation system of Example 2 of the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Before giving examples, it is necessary to explain the application scenarios of the present invention. The present invention is applied to the control of a three-phase six-bridge-arm inverter driving a load motor. The three-phase six-bridge-arm inverter driving a load motor to achieve variable frequency speed regulation has been widely used in the textile field. The textile feed speed can be accurately controlled by driving the load motor with a three-phase six-bridge-arm inverter. Therefore, the stable operation of the three-phase six-bridge-arm inverter driving the load motor is crucial to product quality. In order to ensure the stable operation of the load motor, the three-phase six-bridge-arm inverter adopts SVPWM control. Through SVPWM control, the switch tube of the three-phase six-bridge-arm inverter is turned on and off in a certain regularity. However, since the switch tube of the three-phase six-bridge-arm inverter often uses an insulated gate bipolar transistor, i.e., IGBT, there is a certain randomness in the time delay from receiving the shutdown signal to completely shutting down the device. Therefore, it is often necessary to input a dead time into the SVPWM controller so that the complementary switch tubes in the three-phase six-bridge-arm inverter are turned on with a delay, thereby avoiding the switch tube that is turned on later before the last one is turned off completely, thereby avoiding short circuit. However, after the dead time is introduced, a new problem arises: since the dead time is often a fixed value set artificially, if it is set improperly, the inverter output voltage will be distorted beyond the tolerance range, causing harmonics to be generated in the voltage at the input end of the load motor, causing the speed of the load motor to fluctuate, and affecting the quality of textile products. This embodiment measures the amplitude of the harmonic voltage at the input end of the load motor, establishes a simulation model of a three-phase six-bridge-arm inverter and a load motor, combines the measured data with the simulation data, performs optimization calculation, compensates and adjusts the rated dead time, and inputs it into the SVPWM controller, thereby reducing the magnitude of the harmonic voltage at the input end of the load motor while ensuring the dead time compensation effect.

[0066] Example 1: Figure 1 As shown, this embodiment provides an SVPWM dead zone compensation method based on artificial intelligence, and the method includes the following steps:

[0067] S1, obtaining first data, wherein the first data includes a preset rated dead time, a first sampling time interval, a first interval time, a second interval time, and a truncated harmonic order N.

[0068] S2, input the rated dead time into the SVPWM controller, start the three-phase six-bridge arm inverter, and drive the load motor to run; within the first sampling time interval, use the first interval time as the sampling time interval to collect the input terminal voltage of the load motor, and perform discrete Fourier transform with the second interval time as the time window according to the truncated harmonic order to obtain the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence.

[0069] S3, establish a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, input a preset simulation dead time sequence, and run to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence; use a nonlinear fitting algorithm to fit the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence, recorded as the second function to the Nth function; construct a first objective function and a first constraint condition according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and use a particle swarm optimization algorithm to calculate the optimal opening time compensation value.

[0070] S4, calculating the optimal dead time according to the rated dead time and the optimal on-time compensation value; inputting the optimal dead time into the SVPWM controller.

[0071] For example, a textile factory uses a three-phase six-leg inverter to drive a load motor with a rated power of 7.5kW, a rated voltage of 380V, and a rated current of 17A. The three-phase six-leg inverter uses an SVPWM controller, and the modulation frequency is set to 8kHz. The first data is obtained, the preset rated dead time is 0.000004 seconds, the first sampling time interval is 300 seconds, the first interval time is 0.000125 seconds, the second interval time is 1 second, and the number of truncated harmonics is 13.

[0072] After the rated dead time of 0.000004 seconds is input into the SVPWM controller, the three-phase six-bridge arm inverter is started to drive the load motor to run. In the first sampling time interval of 300 seconds, the voltage at the input end of the load motor is collected with a sampling time interval of 0.000125 seconds to obtain the first end voltage data sequence. According to the set truncated harmonic order 13, the first end voltage data sequence is subjected to discrete Fourier transform with a time window of 1 second to obtain the second harmonic component to the thirteenth harmonic component of the first end voltage data sequence, which are recorded as the second harmonic voltage amplitude sequence to the thirteenth harmonic voltage amplitude sequence respectively.

[0073] The simulation model of the three-phase six-bridge-arm inverter and the load motor is established by the digital twin algorithm, which is recorded as the first simulation model. The preset simulation dead time sequence of 0.000002 seconds to 0.000006 seconds is input into the first simulation model and run to obtain the second harmonic simulation sequence to the thirteenth harmonic simulation sequence. The functional relationship between the second harmonic simulation sequence to the thirteenth harmonic simulation sequence and the simulation dead time sequence is fitted by the nonlinear fitting algorithm, which are recorded as the second function to the thirteenth function respectively. According to the measured second harmonic voltage amplitude sequence to the thirteenth harmonic voltage amplitude sequence and the fitted second function to the thirteenth function, the first objective function and the first constraint condition are constructed. With the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value of -0.0000008 seconds.

[0074] According to the rated dead time of 0.000004 seconds and the optimal on-time compensation value of -0.0000008 seconds, the optimal dead time is calculated to be 0.0000032 seconds. The optimal dead time of 0.0000032 seconds is input into the SVPWM controller to control the on and off of the six switches in the three-phase six-leg inverter.

[0075] Furthermore, the method of establishing a simulation model of a three-phase six-bridge-arm inverter and a load motor by using a digital twin algorithm, inputting a preset simulation dead time sequence, and running to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence includes:

[0076] According to the circuit connection topology of the three-phase six-bridge-arm inverter and the load motor, a simulation model of the three-phase six-bridge-arm inverter and the load motor is established in the simulation platform, which is recorded as the first simulation model; according to the SVPWM controller, an SVPWM controller simulation model is established; and the SVPWM controller simulation model is embedded in the first simulation model.

[0077] The preset simulation dead time sequence is input into the SVPWM controller simulation model respectively to obtain the first simulation scene to the Mth simulation scene; wherein M represents the number of data in the simulation dead time sequence; the first simulation scene to the Mth simulation scene are set into the first simulation model, and the first simulation model is run; the second harmonic component to the Nth harmonic component of the input terminal voltage of the load motor under the first simulation scene to the Mth simulation scene are extracted to obtain the simulation harmonic component matrix ; The simulation harmonic component matrix It is expressed as:

[0078] ;

[0079] in, Indicates The voltage at the input terminal of the load motor in the simulation scenario Subharmonic components; The value range is 1 to integer variable of ; The value is 2 to An integer variable.

[0080] The simulation harmonic component matrix Split into second harmonic simulation sequence to Nth harmonic simulation sequence by column.

[0081] Exemplarily, in the MATLAB / Simulink simulation platform, a first simulation model is established based on the obtained circuit connection topology of the three-phase six-bridge-arm inverter and the load motor. The load motor adopts a 7.5kW asynchronous motor model with a voltage rating of 380V and a current rating of 17A. At the same time, an SVPWM controller simulation model is established and embedded in the first simulation model, and the modulation frequency of the controller is set to 8kHz.

[0082] The preset simulation dead time sequence starts from 0.000002 seconds and increases to 0.000006 seconds at intervals of 0.0000001 seconds, including a total of 41 different simulation dead times. These 41 different simulation dead times are respectively input into the SVPWM controller simulation model to form 41 different simulation scenarios, and the first simulation scenario to the 41st simulation scenario are obtained. The running time of each simulation scenario is set to 10 seconds, and the sampling time step is 0.000125 seconds. The first simulation scenario to the 41st simulation scenario are sequentially set to the first simulation model for operation, and the input voltage data of the load motor under the first simulation scenario to the 41st simulation scenario are collected. The input voltage data of the load motor under the first simulation scenario to the 41st simulation scenario are subjected to discrete Fourier transform, so as to extract the second harmonic component to the thirteenth harmonic component, and obtain a simulation harmonic component matrix with 41 rows and 12 columns. Each row of the matrix corresponds to the simulation result of a simulation dead time, and each column corresponds to the component value of a harmonic order. For example, when the simulation dead time is 0.000002 seconds, the extracted harmonic component values ​​are: second harmonic 3.2V, third harmonic 4.8V, fourth harmonic 2.1V...thirteenth harmonic 1.2V. Similarly, the same harmonic analysis is performed on the other 40 simulation scenarios. Finally, the 41-row and 12-column simulation harmonic component matrix is ​​divided by column to obtain 12 sequences, namely the second harmonic simulation sequence to the thirteenth harmonic simulation sequence. Each harmonic simulation sequence contains 41 data points, corresponding to the harmonic component values ​​of this order under 41 different simulation dead times.

[0083] Furthermore, the method of fitting the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence using a nonlinear fitting algorithm, recorded as the second function to the Nth function, includes:

[0084] Establish the second harmonic simulation sequence to the Nth harmonic simulation sequence from the first data to the The data are the same as the first to the second data in the simulation dead zone time series. The mapping relationship of the data.

[0085] The simulation dead zone time series is from the first data to the The data are taken as independent variables, and the first data to the Nth data in the second harmonic simulation sequence to the Nth harmonic simulation sequence are taken as independent variables. The data are taken as dependent variables, and the nonlinear least squares method is used to fit the functional relationship between the second harmonic simulation value to the Nth harmonic simulation value and the simulation dead time, which are recorded as the second function to the Nth function respectively.

[0086] Exemplarily, after obtaining the second harmonic simulation sequence to the thirteenth harmonic simulation sequence, the mapping relationship between the 41 data in the second harmonic simulation sequence to the thirteenth harmonic simulation sequence and the corresponding 41 data in the simulation dead time sequence is established. Taking the second harmonic simulation sequence as an example, the simulation dead time of 0.000002 seconds is corresponding to the second harmonic amplitude of 3.2V, the simulation dead time of 0.0000021 seconds is corresponding to the second harmonic amplitude of 3.4V, the simulation dead time of 0.0000022 seconds is corresponding to the second harmonic amplitude of 3.5V, and so on, until the simulation dead time of 0.000006 seconds is corresponding to the second harmonic amplitude of 8.1V, forming a complete 41 groups of data mapping relationships. The same method is used to establish the data mapping relationship between the third harmonic simulation sequence to the thirteenth harmonic simulation sequence and the simulation dead time sequence.

[0087] The nonlinear least squares method is used to fit each group of mapping relationships. The simulation dead time is taken as the independent variable, and the corresponding data in the second harmonic simulation sequence to the thirteenth harmonic simulation sequence is taken as the dependent variable. Considering the possible nonlinear relationship between the simulation dead time and the data in the second harmonic simulation sequence to the thirteenth harmonic simulation sequence, a cubic polynomial is selected as the fitting function. The fitting calculation is performed through the fitting toolbox of MATLAB, and 12 functional relationships are obtained, which are recorded as the second function to the thirteenth function respectively. Taking the second function as an example, the functional relationship obtained by fitting shows that when the simulation dead time increases from 0.000002 seconds to 0.0000032 seconds, the second harmonic simulation data rises slowly from 3.2V to 4.8V; when the simulation dead time continues to increase to 0.0000045 seconds, the second harmonic simulation data rises rapidly to 6.9V; when the simulation dead time further increases to 0.000006 seconds, the growth rate of the second harmonic simulation data slows down and finally reaches 8.1V. The functional relationships of other harmonic orders also show nonlinear characteristics, but the specific change trends and directions are different.

[0088] Furthermore, the method of constructing a first objective function and a first constraint condition according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and using a particle swarm optimization algorithm to calculate the optimal opening time compensation value includes:

[0089] A first objective function is constructed according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and the first objective function is expressed as:

[0090] ;

[0091] in, represents the first objective function; Indicates the opening time compensation value in seconds; Indicates the second interval time in seconds; Indicates the first sampling time interval, in seconds; The value is 2 to integer variable of ; The value range is 1 to integer variable of ; For the The first in the harmonic voltage amplitude sequence Data, in volts; Indicates function; Indicates the simulation dead time in seconds; Indicates the rated dead time in seconds.

[0092] Construct a first constraint condition, which is expressed as:

[0093] ;

[0094] in, Indicates the preset lower limit of the dead zone, in seconds; Indicates the preset upper limit of the dead zone, in seconds.

[0095] Taking the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value.

[0096] Exemplarily, the first objective function is constructed based on the second harmonic voltage amplitude sequence to the thirteenth harmonic voltage amplitude sequence obtained within the first sampling time interval of 300 seconds, and the second function to the thirteenth function obtained by nonlinear fitting. The first objective function consists of two parts. The first part measures the size of the second harmonic voltage amplitude sequence to the thirteenth harmonic voltage amplitude sequence obtained by measurement, and the second part measures the product of the derivative of the second function to the thirteenth function with respect to the simulation dead time and the opening time compensation value, reflecting the influence of the opening time compensation value on the harmonic voltage amplitude. The first objective function obtained by combining the two parts comprehensively reflects the size of the total distortion amplitude of the harmonic voltage after the rated dead time and the opening time compensation value are superimposed. The pre-set , pre-set . Therefore, the first constraint condition is obtained: the dead time after compensation must be greater than or equal to 0.000002 seconds and less than or equal to 0.000006 seconds. The first constraint condition ensures that the final optimized dead time is within the practical acceptable range and will not cause a short circuit caused by the simultaneous conduction of the upper and lower IGBT switches of the three-phase six-bridge arm of the inverter. The constructed objective function and constraint conditions are input into the particle swarm optimization algorithm to solve and obtain the optimal opening time compensation value.

[0097] Further, the method for calculating the optimal dead time according to the rated dead time and the optimal opening time compensation value includes:

[0098] The optimal dead time is calculated according to the rated dead time and the optimal opening time compensation value. The calculation formula of the optimal dead time is:

[0099] ;

[0100] in, represents the optimal opening time compensation value, Indicates the optimal dead time.

[0101] For example, the optimal on-time compensation value calculated by the particle swarm optimization algorithm is -0.0000008 seconds, combined with the preset rated dead time of 0.000004 seconds, the optimal dead time is 0.0000032 seconds. The calculated optimal dead time of 0.0000032 seconds is input into the SVPWM controller, and the three-phase six-bridge arm inverter is restarted to drive the load motor, so that the harmonic voltage at the input end of the load motor is reduced while ensuring the dead zone compensation effect.

[0102] Embodiment 2: Based on the same inventive concept, Figure 2 As shown, this embodiment also provides an SVPWM dead zone compensation system based on artificial intelligence, and the system includes:

[0103] The data reading module is used to obtain first data, wherein the first data includes a preset rated dead time, a first sampling time interval, a first interval time, a second interval time, and a truncated harmonic number N.

[0104] The data acquisition module is connected to the data reading module and is used to input the rated dead time into the SVPWM controller, start the three-phase six-bridge arm inverter, and drive the load motor to run; within the first sampling time interval, the first interval time is used as the sampling time interval, the input terminal voltage of the load motor is collected, and according to the truncated harmonic number, the second interval time is used as the time window for discrete Fourier transform, and the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence are obtained.

[0105] The optimization calculation module is connected to the data acquisition module and is used to establish a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, input a preset simulation dead time sequence, and run to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence; a nonlinear fitting algorithm is used to fit the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence, which are recorded as the second function to the Nth function; a first objective function and a first constraint condition are constructed according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and the optimal opening time compensation value is calculated using a particle swarm optimization algorithm.

[0106] The result output module is connected to the optimization calculation module and is used to calculate the optimal dead time according to the rated dead time and the optimal opening time compensation value; and input the optimal dead time into the SVPWM controller.

[0107] Furthermore, the optimization calculation module includes:

[0108] A simulation model building module is used to establish a simulation model of a three-phase six-bridge-arm inverter and a load motor in a simulation platform according to a circuit connection topology structure of the three-phase six-bridge-arm inverter and the load motor, which is recorded as a first simulation model; establish an SVPWM controller simulation model according to the SVPWM controller; and embed the SVPWM controller simulation model into the first simulation model.

[0109] The simulation calculation module is connected to the simulation model construction module, and is used to input the preset simulation dead time sequence into the SVPWM controller simulation model respectively to obtain the first simulation scene to the Mth simulation scene; wherein M represents the number of data in the simulation dead time sequence; the first simulation scene to the Mth simulation scene are set in the first simulation model, and the first simulation model is run; the second harmonic component to the Nth harmonic component of the load motor input terminal voltage under the first simulation scene to the Mth simulation scene are extracted to obtain the simulation harmonic component matrix ; The simulation harmonic component matrix It is expressed as:

[0110] ;

[0111] in, Indicates The voltage at the input terminal of the load motor in the simulation scenario Subharmonic components; The value range is 1 to integer variable of ; The value is 2 to An integer variable.

[0112] A harmonic simulation sequence building module is connected to the simulation calculation module and is used to convert the simulation harmonic component matrix Split into second harmonic simulation sequence to Nth harmonic simulation sequence by column.

[0113] Furthermore, the optimization calculation module also includes:

[0114] The data mapping module is used to establish the first data to the second harmonic simulation sequence to the Nth harmonic simulation sequence. The data are the same as the first to the second data in the simulation dead zone time series. The mapping relationship of the data.

[0115] The function fitting module is connected to the data mapping module and is used to simulate the dead zone time series from the first data to the The data are taken as independent variables, and the first data to the Nth data in the second harmonic simulation sequence to the Nth harmonic simulation sequence are taken as independent variables. The data are taken as dependent variables, and the nonlinear least squares method is used to fit the functional relationship between the second harmonic simulation value to the Nth harmonic simulation value and the simulation dead time, which are recorded as the second function to the Nth function respectively.

[0116] Furthermore, the optimization calculation module also includes:

[0117] The first objective function construction module is used to construct a first objective function according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, wherein the first objective function is expressed as:

[0118] ;

[0119] in, represents the first objective function; Indicates the opening time compensation value in seconds; Indicates the second interval time in seconds; Indicates the first sampling time interval, in seconds; The value is 2 to integer variable of ; The value range is 1 to integer variable of ; For the The first in the harmonic voltage amplitude sequence Data, in volts; Indicates function; Indicates the simulation dead time in seconds; Indicates the rated dead time in seconds.

[0120] The first constraint condition building module is connected to the first objective function building module and is used to construct a first constraint condition, wherein the first constraint condition is expressed as:

[0121] ;

[0122] in, Indicates the preset lower limit of the dead zone, in seconds; Indicates the preset upper limit of the dead zone, in seconds.

[0123] Taking the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value.

[0124] Furthermore, the result output module includes:

[0125] The optimal dead time calculation module is used to calculate the optimal dead time according to the rated dead time and the optimal opening time compensation value. The calculation formula of the optimal dead time is:

[0126] ;

[0127] in, represents the optimal opening time compensation value, Indicates the optimal dead time.

[0128] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0129] Finally, it should be noted that: Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. The SVPWM dead zone compensation method based on artificial intelligence is characterized by: The method comprises the following steps: S1, obtaining first data, wherein the first data includes a preset rated dead time, a first sampling time interval, a first interval time, a second interval time, and a truncated harmonic number N; S2, input the rated dead time into the SVPWM controller, start the three-phase six-bridge-arm inverter, and drive the load motor to run; within the first sampling time interval, use the first interval time as the sampling time interval, collect the voltage at the input end of the load motor, and perform discrete Fourier transform with the second interval time as the time window according to the truncated harmonic order, and obtain the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence; S3, establishing a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, inputting a preset simulation dead time sequence, and running to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence; respectively using a nonlinear fitting algorithm to fit the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence, recorded as the second function to the Nth function; constructing a first objective function and a first constraint condition according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and using a particle swarm optimization algorithm to calculate the optimal opening time compensation value; S4, calculating the optimal dead time according to the rated dead time and the optimal on time compensation value; inputting the optimal dead time into the SVPWM controller; The method of establishing a simulation model of a three-phase six-bridge-arm inverter and a load motor by using a digital twin algorithm, inputting a preset simulation dead time sequence, and running to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence includes: According to the circuit connection topology structure of the three-phase six-bridge-arm inverter and the load motor, a simulation model of the three-phase six-bridge-arm inverter and the load motor is established in the simulation platform, which is recorded as the first simulation model; according to the SVPWM controller, an SVPWM controller simulation model is established; and the SVPWM controller simulation model is embedded in the first simulation model; The preset simulation dead time sequence is input into the SVPWM controller simulation model respectively to obtain the first simulation scene to the Mth simulation scene; wherein M represents the number of data in the simulation dead time sequence; the first simulation scene to the Mth simulation scene are set into the first simulation model, and the first simulation model is run; the second harmonic component to the Nth harmonic component of the input terminal voltage of the load motor under the first simulation scene to the Mth simulation scene are extracted to obtain the simulation harmonic component matrix ; The simulation harmonic component matrix It is expressed as: ； in, Indicates The voltage at the input terminal of the load motor in the simulation scenario Subharmonic components; The value range is 1 to integer variable of ; The value is 2 to integer variable of ; The simulation harmonic component matrix The simulation is divided into second harmonic simulation sequence to Nth harmonic simulation sequence by column.

2. The SVPWM dead zone compensation method based on artificial intelligence as claimed in claim 1, characterized in that: The method of fitting the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence using a nonlinear fitting algorithm, recorded as the second function to the Nth function, includes: Establish the second harmonic simulation sequence to the Nth harmonic simulation sequence from the first data to the The data are the same as the first to the second data in the simulation dead zone time series. The mapping relationship of data; The simulation dead zone time series is from the first data to the The data are taken as independent variables, and the first data to the Nth data in the second harmonic simulation sequence to the Nth harmonic simulation sequence are taken as independent variables. The data are taken as dependent variables, and the nonlinear least squares method is used to fit the functional relationship between the second harmonic simulation value to the Nth harmonic simulation value and the simulation dead time, which are recorded as the second function to the Nth function respectively.

3. The SVPWM dead zone compensation method based on artificial intelligence as claimed in claim 2, characterized in that: The method of constructing a first objective function and a first constraint condition according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and using a particle swarm optimization algorithm to calculate an optimal opening time compensation value includes: A first objective function is constructed according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and the first objective function is expressed as: ； in, represents the first objective function; Indicates the opening time compensation value in seconds; Indicates the second interval time in seconds; Indicates the first sampling time interval, in seconds; The value is 2 to integer variable of ; The value range is 1 to integer variable of ; For the The first in the harmonic voltage amplitude sequence Data, in volts; Indicates function; Indicates the simulation dead time in seconds; Indicates the rated dead time in seconds; Construct a first constraint condition, which is expressed as: ； in, Indicates the preset lower limit of the dead zone, in seconds; Indicates the preset upper limit of the dead zone, in seconds; Taking the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value.

4. The SVPWM dead zone compensation method based on artificial intelligence as claimed in claim 3, characterized in that: The method for calculating the optimal dead time according to the rated dead time and the optimal opening time compensation value comprises: The optimal dead time is calculated according to the rated dead time and the optimal opening time compensation value. The calculation formula of the optimal dead time is: ； in, represents the optimal opening time compensation value, Indicates the optimal dead time.

5. The SVPWM dead zone compensation system based on artificial intelligence is characterized by: The system comprises: A data reading module, used for acquiring first data, wherein the first data includes a preset rated dead time, a first sampling time interval, a first interval time, a second interval time, and a truncated harmonic number N; The data acquisition module is connected to the data reading module and is used to input the rated dead time into the SVPWM controller, start the three-phase six-bridge arm inverter, and drive the load motor to run; within the first sampling time interval, the first interval time is used as the sampling time interval, the input terminal voltage of the load motor is collected, and according to the truncated harmonic number, the second interval time is used as the time window to perform discrete Fourier transform, and obtain the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence; The optimization calculation module is connected to the data acquisition module and is used to establish a simulation model of a three-phase six-bridge-arm inverter and a load motor through a digital twin algorithm, input a preset simulation dead time sequence, and run to obtain a second harmonic simulation sequence to an Nth harmonic simulation sequence; a nonlinear fitting algorithm is used to fit the functional relationship between the second harmonic simulation sequence to the Nth harmonic simulation sequence and the simulation dead time sequence, which are recorded as the second function to the Nth function; a first objective function and a first constraint condition are constructed according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, and a particle swarm optimization algorithm is used to calculate the optimal opening time compensation value; A result output module is connected to the optimization calculation module and is used to calculate the optimal dead time according to the rated dead time and the optimal opening time compensation value; and input the optimal dead time into the SVPWM controller; The optimization calculation module comprises: A simulation model building module is used to establish a simulation model of a three-phase six-bridge-arm inverter and a load motor in a simulation platform according to a circuit connection topology structure of the three-phase six-bridge-arm inverter and the load motor, which is recorded as a first simulation model; establish an SVPWM controller simulation model according to the SVPWM controller; and embed the SVPWM controller simulation model into the first simulation model; The simulation calculation module is connected to the simulation model construction module, and is used to input the preset simulation dead time sequence into the SVPWM controller simulation model respectively to obtain the first simulation scene to the Mth simulation scene; wherein M represents the number of data in the simulation dead time sequence; the first simulation scene to the Mth simulation scene are set in the first simulation model, and the first simulation model is run; the second harmonic component to the Nth harmonic component of the load motor input terminal voltage under the first simulation scene to the Mth simulation scene are extracted to obtain the simulation harmonic component matrix ; The simulation harmonic component matrix It is expressed as: ； in, Indicates The voltage at the input terminal of the load motor in the simulation scenario Subharmonic components; The value range is 1 to integer variable of ; The value is 2 to integer variable of ; A harmonic simulation sequence building module is connected to the simulation calculation module and is used to convert the simulation harmonic component matrix The simulation is divided into second harmonic simulation sequence to Nth harmonic simulation sequence by column.

6. The SVPWM dead zone compensation system based on artificial intelligence as claimed in claim 5, characterized in that: The optimization calculation module also includes: The data mapping module is used to establish the first data to the second harmonic simulation sequence to the Nth harmonic simulation sequence. The data are the same as the first to the second data in the simulation dead zone time series. The mapping relationship of data; The function fitting module is connected to the data mapping module and is used to simulate the dead zone time series from the first data to the The data are taken as independent variables, and the first data to the Nth data in the second harmonic simulation sequence to the Nth harmonic simulation sequence are taken as independent variables. The data are taken as dependent variables, and the nonlinear least squares method is used to fit the functional relationship between the second harmonic simulation value to the Nth harmonic simulation value and the simulation dead time, which are recorded as the second function to the Nth function respectively.

7. The artificial intelligence-based SVPWM dead zone compensation system according to claim 6, characterized in that: The optimization calculation module also includes: The first objective function construction module is used to construct a first objective function according to the second harmonic voltage amplitude sequence to the Nth harmonic voltage amplitude sequence and the second function to the Nth function, wherein the first objective function is expressed as: ； in, represents the first objective function; Indicates the opening time compensation value in seconds; Indicates the second interval time in seconds; Indicates the first sampling time interval, in seconds; The value is 2 to integer variable of ; The value range is 1 to integer variable of ; For the The first in the harmonic voltage amplitude sequence Data, in volts; Indicates function; Indicates the simulation dead time in seconds; Indicates the rated dead time in seconds; The first constraint condition building module is connected to the first objective function building module and is used to construct a first constraint condition, wherein the first constraint condition is expressed as: ； in, Indicates the preset lower limit of the dead zone, in seconds; Indicates the preset upper limit of the dead zone, in seconds; Taking the minimum value of the first objective function as the goal and the first constraint condition as the constraint, the particle swarm optimization algorithm is used to calculate the optimal opening time compensation value.

8. The artificial intelligence-based SVPWM dead zone compensation system according to claim 7, characterized in that: The result output module includes: The optimal dead time calculation module is used to calculate the optimal dead time according to the rated dead time and the optimal opening time compensation value. The calculation formula of the optimal dead time is: ； in, represents the optimal opening time compensation value, Indicates the optimal dead time.

Images